Citation
Finding the components of the primary language focusing on the algorithm of discovery

Material Information

Title:
Finding the components of the primary language focusing on the algorithm of discovery
Creator:
Alharbi, Nesreen
Place of Publication:
Denver, CO
Publisher:
University of Colorado Denver
Publication Date:
Language:
English

Thesis/Dissertation Information

Degree:
Doctorate ( Doctor of philosophy)
Degree Grantor:
University of Colorado Denver
Degree Divisions:
Department of Computer Science and Engineering, CU Denver
Degree Disciplines:
Computer science and information systems
Committee Chair:
Alaghband, Gita
Committee Members:
Stilman, Boris
Altman, Tom
Karimi, Jahangir
Biswas, Ashis

Record Information

Source Institution:
University of Colorado Denver
Holding Location:
Auraria Library
Rights Management:
Copyright [name of copyright holder or Creator or Publisher as appropriate]. Permission granted to University of Colorado Denver to digitize and display this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.

Downloads

This item has the following downloads:


Full Text
FINDING THE COMPONENTS OF THE PRIMARY LANGUAGE
FOCUSING ON THE ALGORITHM OF DISCOVERY
by
NESREEN ALHARBI
B.S., King Abdulaziz University-Jeddah Saudi Arabia, 2001 M.S., University of Colorado- Denver USA, 2009
A thesis submitted to the Faculty of the Graduate School of the University of Colorado in partial fulfillment of the requirements for the degree of Doctor of Philosophy
Computer Science and Information Systems Program
2018


This thesis for the Doctor of Philosophy degree by Nesreen Alharbi has been approved for the
Computer Science and Information Systems Program
by
Gita Alaghband, Chair Boris Stilman, Advisor Tom Altman Jahangir Karimi Ashis Biswas
Date May 12, 2018


Alharbi, Nesreen. (Ph.D., Computer Science and Information Systems Program)
Finding the Components of the Primary Language - Focusing on the Algorithm of Discovery Thesis directed by Professor Boris Stilman
ABSTRACT
In the late nineteen fifties, Professor John von Neumann suggested that the human brain uses an internal language for mental calculation. He named that language the Primary Language. He suggested that the Primary Language differs from all the human (Secondary) languages used for communication. We consider Primary Science to be the science done with the Primary Language, that is differs from the familiar, conventional science.
The ultimate goal of this research is to reveal the nature of the Primary Language. To accomplish that, it is important to find those algorithms, used by the human brain, based directly on this language. The results of previous research revealed the existence of at least two ancient algorithms critical for the development of human intelligence. It is our assumption that in absence of the secondary languages, during times long past, those algorithms would have directly utilized the Primary Language. These two algorithms are Linguistic Geometry (LG) the algorithm for optimizing warfighting strategies, and the Algorithm of Discovery (AD) the algorithm for inventing new algorithms.
The main hypothesis of this research is that the Primary Language is the “language” of visual streams (mental “movies”). Another hypothesis states that the AD is a universal algorithm used for making discoveries. It suggests that the AD is based on multiple thought experiments, which manifest themselves via visual streams. It appears that visual streams are the only interface to the AD. The AD operates with three classes of visual streams: observation, construction, and validation. These visual streams can run concurrently and exchange information. Each stream may initiate additional thought experiments, program
in


them, and then execute them. Visual streams are used by the AD to construct new algorithms and, in this way, make discoveries.
The current objective is to further investigate the AD by applying it to various discoveries, from different domains, such as Computer Science, Molecular Biology, etc., in order to reveal its inherent details. During this research, it was found that some components of the AD are utilized for every discovery, while others may be utilized for specific discoveries only. Once the major elements of the AD are revealed, generalizing them will lead to a complete understanding of the AD.
A comprehensive understanding of the AD and its components will lead to the implementation of the generalized AD, i.e., the final goal of making discoveries on demand. The expectation is that the implemented AD will have a profound impact on all branches of science, including Computer Science, and, in particular, Artificial Intelligence (AI).
The form and content of this abstract is approved. I recommend its publication.
Approved: Boris Stilman
IV


I Dedicate This Thesis to My Husband, Yousef My Parents, Ghuzaiel and Mohammed My Daughters, Maria, Lin, and Dana Without your support and love I could not have done it.
v


ACKNOWLEDGMENTS
First and foremost, I would like to thank my advisor, Professor Boris Stilman, to whom I am so grateful for the amount of time that he spent guiding and supporting me through my research. I thank him for his patience, enthusiasm, motivation, and guidance throughout this entire process. I would never have been able to finish my dissertation without his continuous advising and immense knowledge.
My sincere thanks to my committee members Professor Gita Alaghband, Professor Tom Altman, Professor Jahangir Karimi, and Professor Ashis Biswas for their encouragement and valuable feedback. You all have made my defense an enjoyable event. I would like to especially thank Professor Gita Alaghband, the chair of my committee, from whom I have been receiving extensive guidance in both personal and professional aspects since the beginning of my PhD studies. Thank you Professor Alaghband, nothing would have been achieved without your support.
I would like to express my extreme thanks to Professor Christopher Miller, Department of Integrative Biology at University of Colorado at Denver, for all the time and effort he spent reviewing the accuracy of the analysis of the genetic code and the protein translation in this research. Also, I would like to thank Professor Altman, from the bottom of my heart, for reviewing my proposal and providing me with valuable feedback that helped improve my work. In addition, I would like to thank Mrs. Hanna Altman for helping me check the accuracy of the biology information in my proposal, and I would like to thank Diane Yoha, Sarah Mandos, and all the staff of the Computer Science department for their assistance and support.
vi


In addition, my appreciation and gratitude to the government of Saudi Arabia for granting me a full scholarship that enabled me to pursue my PhD studies and to attend this great university. Finally, and most importantly, my sincere thanks to my amazing family. My husband Yousef who had to live far away from us in order to give me the opportunity to study. My daughters Maria, Lin, and Dana who picked up the slack at home and helped each other, becoming independent little ladies and allowing mom to study. Big thanks to my father Mohammed who left home to come with me and helped with the upbringing of my daughters while they were young. Another big thanks to my mother Ghuzaiel for traveling all the way from Saudi Arabia to the United stated for my sake. She spent four months taking care of my little daughter so that I would have enough time to prepare for the preliminary exam; her presence was such a huge relief. Their continuous love and support for my studies is the reason why this research is completed.
vn


TABLE OF CONTENTS
CHAPTER
I. INTRODUCTION................................................................1
1.1.0 Overview...............................................................1
1.2.0 Method of Study........................................................2
1.3.0 Motivation.............................................................3
1.4.0 Objectives.............................................................3
1.5.0 Work Domain............................................................4
1.6.0 Challenges.............................................................5
1.7.0 C ontributi on.........................................................6
1.8.0 Dissertation Structure.................................................8
II. THE PRIMARY LANGUAGE..................................................... 10
2.1.0 Overview..............................................................10
2.2.0 The Primary Language (PL).............................................10
2.3.0 Visual Streams and Thought Experiments................................11
2.4.0 Types of Visual Streams...............................................13
2.4.1 Discovery Streams and the Algorithm of Discovery (AD)..............14
2.5.0 Proximity and Mosaic reasoning........................................16
III. COMMUNICATION STREAMS AND ALGORITHMIC MOSAICS............................19
3.1.0 Overview..............................................................19
3.2.0 Motivation............................................................19
3.3.0 Communication Streams.................................................20
3.4.0 Algorithmic and Symbolic Mosaics......................................22
viii


3.5.0 Results....................................................................24
IV. INTRODUCTION TO THE INVESTIGATED DISCOVERIES..................................26
4.1.0 Introduction to Linguistic Geometry........................................26
4.1.1 Trajectories............................................................26
4.1.2 Introduction to the Central Dogma.......................................28
4.2.0 DNA and RNA................................................................29
4.3.0 The Genetic Code...........................................................30
4.4.0 Protein Synthesis..........................................................31
V. REVISITING THE ALGORITHM OF ADMISSIBLE TRAJECTORIES...........................33
5.1.0 Overview...................................................................33
5.2.0 Motivation.................................................................34
5.3.0 FindMid Experiment.........................................................34
5.3.1 The Observation Stream..................................................34
5.3.2 The Construction Stream and the Expression Stream.......................36
5.4.0 Define the Mosaic's Aggregates and Matching Rules..........................38
5.5.0 GlueTree Experiment........................................................40
5.5.1 The Observation Stream..................................................40
5.5.2 The Construction Stream.................................................40
5.6.0 Results....................................................................45
VI. REVISITING THE DISCOVERY OF GENETIC CODE......................................46
6.1.0 Overview...................................................................46
6.2.0 Motivation.................................................................47
6.3.0 The Gamow Thought Experiments (The Diamond Code)...........................47
IX


6.3.1 Building the Input Mosaic..................................................48
6.3.2 Generating the “Construct Input Mosaic” Shell..............................51
6.3.3 Identifying the Output Mosaic..............................................53
6.3.4 Generating the “Construct Output Mosaic” Shell.............................54
6.3.5 Defining the Intermediate Mosaic...........................................55
6.3.6 Assigning the Reading Rules................................................57
6.3.7 Identifying the Matching Rules.............................................60
6.3.8 Defining the Code and Amino Sets...........................................61
6.3.9 Constructing the Intermediate Mosaic.......................................63
6.4.0 The Crick Thought Experiments................................................65
6.4.1 Validating the Diamond Code................................................65
6.4.2 Redefining the Input Mosaic................................................67
6.4.3 Evaluating the Reading Rules...............................................69
6.4.4 Rebuilding the Amino Set...................................................71
6.5.0 Modifying the Reading Rules..................................................72
6.6.0 Deciphering the First Triple.................................................73
6.6.1 Defining the Input Mosaic..................................................74
6.6.2 Generating the “Construct information mRNA” Shell..........................75
6.6.3 Translation Components.....................................................76
6.6.4 The Mapping Thought Experiments............................................77
6.7.0 Results......................................................................80
VII. REVISITING THE DISCOVERY OF PROTEIN TRANSLATION.................................82
7.1.0 Overview.....................................................................82
x


7.2.0 Motivation......................................................................83
7.3.0 Applying the AD on the Discovery of tRNA Activation.............................83
7.3.1 The Adapter Hypothesis Thought Experiments....................................84
7.3.2 Identifying the Process of Activating Amino Acids.............................85
7.3.3 Identifying the Process Rules to Link Amino Acids to tRNA.....................95
7.4.0 Finding the Rules to Attach tRNA and mRNA.......................................99
7.4.1 Defining Matching rules to link mRNA and tRNA Mosaics........................100
7.4.2 Discovering Matching Rules...................................................101
7.5.0 The Watson Model...............................................................Ill
7.5.1 Defining the Structure of the Ribosome.......................................117
7.5.2 Identify the Matching Rules between the mRNA and the Ribosome................118
7.5.3 Finding the Plugging Sites for tRNA into the Ribosome Mosaic.................119
7.5.4 Constructing the Mechanism of Elongation.....................................120
7.6.0 The Watson-extension Model.....................................................128
7.6.1 Validating the Nierhaus model................................................131
7.6.2 Declaring the Matching Rules for Occupying the E-Site........................133
7.6.3 Reconstructing the Translation Process.......................................136
7.7.0 The Alternative Model: Rebuilding the Elongation Model.........................143
7.7.1 Preparing the Construction Set for Elongation: the Observation Stream........146
7.7.2 Reconstructing Elongation....................................................148
7.7.3 Verifying the New Three-site Model: The Validation Stream....................156
7.8.0 The Hybrid Model of Elongation.................................................157
7.8.1 Declaring the First Process Rule: The Observation Stream.......................159
xi


7.8.2 The Observation Stream: Declaring the Second Process Rule.................161
7.8.3 The Observation Stream: Declaring More Process Rules......................163
7.8.4 Building the Model: The Construction Stream...............................165
7.9.0 Results...................................................................173
\ 111. CONTRIBUTIONS AM) FUTURE WORK............................................. 178
8.1.0 Overview..................................................................178
8.2.0 Contributions.............................................................178
8.2.1 New Matching Rules.....................................................178
8.2.2 Investigation of New Streams...........................................181
8.2.3 New Types of Mosaics...................................................181
8.2.4 New Transformation Phases of Protein Synthesis.........................184
8.2.5 New Scope Sets.........................................................185
8.2.6 Identifying a New Stage of the Observation Stream......................185
8.2.7 Identifying New Thought Experiments...................................186
8.3.0 Towards Automation of the AD..............................................188
8.4.0 Future Work...............................................................191
BIBLIOGRAPHY.....................................................................193
xii


LIST OF TABLES
TABLE
1. Different types of internal streams...............................................14
2. Several instruction aggregates introduced during this research....................23
3. One-dimensional letter tiles mosaic based on the DNA base tiles in Figure 9.......51
4. The code mosaic generated by the expression stream as Shown in [42], [7]..........63
5. One-dimensional mosaic of the mRNA letter tiles converted from the DNA base tiles.75
6. The final genetic code mosaic, [37]...............................................80
7. The RNA complementarity matching mosaic between the first two tiles in the mRNA and
tRNA.............................................................................103
8. The wobble complementarity matching mosaic between the third tile in the mRNA and tR 103
9. The updated wobble complementarity matching mosaic................................106
10. The updated wobble complementarity matching mosaic...............................107
11. The final wobble complementarity matching mosaic.................................110
xiii


LIST OF FIGURES
FIGURE
1. The communication streams pass information between the internal streams and the outer world
and vice versa.......................................................................21
2. Two different trajectories between points x; and xj...................................27
3. Visual representation of the initial input model......................................36
4. The visual model after being morphed by the construction stream.......................37
5. Visual simulation of the GlueTree thought experiment..................................41
6. Building a Mid-aggregate from the To-Mid and From-Mid aggregates......................42
7. The visual representation of the mosaic of the tree of the admissible trajectories....44
8. The final algorithmic mosaic corresponded to the construction of the admissible trajectories
tree.................................................................................44
9. The schematic representation of the double helix, [10]................................49
10. Part of the input mosaic as shown in [64].............................................50
11. “Construct Input mosaic” shell generated by the expression stream.....................53
12. “Construct Output mosaic” shell generated by the expression stream....................55
13. The code mosaic sits between the cavities of the input mosaic as shown in [42], [7]...57
14. Four symmetrical code aggregates could be generated from the genetic code 123 by the
construction stream, [7].............................................................58
15. Part of the intermediate mosaic constructed form the genetic sequence “123123”, [7]...65
16. The code aggregates generated from the input sequence 1112n...........................67
17. Constructing the RNA mosaic from a single strand of the DNA mosaic....................69
18. “Construct information mRNA” shell generated by the expression stream.................76
19. An mRNA mosaic of Poly-U code aggregates, [7].........................................78
20. The mapping between the code and amino aggregates, [7]...............................79
21. The input aggregate: Acetate.........................................................87
xiv


87
22. The input aggregate: Acetyl-CoA.................................................
23. CO-NH peptide bond in peptide mosaic retrieved from a previous thought experiment...88
24. General representation of amino aggregates..........................................90
25. The ATP aggregate....................................................................90
26. First scenario CO binds with Ad.P...................................................90
27. Second scenario CO binds with Ad.P~P................................................91
28. Third scenario CO binds with P~P....................................................91
29. Fourth scenario CO binds with P......................................................91
30. The shell corresponded to constructing the ATP aggregate.............................92
31. The shell corresponded to constructing the amino aggregate...........................93
32. The shell generated to construct the Active amino aggregate..........................93
33. Active amino aggregate carried by the enzyme backbone as visualized by the construction
stream...............................................................................94
34. The final shell to construct the active amino aggregate..............................95
35. The first step of the translation process: constructing the amino-tRNA...............98
36. “Construct amino-tRNA” shell.........................................................99
37. The third tile’s matching rule is unknown, [6]......................................102
38. The construction stream failed to build the G-A aggregate, [6]......................105
39. The construction stream succeeded in building the U-C aggregate, [6]................106
40. The G-U aggregate, [6]..............................................................107
41. The I-A aggregate, [6]...............................................................107
42. The physical position of the bonds for the pair U-C.................................108
43. The physical position of the bonds violates the physical rule: too close bonds......109
44. The physical position of the bonds obeys the physical rule..........................110
45. The complementarity rules that control the binding of mRNA and tRNA, [6]............110
46. The tRNA mosaic before folding as represented by the construction stream............112
xv


47. The tRNA mosaic after folding as represented by the construction stream.........113
48. “Construct tRNA body” shell.....................................................113
49. The modified “Construct tRNA body” shell.......................................114
50. Three types of tRNA mosaics built by the construction stream: (a) amino-tRNA, (b) peptide-
tRNA, and (c) deacylated-tRNA..................................................114
51. “Construct amino-tRNA” shell....................................................115
52. “Construct peptide-tRNA” shell..................................................115
53. Constructing the mRNA mosaic....................................................116
54. The translation process as defined by previous thought experiments, [46]........117
55. A general shape of the 50-s and 30-s mosaics....................................121
56. Constructing the ribosome mosaic................................................121
57. The shell corresponded to the construction of the ribosome mosaic...............122
58. The mRNA enters the ribosome mosaic.............................................122
59. The starting state in the mechanism of elongation identified by the construction stream.... 123
60. The amino-tRNA enters into the ribosomal A-site.................................124
61. The amino-tRNA occupies the A-site..............................................125
62. The peptide formation...........................................................125
63. The tRNA leaves the ribosome....................................................126
64. The peptide-tRNA translocates into the P-site...................................126
65. The next amino-tRNA enters the ribosome.........................................127
66. The “Elongation” shell for the Watson model.....................................127
67. The mRNA mosaic.................................................................129
68. Three Different Types of tRNA Mosaics...........................................129
69. The modified “Construct tRNA body” shell.......................................130
70. The ribosome mosaic built by the construction stream and tagged by the expression stream.
...............................................................................135
xvi


71. The shell corresponded to the construction of the ribosome mosaic.................135
72. The mRNA mosaic takes a U-turn shape inside the ribosome..........................136
73. The “Link the ribosome to the mRNA” shell generated by the expression stream......136
74. Different snapshots of the initiation process by the construction stream: (1) The current
codon is not an initiation codon the construction stream will read the next codon, (2) The current codon is an initiation codon, (2) call the initiator tRNA the fMet-tRNA, and (4) place the fMet-tRNA in the P-site and create the codon-anticodon aggregate.........138
75. The “Initiate” shell corresponded to the process in Figure 74......................138
76. One cycle of the elongation process as visualized by the construction stream.......140
77. The “Elongation” shell corresponded to the process in Figure 76....................140
78. The first scenario for a new version of the “Elongation” shell.....................141
79. The second scenario for a new version of the “Elongation” shell....................142
80. Reading a termination codon........................................................142
81. Terminating protein synthesis......................................................143
82. The “Termination” shell............................................................143
83. The new structure of the ribosome mosaic...........................................149
84. The “Construct ribosome” shell generated by the expression stream..................149
85. Three different types of tRNA mosaics as visualized by the construction stream.....150
86. The mRNA enters the ribosome.......................................................151
87. The starting cycle of elongation...................................................152
88. The amino-tRNA is ready to accept the peptide mosaic...............................153
89. Peptidyltransferase................................................................153
90. The tRNA mosaics translocation.....................................................154
91. The new amino-tRNA is ready to enter the elongation cycle and the deacylated-tRNA is
ready to move out of the cycle......................................................154
92. The beginning of a new elongation cycle............................................155
xvii


93. The algorithmic mosaic for the alternative elongation model built by the expression stream.. 155
94. The three classical states the tRNA mosaics can move to during elongation........160
95. The P/E state where deacylated-tRNA occupied the P-site of the 30-s mosaic and the E-site
of the 50-s mosaic at the same time.............................................161
96. The A/E state where peptide-tRNA occupied the A-site of the 30-s mosaic and the P-site of
the 50-s mosaic simultaneously..................................................162
97. The ribosome mosaic built by the construction stream.............................166
98. The initial scene to start elongation............................................166
99. The EF-Tu-GTP-amino-tRNA is in the A/T state.....................................167
100. The algorithmic mosaic explaining the process in Figure 99......................167
Figure 101. The amino-tRNA moves from the A/T state to the A/A state.................168
102. The elongation shell after translating the operations from Figure 101...........169
103. The peptide mosaic links to the amino aggregates and the deacylated-tRNA and the peptide-
tRNA moves to the states P/E and A/P respectively...............................169
104. The expression stream continues building the elongation shell by watching the visual movie
run by the construction stream...................................................170
105. The deacylated-tRNA and the peptide-tRNA move completely to the E-site and the P-site,
respectively.....................................................................171
106. “Elongation” Shell...............................................................171
107. The previous deacylated-tRNA must leave the ribosome before the new deacylated-tRNA
can enter the E-site.............................................................172
108. The final “Elongation” shell.....................................................172
109. The final algorithmic mosaic for the translation process.........................177
110. The stages of protein synthesis..................................................184
111. The components of the AD utilized in each case study.............................188
xviii


CHAPTER I
INTRODUCTION
1.1.0 Overview
In the late 1950’s, Professor John von Neumann [47] proposed the existence of an internal language used by the human brain for mental calculation. He named the brain’s language the Primary Language (PL). He assumed that the PL differs from other external languages humans use for communication, which he defined as secondary languages. The secondary languages include the natural languages and the languages of science, such as mathematics, computer science, etc. Von Neumann suggested that the PL is used for thought.
At the beginning of his research into the PL, Professor Stilman’s originally hypothesized that the PL included all major algorithms essential for humanity’s development and survival. During his investigations, Stilman discovered two algorithms that are critical to the development of human intelligence and revealed their relationship to the PL. They are Linguistic Geometry (LG), the algorithm for optimizing warfighting and the Algorithm of Discovery (AD), the algorithm for inventing new algorithms, [27], and [16],
Stilman concluded that the AD is based on multiple thought experiments, Section 2.3.0, which manifest themselves via visual streams (mental movies). It appears that, in the human brain, visual streams are the only interface to the AD. The AD operates with three classes of visual streams: observation, construction, and validation. Another class of visual streams, the communication streams, along with the AD, runs to translate the work of the AD’s streams. All classes of visual streams can run concurrently and exchange information. Each stream may initiate additional thought experiments, program, and then execute them, [27],
1


After extensive investigation, Stilman revised his original hypothesis, [27], and [16], The PL is not a collection of major algorithms; it is the Language of Visual Streams (LVS). As predicted by von Neumann, the PL is the base for all external languages. Stilman assumed that the PL is the engine that generates visual streams.
This research is based on the main hypothesis that the AD is used for inventing, or discovering, all other algorithms. The AD is based directly on the PL, the Language of Visual Streams. The visual streams are used by the AD to construct new algorithms and, in this way, make discoveries.
1.2.0 Method of Study
The method of study can be divided into two areas: case studies and software implementation. To reveal the nature of the AD, it was applied (as currently understood) to many case studies from two classes of discoveries: the field of LG, focusing on the admissible trajectories algorithm, and the field of Molecular Biology/Biochemistry, focusing on the discoveries of the genetic code and the protein synthesis. By applying the AD and its components, the various stages the AD followed on its way to making these discoveries were defined.
When the AD uncovers all stages of any innovation, the expression stream, a subset of the communication streams, generates the algorithm that guides the discovery. Then, the generated algorithm can be implemented as a computer program. The basic assumption is that if this program finds the correct solution to a class of diverse problems, then the discovered algorithm is correct. Eventually, the goal is to automate future discoveries, i.e., produce discoveries on demand.
2


1.3.0 Motivation
Every discovery starts with an idea, which motivates the discoverer. The spark, that motivated the work in this field, was the idea of having an algorithm that generates other algorithms, i.e., the algorithm that makes discoveries. The main question to be answered is “How, by following the work of others, can the entire nature of the AD be revealed and its components be found?” It was possible to answer this question, after reading the works of Stilman and various other papers related to this area of study. If computer programs could produce discoveries routinely as an output, a huge leap would be made by the humanity. Achieving the goal of making discoveries on demand would be the final outcome of this work.
This research started by working in the LG field, finding out that the AD led to the discovery of the algorithm for generating admissible trajectories. During this research, the AD is assumed to be universal, which means that any discovery will happen if the correct inputs, rules, and procedures are identified. Following this approach, it was found that the major innovations were made in stages employing various thought experiments and visual streams. During the thought experiments, the visual streams' engine may drop one algorithm and introduce another one according to the analysis phase. However, the AD that controls those streams is the same; it makes the discovery if the inputs and analysis are correct. Accurate analysis will generate new algorithms, which are the essence of any discovery.
1.4.0 Objectives
The objective of this research is to apply the AD to various discoveries to reveal its inherent details. In so doing, several components of the AD were investigated. In all of the case studies in this research, the AD was applied to two different fields of science, LG and
3


Molecular Biology/Biochemistry. Once the elements of the AD in any discovery are revealed, those elements can be generalized. Then, it is possible to implement and simulate the algorithm for that discovery. By executing the implementation of the algorithm, the AD’s reasoning components can be further analyzed and verified.
In every case study, usually there are short- and long-term goals. The short-term goal is to better understand the AD by deeply studying its applications, while applying it to selected case studies. The long-term goal is to automate future discoveries, i.e., develop a program for making them on demand.
1.5.0 Work Domain
The research’s work domain is the PL; the PL domain includes investigating all the different types of visual streams employed by the AD to make any discovery. As mentioned in Section 1.2.0, the case studies are from two different fields of science, the field of LG and the field of Molecular Biology/Biochemistry. When employing the AD, it is required to revisit all attempts to further refine the algorithm responsible for a specific discovery, including any known failed attempts. The failed attempts reveal as much information about the nature of the AD as do the successful ones. The AD analyzes the reason(s) for its unsuccessful execution, then it either returns to a specific waypoint or re-initializes itself to try to find new inputs and procedures guiding the construction in a different direction in an attempt to make the correct discovery.
During the analysis phase, the most difficult step was finding a clear description of all attempts made by the discoverers. For that reason, only discoveries with clear descriptions were considered as candidates for the AD's application and selected as case studies for this research. The first case study is the application of the AD to the discovery of the algorithm
4


for generating admissible trajectories in LG. Unfortunately, not all experiments were recorded, but by discussing these discoveries with Stilman it was possible to apply the AD and obtain good results.
The second case study is the discovery of the genetic code. Many scientists were involved in cracking the genetic code. Multiple written descriptions of failed experiments to decipher the code were found and studied. After finding a clear description of several recorded trials, the process of discovery was broken into smaller problems. The AD was applied to those smaller problems to reveal the components of the algorithm that guided the discovery of deciphering the genetic code.
The third case study is the discovery of the algorithm of translation. Again, multiple scientists participated in revealing the process of translation, the last step of protein synthesis. Only major experiments were examined in an effort to discover the algorithm used in protein synthesis. The descriptions of several trials were followed, including failed ones, and the problem was broken into several smaller problems. By applying the AD to these smaller problems, the components of the AD that led to the discovery of the algorithm of translation were successfully revealed.
1.6.0 Challenges
During this research, one goal was to study the PL by applying the AD to a non-computer science field. While analyzing the case studies in the Molecular Biology and Biochemistry fields, the main concepts of these fields had to be studied to better understand the reasoning that led to the conclusions of those case studies. Also, due to the nature of the discoveries in this field, many papers with different conclusions were found. Therefore, it was important to select only those papers that eventually led to the discovery of the algorithm of translation.
5


The most important part, of this area of research, was to determine which papers were instrumental in following the work of the AD. Otherwise, there was a risk of analyzing many papers only to discover those papers did not add any value to this research and, thus, wasted time.
After applying the AD on a series of thought experiments and revealing more components of the AD or finding a pattern of visual streams, the newly discovered components had to be applied to previous work. First, it was necessary to understand the nature of the new components, and second, to see if the new components were applicable to all discoveries or associated with discoveries of a special nature. This recursion significantly increased the research time because the AD had to be reapplied to previous work every time a new component of the AD was discovered. This, in turn, could lead to new findings, which must also be applied to previous work.
1.7.0 Contribution
While applying the AD to the genetic code and translation discoveries, new components of the AD were revealed. These components were found to be applicable to any similar discovery. New transformation rules that played a major role in revealing the genetic code and the translation algorithms were introduced. Movement rules were identified; these are part of the transformation rules guiding the movement of one mosaic (constructed object), while copying another to create a new mosaic. Another subset of the transformation rules are the reading rules, which identify the reading of information from one mosaic to construct another mosaic, based on that reading. Along with the movements rules come the alignment rules that define how to align two mosaics, if necessary.
6


The process rules are another subset of the transformation rules and appeared initially during the analysis of the translation discovery. The process rules must be identified in any discovery requiring multiple mosaic actions, whether in sequence or in parallel. Such rules are responsible for creating the order of the actions for every mosaic inside the thought experiment. The identification of the correct process rules is the basis of defining the correct algorithm for any discovery.
While working on the protein synthesis problem, two transformation phases were identified: transcription and translation. The transcription phase transforms the DNA mosaic into the RNA mosaic. The translation phase transforms an RNA mosaic into an amino acid mosaic.
Multiple scope sets were declared while analyzing many thought experiments. The observation stream defines scope sets to focus the run of visual streams while working on any problem and reducing the time needed to reach a solution. Physical rules were identified while working on the translation problem. The physical rules are part of the environmental rules that represent the laws of the real-world environment.
Also, a new type of visual stream, the communication stream, was studied and applied to all case studies within this research. An expression stream is a subset of the communication streams, and is responsible for generating the communication thought experiments.
Identified for the first time, these streams translate the internal thought experiments into a secondary language understood by the outer world. The PL applies the internal streams to develop the internal thought experiments. The different types of visual streams are explained in Section 2.4.0.
7


Several new types of data storage mosaics were identified. The first type is the information mosaic (one-dimensional and two-dimensional). Information mosaics are created via a sequence of transformation phases by the internal streams. Other new types of mosaics are the algorithmic and symbolic mosaics that include information to be passed to the outer world. The algorithmic and symbolic mosaics are constructed via the communication thought experiments by the expression stream.
1.8.0 Dissertation Structure The chapters of this dissertation are organized as follows:
• Chapter II: Review of the literature discussing the PL, thought experiments, the AD, including a brief description of all its streams and components, and LG and its components.
• Chapter III: Explanation of the communication streams and the algorithmic mosaics.
• Chapter IV: Review of the literature discussing the discoveries of the admissible trajectories, the genetic code, and the protein synthesis.
• Chapter V: The first case study is the application of the AD to the rediscovery of the algorithm for generating admissible trajectories in LG.
• Chapter VI: The second case study is the application of the AD to the rediscovery of the genetic code.
• Chapter VII: The third case study is the application of the AD to the rediscovery of the translation algorithm.
• Chapters V, VI, and VII contain the essential parts of this research. In every case study, the original descriptions of the discoveries are analyzed, the AD is
8


applied, and the components of the AD guiding those discoveries are presented.
• Chapter VIII: Contributions and Future Work. It includes contributions of this research in the fields of Artificial Intelligence and Computer Science.
9


CHAPTER II
THE PRIMARY LANGUAGE
2.1.0 Overview
This chapter includes a PL literature review. As mentioned in Section 1.1.0, von Neumann introduced the PL as the language the human brain uses for mental calculation. While investigating PL, two of its major algorithms essential for the development of humanity were revealed, LG and the AD. According to the hypothesis of this research, the PL is the Language of Visual Streams (LVS) that manifests itself via various means including invention of new algorithms.
2.2.0 The Primary Language (PL)
The two types of languages that the human brain uses were suggested by von Neumann in 1957, [47], He introduced the concept that the PL is used by the brain for computation and mental calculation. Von Neumann suggested that the PL was developed “beneath” the secondary languages, i.e., the external languages. When von Neumann declared the existence of the PL, he argued that the nature of the PL was unknown. He writes, “It is only proper to realize that [human] language is largely a historical accident. The basic human languages are traditionally transmitted to us in various forms, but their very multiplicity proves that there is nothing absolute and necessary about them. Just as languages like Greek or Sanskrit are historical facts and not absolute logical necessities, it is only reasonable to assume that logic and mathematics are similarly historical, accidental forms of expression. They may have essential variants, i.e., they may exist in other forms than the ones to which we are accustomed ... The language here involved may well correspond to a short code in the sense described earlier, rather than to a complete code: when we talk mathematics, we may be
10


discussing a secondary language, built on the primary language truly used by the central nervous system [47].” For more than 60 years, the nature of the PL was still unknown.
Following von Neumann's hypothesis, the investigation of the PL started, [27]-[6], with the assumption that the PL is the basis of all algorithms crucial for the development of human intelligence. According to Stilman [16], the PL appeared long before the time when the external languages were developed. During our investigation, it was assumed that the PL generates mental movies called visual streams that interact with each other. It was suggested that the PL is the Language of Visual Streams. Visual streams were introduced as imaginary animated movies rather than sets of strings of symbols (called languages in Mathematics and Computer Science). In particular, visual streams are generated by the PL to find solutions for many problems by focusing them in proper direction, and reasoning about them. The PL uses those visual streams to interact with each other. The PL is hypothesized to be the language utilized for the development of all algorithms discovered by humans; it is the foundation for all external, i.e., spoken languages.
2.3.0 Visual Streams and Thought Experiments
Thought experiments are the methods created by an investigator’s imagination to visualize several scenarios to analyze a problem, “solve it,” and “see” the results of different solutions. Thought experiments provide visual reasoning about the laws of nature in different fields. Hobbes and Locke introduced the state of nature using philosophical thought experiments, [52], Also, thought experiments can be used to examine other hypotheses without actually testing these hypotheses. For example, Galileo created a thought experiment to refute Aristotle’s theory about the free fall of bodies. Without dropping a single object, Galileo concluded that bodies with unequal mass fall at the same speed, [60], and [1],
11


The possibility of thought experiments shows that the human brain can mix a variety of animated models reflecting reality and its natural (or artificial) laws. By studying the brain's cortex, scientists have successfully detected some of those mental images. They found that human mind converts any shape or symbol into a mental image. Such images have an analogical nature because, according to the studies, the visual areas of the brain sketch the contours of such imagined objects, [60],
It appears that for any species to survive, it is very important for said species to understand the laws of nature. Humans can represent those laws in their brains by simulating them in the form of mental images. Neuropsychologists found that people carry with them mental universes that simulate the laws of the world around them, [60], Moreover, other studies proposed that humans also carry the laws of major human relations, including the laws of warfighting. Those laws manifest themselves in several ways; for example, the sensorimotor system of the human brain can understand kinematics when predicting an object’s trajectory. Humans use those laws routinely and subconsciously while visualizing an object in a mental movie or predicting its trajectory on a map, [1], and [61],
The visual streams may simulate different artificial worlds; each of which is ruled by artificial laws of nature and populated with animated entities, which represent real or simulated objects. The visual streams are constructed and run by the human brain to observe realistic or artificial events in those mentally constructed worlds. According to this research’s hypothesis, every human invention is made by employing visual streams, [27],
By observing the effects of the physical laws in the environment around it, the brain simulates them and constructs the animated mental events accordingly. The brain has the power to alter the mental worlds by redefining their components according to the problem
12


statement. In this research, a visual stream is assumed to be a mental movie that is played in the human brain to show, at its conclusion, a solution of a staged problem. Stilman proposed that it is possible to model and eventually implement such mental worlds and visual streams by employing conventional algorithms. Stilman further proposed that the PL is the LVS, [8], [17], [9], and [28],
Visual streams can trigger additional thought experiments and schedule them to be run by other visual streams. The objective, structure, and outcome of the created thought experiments should become known to the triggered visual stream. Moreover, the visual streams that schedule additional thought experiments (to be executed later) are usually are not aware of the algorithms to be executed in those experiments. The structure of these algorithms will be developed and, thus, will become known during the runs of the associated thought experiments, [23], and [10],
2.4.0 Types of Visual Streams
Visual streams can be broken into groups according to their application. Most common, there are visual streams that interact internally only with other visual streams. They have no interaction with the outer world. This type of visual streams is called the internal stream. A different type of visual stream that interact with the outer world is the communication stream, [27],
The visual streams can be classified by the information the streams process. Mundane streams deal with ordinary information while scientific streams deal with scientific information. Moreover, the visual streams used by the AD can be broken into three groups based on their function: observation, construction, and validation streams. Visual streams
13


may also be classified based on the type of reasoning they utilize. So far, two major types of reasoning have been revealed: proximity reasoning and mosaic reasoning.
Additionally, visual streams can be classified according to their programming structures (parallel, sequential, or nested). Parallel streams run concurrently, while sequential streams run consecutively. A typical example of sequential streams is how they are utilized by the AD: observation to construction to validation streams. The nested streams call each other in no specific order. Yet another type of classification is related to the stream’s common theme, which unifies a group of streams with a common set of constraints that limit possible directions of the streams’ morphing. Table 1 represents the different types of the identified internal streams.
Table 1. Different types of internal streams.
Classification Criteria
The AD Streams Information Reasoning Programming Structure Common Theme
Observation, construction, and validation Mundane and scientific Proximity and mosaic Parallel, sequential, and nested Any type of internal streams has similar set of constrains
2.4.1 Discovery Streams and the Algorithm of Discovery (AD)
In this research, it is assumed that the AD is the algorithm for inventing new algorithms; in other words, it is the algorithm that makes discoveries. Based on multiple thought experiments involved in the discoveries in LG and Molecular Biology/Biochemistry, in [16] through [59], it is argued that the AD does not utilize a tree-based search to solve a problem. Instead, it is theorized that the PL generates visual streams to build a mental movie, that is played in the human brain, and which concludes with a solution to the staged problem. The
14


PL will use those visual streams as inputs to the AD. An input visual stream is usually built from multiple imaginary instances of the object under investigation. The first visual streams the AD receives are usually a visual replication of natural or imaginary entities. The AD then runs this visual stream, several times, to understand these objects, find their structure, and the rules of their construction, [16],
Stilman categorized visual streams based on their nature or the work for which they are responsible, [27], He proposed that streams could be divided into groups based on their purpose. He identified three classes of discovery visual streams the AD can execute: observation, construction and validation, [12], These discovery streams can be sequential, parallel or nested. At the beginning of every AD run, the AD first initiates the observation stream. In this research, it is proposed that the observation stream has two stages, the preliminary stage, and the analysis stage. The main job of the preliminary observation stream is to collect the required inputs from a stream of inputs including all known data about the problem. The analysis stage of the observation stream requires it to investigate and build at least two tools: the construction set and the visual model, usually by “erasing the particulars”, [24],
Erasing the particulars means to remove any unnecessary specifics of the objects inside the visual stream by executing the visual stream several times in attempt to reduce multiple instances of the object to the one or two simple instances, considering the outputs from previous thought experiments as well as incoming inputs from the impression stream. By doing so, the observation stream modifies objects on hand to “observe” their real structure and “find” the rules that control it. Usually, in order to discover the structure of an object or a process, the observation stream has to construct it. Once the observation stream reveals the
15


rules that direct the construction of an object, it builds a visual prototype of this object and saves the rules in the construction set. The observation stream then triggers another visual stream, the construction stream, to start building the object, or staging the process, and, eventually, reveals the algorithm that guided the construction.
The validation stream tests the correctness of the constructed object, or staged process, and compares it to the visual model identified by the observation stream. The validation stream also examines the rules for building the object, or the staging process, by comparing them to any new data coming from the outer world. If a contradiction is found, the validation stream will reject the results of that construction and call the observation stream to rebuild the construction set or modify the visual model. In such cases, the construction stream must be called again to reconstruct the object or restage the process. These multiple executions are needed for the AD to correctly solve the problem, [16],
In many cases, the AD may utilize a virtual moving entity, the Ghost, to move in the artificial world to investigate other objects in this artificial world. The observation stream employs the Ghost to “walk” and “observe” the structure of an object or to find the rules that control objects in the artificial world. The construction stream can also use the Ghost to build an object by assembling its parts and placing these parts in their correct positions. The validation stream triggers the Ghost to “scan” through the artificial world and “test” the rules and objects identified by previous visual streams in order to “find” any contradictions that might invalidate these rules or objects, [16],
2.5.0 Proximity and Mosaic reasoning The AD sometimes needs to apply special types of reasoning to find solutions for certain categories of problems. Some problems require optimization components. In this case, the
16


AD tries to find the solution by running a sequence of thought experiments that run via proximity reasoning, [23], In many cases, to find a solution, the AD has to run visual streams that have the ability to assemble objects of different shapes and types, i.e., run under the influence of mosaic reasoning, [16],
In [10], Stilman introduced mosaic reasoning to describe the reasoning that helps visual streams build a mosaic picture of any object out of small colorful tiles. The major components of mosaic reasoning, revealed to date, are tiles, aggregates, and matching rules.
A tile is the smallest unit of the mosaic. In order to construct any object correctly, tiles must be accurately inserted in the mosaic. An incorrect insertion of any tile will ruin the final structure of the mosaic leading to an incorrect solution of the problem. Two or more tiles linked together form one aggregate.
In general, mosaic reasoning runs via multiple visual streams using tiles and aggregates. The observation stream starts mosaic reasoning by identifying the tiles and aggregates that will be used to build the final mosaic. This identification process might take several executions and morphing of the visual streams. Once the tiles and aggregates are identified, the observation stream must investigate the matching rules.
The matching rules determine how tiles and aggregates link together. For the mosaic of processes, those rules control how different objects move within the artificial world. These rules can affect the construction of any mosaic locally or globally. The global matching rules are the global complementarity rule, the environmental rules, and the transformation rules, while the local matching rules are the local complementarity rule and the interchangeability rule.
17


Global complementarity rules control the insertion of one aggregate into the mosaic according to the placement of its adjacent aggregate. Further, visual streams can use global transformation rules to construct new mosaics from one or more existing mosaics. Environmental rules represent all the rules of the real environment. The environmental rules must be obeyed while building any mosaic. These rules include physical constraints, chemical rules, etc.
Transformation rules include the movement rules that guide the streams to move one or more mosaics. As a result of this movement, a new mosaic is generated. The process rules are also part of the transformation rules. They identify the order of the mosaic’s actions within the thought experiment. For example, if there exists three different mosaics that must interact together to build another mosaic, the process rules control which mosaic moves first, while the movement rules guide the motion of a mosaic.
On the other hand, local complementarity rules control the placement of a tile based on the adjacent tiles. Interchangeability rules allow visual streams to remove one aggregate from the mosaic and replace it with an interchangeable one that has the same general structure. Interchangeable aggregates that share the same features are called plug-ins. This interchangeability changes the final picture of the mosaic, but the correctness of the final structure of the mosaic will not be affected.
18


CHAPTER III
COMMUNICATION STREAMS AND ALGORITHMIC MOSAICS
3.1.0 Overview
While studying the PL, several types of visual streams were introduced. As mentioned in Chapter II, it is assumed that the PL is the engine that generates visual streams to visualize and investigate any problem. During this research, visual streams were divided into two major groups based on their interaction with the outer world. The internal streams are the streams interact internally, i.e., only with each other. An example of an internal stream is a discovery stream, which could be an observation, construction, or a validation stream. On the other hand, the communication streams are the streams that operate between the internal streams and the outer world. Communication streams transform the information from the world around us to visual inputs to be processed by the internal streams and vice versa. The subset of the communication streams that converts the outer world information into the internal streams is the impression streams. The subset of the communication streams that converts information of the internal streams to secondary languages, i.e. human languages, is called the expression streams. This chapter includes a detailed description of the communication streams and the associated results of this research.
3.2.0 Motivation
After several investigations of the PL, a new class of visual streams was discovered, the communication streams, [27], These streams were never applied to any of the previous work on the PL. Thus, it was important to investigate the nature of the commination streams to reveal additional structures of the PL, if any.
19


3.3.0
Communication Streams
The main purpose of the communication streams is to exchange dynamic pictorial information between the internal streams of one entity such as a human or a robot and the outer world, i.e., other entities (humans or robots), and the nature. These streams utilize a symbolic language that other entities understand. In addition to symbols represented as letters, digits, scientific symbols, sounds, etc., this language could use drawings, which are snapshots of the internal streams. Another group of communication streams passes symbolic and pictorial information back to the internal streams.
Communication streams can be divided into groups based on their functionality. One group is the set expression streams. The main function of these is to convert information from internal streams into a form acceptable for passing to the outer world. Extensive case studies in [16]-[6] revealed two types of expression streams, pictorial and symbolic, [25],
The pictorial expression stream translates an internal visual stream into a sequence of snapshots of this internal stream. A typical example of the pictorial stream is the set of illustrations that scientists draw to explain their ideas. On the other hand, the symbolic expression stream translates the internal visual streams by “watching” the morphing of this stream. This “watching” also includes “tagging” or naming the important objects, actions, and events to create the symbolic shell. Then, employing a set of rules, i.e., a grammar [27], the symbolic expression stream converts this symbolic shell into a string of symbols in a speech, an algorithm, a scientific theory, etc. In such a case, the tags are utilized as a set of terminal symbols for the grammar. The symbolic expression stream also has the ability to add captions to the illustrations generated by the pictorial visual streams. In particular, the
20


output of the expression stream could serve as a solution for some problem to be presented to the scientific community.
The other subset of the communication streams is the set of impression streams. While the expression stream links the internal streams with the symbolic languages, the impression stream works in the opposite direction. Impression stream converts the external languages into visual objects that interact based on the description provided by the symbolic language. The impression stream converts the symbolic strings of an external language into the animated visual objects.
All visual streams operate in the form of thought experiments, [16], In that sense, the thought experiments generated by the communication visual streams provide a link between the PL and the outer world. The internal visual streams operate employing internal thought experiments, while the communication streams, both expression and impression streams, utilize communication thought experiments. These thought experiments work in mutually opposing directions, as shown in Figure 1.
Internal Thought Experiments
1 Established Science Discovery Science

Internal Streams
Communication Thought Experiments
£§ i
LM
1U2X6)
5678Science MathOO J»XLO0 4t €¥ E12340t2

Figure 1. The communication streams pass information between the internal streams and the
outer world and vice versa.
21


According to the hypothesis in Section 1.1.0, the expression of an idea is the transformation of the internal visual streams into symbolic language via the expression stream. As already mentioned, a symbolic language can be verbal, written, code-based, e.g., Morse code, etc. It can also be a technical or natural language. Technical languages usually include specific symbols and rules developed within a given branch of science.
In some cases, it is very difficult to develop proper communication streams, especially those related to science. However, when those streams have been developed, such as those for natural languages, they could run in real time or even faster. This is what happens in the human brain. Indeed, humans usually do not have gaps between their thoughts and the expression of them (unless they are not familiar with the language they are speaking). In the latter case, humans create the following chain of streams. First, they use an expression stream based on their native language. This stream is then converted into an internal stream of another entity, using its impression stream. This new internal stream is translated into another expression stream using the language familiar to another entity (and non-familiar to the first one). In other words, one expression stream initiates another impression stream to create a new “different language” internal stream to be linked to another expression stream, thus creating a chain of internal and communication streams.
3.4.0 Algorithmic and Symbolic Mosaics While working on the different problems in Chapters 5, 6, and 7, the expression stream translated the discovery streams into an algorithmic or a symbolic mosaic. The expression stream creates algorithmic mosaics by attaching several instruction aggregates. Every such aggregate represents one of more actions made by other visual streams. The instruction aggregate consists of two information aggregates, the operation aggregate and the parameters
22


aggregate. For example, the instruction aggregate “Add (number 1, number2)” tagged the action of the construction stream when it adds two numbers together. During this research we discovered several instruction aggregates introduced by the expression stream. Each aggregate represented an operation executed by the construction stream while building a part of a mosaic. Table 2 shows a general form of an instruction aggregate utilized by the expression stream, as revealed in Chapters V, VI, and VII. The set of various instruction aggregates is expected to increase based on the subsequent PL investigations.
Table 2. Several instruction aggregates introduced during this research.
Instruction Aggregate Meaning
Overlay (first mosaic, second mosaic) Overlay two mosaics
Add (number 1, number2) Add two numbers
Read (information mosaic, reading rules) Read information in one mosaic based on defined reading rules
If condition then... Check a condition during construction
Repeat (number of cycles, repeating from) Repeat several instruction aggregates. This operation aggregate takes two parameter aggregates: 1. Number of cycles: how many times this operation should be repeated. 2. Repeating from: the starting place to repeat
Choose (mosaic/ tile/ aggregate, set) Take a tile, aggregate, mosaic from a set of tiles, aggregates, mosaics, respectively
Copy (mosaic 1, mosaic 2) Take an identical copy of one mosaic and save it into another
Align (first mosaic, second mosaic, alignment rules) Align two mosaics according to a defined alignment rules
Attach (first mosaic, second mosaic, Complementarity rule) Link two mosaics together based on the complementarity rules
Scan (mosaic, movements rules) Send the Ghost to walk through a whole mosaic based on the movement rules
Remove (aggregates, tiles, or mosaics) Delete tiles, aggregates, or mosaics from the visual stream
23


Table 2 cont’d
Instruction Aggregate Meaning
Detach (mosaic, aggregate) Detach an aggregate from a mosaic
Generate (aggregate, construction rules, tiles set) Build an aggregate out of a set of tiles based on the construction rules
Call algorithmic shell Call any part of the algorithmic mosaic
Move (mosaic, current location, destination location, movement rules) Move one mosaic from the start location to the end location according to the movement rules
Erase (particulars, mosaic/aggregate, general shape) Erase the specifics from one mosaic and convert it to a general shape.
The expression stream builds the symbolic mosaics to save the information declared by an internal mosaic. During this research, many symbolic mosaics were generated by the expression stream, in the form of tables, to save the different matching rules announced by the discovery streams. In Chapter VI, the table that maps between the DNA base and letter tiles was constructed by the expression stream, shown in Table 3. As the AD was applied to the protein synthesis problem more symbolic mosaics were identified, Chapters V, VI, and VII.
3.5.0 Results
The algorithmic mosaic is yet another type of information mosaic built to communicate written information. However, algorithmic mosaics are constructed by the expression stream using a secondary language. These mosaics are a direct translation of the internal streams in order to be understandable by the outer world. They can also be read by the impression stream and translated into a visual input to be passed to the internal streams.
In the next chapters, we consider different algorithmic mosaics generated by the expression streams. Specifically, at the end of every discovery, a complete algorithmic mosaic representing a solution to the investigated problem is generated. The work on
24


communication streams and algorithmic mosaics will be continued in the future based on the results obtained in this research.
25


CHAPTER IV
INTRODUCTION TO THE INVESTIGATED DISCOVERIES
4.1.0 Introduction to Linguistic Geometry
Linguistic Geometry, LG, is another algorithm based directly on the PL, i.e., on visual streams, [28], and [26], Currently, it is understood that the development of LG was not an invention, but the discovery, or rediscovery, of an ancient algorithm developed by the human brain in an attempt to optimize warfighting, [29], and [30], LG was rediscovered, and further developed over the last 40 years, as an approach to solving opposing games, such as chess, in real-time. This research, started in 1972 as a series of experiments for the analysis and modeling of the chess experience, to include previously developed strategies of advanced players or masters of the game. By generalizing this experience, a computer program, PIONEER, was developed, [28], PIONEER solved a number of well-known complex chess positions and endgames with only 100 variations included on the search tree.
Studies following PIONEER’S approach led to the general LG theory, a new type of game theory for solving Abstract Board Games (ABG). The mathematical foundation of LG is the Hierarchy of Formal Languages. One of the major results in LG permits the solving of classes of games without any kind of search and simultaneously proving the optimality of the proposed solutions. The analysis of this discovery played an important role in the development of the AD, [8], [17], [12], and [13]-[18],
4.1.1 Trajectories
Before LG generates optimal gaming strategies, the problem must be defined as an ABG. This requires the definition of all the ABG terms, e.g., the definition of the abstract board, the players, the pieces, and the game rules.
26


Assume that an ABG is defined including its reachabilities, which describe all the locations reachable from an arbitrary location in one step, for an arbitrary piece p, [28], Trajectories for piece p from location x, to xj of length /, is the set of strings of symbols that represent all possible paths for piece p between x; and xj of the length /. Any trajectory of length / is the shortest trajectory if / is the minimum possible length among all the trajectories of piece p between x, and xj. It is possible to have more than one shortest trajectory between two locations. Admissible trajectories of degree k are those trajectories that can be divided into k shortest trajectories. Hence, any shortest trajectory is an admissible trajectory of degree 1. In LG, the set of all the trajectories of the length less than H (H is an integer and H > 1) for the current state of an ABG is formally represented as the Language of Trajectories LtH, [11], and [28], Figure 2 shows two different trajectories that piece p can take between x, and xj, one - of length 4 and the other - of length 6. Following a series of previous works by Stilman and Aldossary to apply the AD on different discoveries in the field of LG, [8]-[4], the AD was applied to rediscover the admissible trajectories algorithm and revealed all the AD components led to this discovery, which is shown in Chapter V.
Figure 2. Two different trajectories between points x; and xj.
27


4.1.2
Introduction to the Central Dogma
DNA is a repository of genetic information for all living organisms, and since the discovery of its 3D structure scientists have been puzzled by the genetic code, [3], One of the areas that puzzled scientists is the generation of proteins; specifically, understanding how the genetic information is utilized for generating proteins. In 1958, Francis Crick hypothesized that the genetic information travels from DNA to another repository, RNA, and then it is translated into proteins, [35], Today, this process is considered the Central Dogma of Molecular Biology.
As the first step of applying the AD to rediscovering the solution of the two major problems in the Central Dogma, i.e., the genetic code and protein synthesis, the AD broke them into a sequence of thought experiments (including successful and failed experiments). This sequence did not appear all at once, i.e., every experiment generated the next one until a solution was found. The information about the experiments that historically took place, was obtained from the following papers [3]-[54], In order to develop this sequence of thought experiments, the inputs and outputs of the original experiments must be considered, in order to include the information available at the time of the original discovery. In terms of input, the actual components of DNA (the DNA bases) were known, but the meaning of the sequence of those bases (the genetic code) was not completely understood. For the output, some fragments of protein sequences were known, such as the amino sequence of insulin. The total number of amino acids, which are the fundamental components of proteins in all living organisms, was known as well. However, some of the amino acids that built the protein mosaic were still unknown at that time, [3],
As mentioned in Chapter II, the AD was expected to construct the sequence of thought experiments to reveal the algorithm for translation of genetic information into the structure of proteins. To construct this sequence, it had to utilize information about the components of the
28


algorithm of translation known at the time of the original discovery. For example, the mechanism for DNA replication was hypothesized as follows: unzip the double helix and then duplicate each strand independently. Each strand should form a new copy of the DNA. This algorithm could also be applied for creating the RNA mosaic from a DNA template, [50],
On the other hand, the complete algorithm of translating DNA into protein was unknown.
Even the most basic questions were yet to be answered. The correct method of reading the DNA sequence was still a mystery. It was not clear whether to read the genetic code on the two strands of the DNA simultaneously or only read one strand at a time. If the genetic code was read only from one DNA strand, then the dilemma was which one should be read.
4.2.0 DNA and RNA
In order to find the essential algorithms that guide protein synthesis, the AD had to define the visual model required as the initial input. The AD reasoned that since the genetic information resides in the DNA, the initial input must be a complete 3D structure of the DNA, the double helix. Visual streams, especially the observation stream, had to morph this input to create the informational representation of the DNA necessary to start transcription, the first step in protein synthesis that creates the RNA intermediate mosaic. In order to do so, the AD retrieved the output resulting from the experiments that revealed the structure of DNA, [10], Such information stated that the DNA mosaic consists of several types of tiles. They include four different nitrogenous bases, Adenine (A), Guanine (G), Thymine (T), and Cytosine (C). One of those nitrogenous bases is included in every nucleotide aggregate, which also includes a sugar tile and a phosphate tile. Each strand of the double helix is a chain of nucleotide aggregates. The strands are linked together by hydrogen bonds via nitrogenous bases (inside the structure of the DNA), whereas the outside structure of the double helix is formed by two sugar-phosphate backbones.
29


Following the complementarity matching rule, the nitrogenous bases that link the DNA strands are paired together as follows, A with T and C with G.
At the beginning of investigating protein synthesis, the existence of RNA was known, but its role in the process was not clear. The nucleotide bases of RNA matched well those in the DNA; A, G and C were exactly the same, while the base of U can be considered as a replacement for T. Inside the cells, special molecules copy one strand of the DNA to create the RNA molecule. While copying, every T nucleotide is transformed into a U nucleotide. During the discovery of protein synthesis, different kinds of RNA were identified. They include mRNA and tRNA. The mRNA is the RNA that carries the genetic code are discussed in Chapters VI, and VII. The tRNA is the RNA are responsible for carrying different amino acids and moving them to be assembled next to each other to form the peptide mosaic, [50],
4.3.0 The Genetic Code
This section discusses the genetic information that was revealed during the application of the AD to the genetic code problem. At the end of revisiting genetic code discovery, the AD concluded that the genetic alphabet consists of the four letter tiles. Every three letters would be connected to each other to form one code (codon) aggregate that signals synthesis of the one amino aggregate. The genetic alphabet contains 64 different code aggregates, 61 of which represent the existing amino acids and three are the “Stop” aggregates that signal the termination of the current genetic reading, [50],
After discovering the DNA structure, the AD triggered thought experiments to decipher the genetic code as shown Chapter VI. A series of thought experiments was initiated to redefine this problem, i.e., to switch visual streams from one theme to another, see Sections 6.3.0, 6.4.0, 6.5.0, and 6.6.0. These experiments resulted in the conversion of a 3D chemical theme into an
30


informational one, leading to the AD's first attempt to crack the genetic code. When this attempt failed, the AD defined and redefined the inputs and assumptions in order to decode each genetic code symbol, one by one, until the meaning of all of the genetic symbols was finally discovered, [42], and [55],
4.4.0 Protein Synthesis
The AD preliminary analysis showed that the algorithm of protein synthesis consists of two main phases: transcription and translation. Transcription is the first stage needed to turn the DNA mosaic into an RNA mosaic. This algorithm converts the four-letter tiles in the DNA mosaic, A, T, C and G into A, U, C and G, respectively, to form the letter tiles in the RNA mosaic. At the beginning of transcription, RNA is generated by a special RNA polymerase that binds to a specific region of DNA, which contains the genetic information. First, the polymerase unzips the two DNA strands (from each other) to start the transcription. This polymerase then reads the base tiles in the DNA and builds the RNA mosaic accordingly. It will continue to read through the DNA strand until it reaches a termination region that stops the transcription. Once transcription is terminated, the polymerase releases the DNA mosaic. The result of this process is an RNA mosaic that carries the genetic information required to form one protein chain, [50],
During translation, the genetic code on the generated mRNA resulting from transcription is read to construct the protein peptide mosaic. Many molecules work together to read the mRNA strand in order to transform it into a protein mosaic. In this research, the AD was applied to rediscover the translation algorithm. A number of thought experiments essential to rediscovering the process of activating the tRNA mosaic were defined, as shown in Section 7.3.0. Then, the AD was applied to rediscover different models of the translation process, given in Sections 7.5.0, 7.6.0, 7.7.0, and 7.8.0.
31


By now it is known that during the translation, ribosome molecules are constructed and linked to the mRNA mosaic. Once a ribosome is attached to an mRNA mosaic, the ribosome starts reading the genetic code of the mRNA in triples, i.e., three mRNA tiles at a time. Based on the mRNA reading, the ribosome will call for the corresponding tRNA mosaic, yet another type of the RNA molecule. Every tRNA mosaic is attached to a specific building block, i.e., the specific amino aggregate. The ribosome will continue reading the genetic code on the mRNA mosaic and calling for the correct tRNA mosaics. Every time a new tRNA mosaic enters the ribosome, the ribosome will link its attached amino aggregate to the growing peptide mosaic. The ribosome will stop reading when a “Stop” code aggregate is reached. Chapter VI describes the beginning of the research to apply the AD to redefine the rules that control the translation algorithm. At the end of Chapter VI the rules that control the reading of the mRNA mosaic were rediscovered. Chapter VII continues the investigation to rediscover the entire mechanism of translation.
32


CHAPTER V
REVISITING THE ALGORITHM OF ADMISSIBLE TRAJECTORIES
5.1.0 Overview
In this chapter, the work of applying the AD to rediscovering the algorithm for generating admissible trajectories is described. For this purpose, all the thought experiments that led to this discovery were replayed by the AD, [5], The role of mosaic reasoning was emphasized, for the first time, in this series of thought experiments. Specifically, all the components of mosaic reasoning that participated in constructing the algorithm for generating admissible trajectories of degree 2 between arbitrary locations on an abstract board were found. Also, the communication thought experiments were identified. These thought experiments are run by the expression stream to translate the internal streams into a symbolic language. The final result of the expression stream is an algorithmic mosaic that describes the algorithm for discovering admissible trajectories, [5], and [28],
In LG, admissible trajectories of degree 2 are all the trajectories that are constructed from two shortest trajectories. The AD assumed that the algorithm for generating the shortest trajectories has already been discovered, [8], Via the impression stream, the AD received the following inputs: the start and end points, the length of the desired trajectory (1) the type of reachability 1, as well as the output from the algorithm for generating the shortest trajectories, if any exist. The reachability relation governs the movement from one point to another. The output of the algorithm for generating admissible trajectories is a tree mosaic that represents all possible admissible trajectories from start to end, [11],
1 The rules that control the movement from the start to the end points.
33


5.2.0
Motivation
Rediscovering the algorithm to generate the admissible trajectories was the next step, after applying the AD to rediscover the shortest trajectories algorithm, [4], The goal was to emphasize the components of the AD that were found while rediscovering the shortest trajectories and to find new components, if possible. Also, it was important to study the applicability of mosaic reasoning in this discovery. Further, a simple straightforward discovery such as admissible trajectories was needed to break down all the communication thought experiments that led to the identification of the admissible trajectories algorithmic mosaic.
5.3.0 FindMid Experiment
5.3.1 The Observation Stream
In its preliminary stage, the observation stream received the definition that any admissible trajectory of degree 2 consists of a pair of shortest trajectories. From this definition, the observation stream concluded that there must exist a mid-point between these shortest trajectories. It focused itself to save the algorithm for generating the shortest trajectories. It also saved the incoming admissible trajectory and its length, /. The observation stream, by analyzing the current inputs, found that all components of the admissible trajectory were already known. The only missing piece of information was the set of all mid-points. The observation stream tasked itself to investigate the construction of this set.
To schedule the required experiments, the observation stream modeled this problem as a backyard to be paved with tiles and introduced an appropriate construction set. In this case, the bundle of admissible trajectories to be constructed represents a complete mosaic of tiles,
34


which is similar to the process of paving a backyard. The stream also assumed that adjacent tiles along the trajectories would be linked by the reachability relation. It then introduced the set of trajectory tiles (T-tiles). The observation stream declared special subsets of the T-tile set: the mid-point tiles (Mid-tiles), the start tile (Start-tile) and the end tile (End-tile). The stream reasoned that for every admissible trajectory of degree 2, a mid-point is always the end-point of one shortest trajectory and the start-point of another one.
Figure 3 shows the initial visual model created by the observation stream. At the beginning, the start-point (blue square) and end-point (orange square) were marked on the ABG board. The observation stream waited for the length of the admissible trajectories, /. Once received, the forward map (bottom-left of a given block) and the backward map (top-right of a given block) were retrieved. These maps include all the forward and backward distances for each square in Figure 3. Forward distances are those distances from the starting to the end points. Backward distances are the reverse.
The observation stream scheduled an experiment, FindMid, to study the nature of the mid-points and, possibly, to find and store them as a set of all possible mid-points (DOCK). The input of FindMid was a visual model that includes the start and end points and the construction set that contained: the length of the trajectory (l), the type of reachability relation, and the algorithm for generating the shortest trajectories.
35


Figure 3. Visual representation of the initial input model.
5.3.2 The Construction Stream and the Expression Stream
The construction stream started the FindMid experiment after receiving the visual model and the construction set. As this stream searched for the mid-points, it reasoned that every mid-point is an end-point of a given shortest trajectory. Therefore, the length from the start point to the midpoint must be equal to the length of the shortest trajectory. Similarly, this mid-point would be the start-point of another shortest trajectory and its distance from the end-point had to be equal to the length of the second trajectory. Consequently, the total length / of the admissible trajectory, which is equal to the sum of the lengths of the first and second components, must be equal to the sum of distances described above. The observation stream reasoned that it could find those mid-points as the locations where the sum of forward and backward distances is equal to /. The reading rules needed to read the forward and backward distances, as well as the algorithm for calculating the sum, comes to the


construction stream through its input via the impression stream, as a part of the algorithm for generating the shortest trajectories.
The construction stream retrieved the forward and backward distance maps created originally for generating the set of shortest trajectories. The construction stream overlaid the two maps to create the sum map. It concluded that all the mid-points must be the points where the sum equals to /. The construction stream saved this conclusion as the reading rule to mark the mid-points. These points would be marked as members of DOCK. Figure 4 shows the visual model, after morphing by the construction stream. This model included all Mid-tiles (green squares) that belong to the set DOCK after "erasing the particulars", i.e., eliminating unnecessary locations on the ABG board.
Figure 4. The visual model after being morphed by the construction stream.
The expression stream monitored the actions of the construction stream and tagged the events needed to find the mid-point. The stream generated a “Find Mid” shell that included all steps, or instruction aggregates, needed to build the DOCK set. Such a shell, once finalized, could run independently from the internal streams to identify all the mid-points for
37


any incoming admissible trajectories. At the beginning, the expression stream saved the instruction aggregate “Call generate-forward-distance().” The expression stream then added to the current shell “Call generate-backward-distance().” The algorithms for building the forward distance and backward distance maps were inputs from the “shortest trajectories experiments.” The expression stream saved the action of overlaying the two maps. Based on that action, the expression stream added a new instruction aggregate “Overlay (forward distance, backward distance)” The last instruction aggregates added by the expression stream are “Add (forward distance, backward distance)”, “Read (current-sum, Null)”, and “If current-sum := / then Save current location in DOCK.” The expression stream saved the “Find Mid” shell to be added later to the final algorithmic mosaic corresponding to the construction of the admissible trajectories.
After defining the DOCK set, the visual streams, both observation and construction, focused on the investigation needed to solve the problem into an examination of the members of the DOCK set only. The DOCK set is a special type of scope set that the AD sometimes identifies to help focus the streams and minimize the time required to solve a problem.
5.4.0 Define the Mosaic's Aggregates and Matching Rules The observations stream then moved to the definition of the mosaic’s aggregates and used the previous conclusion that for every admissible trajectory of degree 2, a mid-point is always the end-point of the first trajectory and the start-point of the second one. The stream identified the From-Mid and To-Mid aggregates. The To-Mid aggregate leads from the Start-tile to the Mid-tile and the From-Mid aggregate leads from the Mid-tile to the End-tile of an admissible trajectory. Each of these additional aggregates represents a bundle of shortest trajectories that either end at or start from the chosen mid-point. However, they are
38


represented not just as lists of trajectories, but also as tree data structures, taking into account that many trajectories in a bundle have a substantial number of parts in common. The observation stream also recognized a third aggregate, the Mid-aggregate. The two Mid-tiles are used to build the Mid-aggregate, the major component of the bundle of complete admissible trajectories associated with a specific mid-point. This bundle is also represented as a tree of admissible trajectories.
To construct the From-Mid and To-Mid aggregates out of a set of tiles, the observation stream identified the global transformation rule. This rule guided the construction stream while building the To-Mid and From-Mid aggregates. These aggregates turned the pair, Start-tile and Mid-tile, into the To-Mid aggregate, and the pair, Mid-tile and End-tile, into the From-Mid aggregate. This global transformation rule is the algorithm for generating the shortest trajectories and is included as an input to the AD via the impression stream. Another transformation rule turned each triple, To-Mid, Mid-tile and From-Mid, into the Midaggregate of admissible trajectories. The algorithm which accomplishes this has yet to be discovered, see Section 5.5.0.
The observation stream then specified the global complementarity rule, which controlled the proper construction of the Mid-aggregate out of the To-Mid and From-Mid aggregates. The stream found that the To-Mid and From-Mid aggregates could be merged, or "glued", together at a specific tile, the Mid-tile.
Next, the observation stream started to assign interchangeability rules. Global interchangeability rules allowed the construction stream to take one Mid-aggregate of the admissible trajectory tree mosaic and replace it with another one from the set DOCK. This new Mid-aggregate had a different mid-point, which was a Mid-tile. This would certainly
39


change the picture, but the whole structure would stand. The observation stream defined the Mid-aggregates to be the global plug-ins for the admissible tree mosaic. Plug-ins are the interchangeable aggregates in any mosaic that can be replaced by another without affecting the final structure of the mosaic.
The local interchangeability rule permitted the construction stream to take any Mid-tile and interchange it with any other Mid-tile, which may lead to the construction of a different Mid-aggregate. After passing through all the Mid-tiles of DOCK, the stream generated a complete mosaic composed of the bundle of admissible trajectories.
5.5.0 GlueTree Experiment
5.5.1 The Observation Stream
The observation stream concluded that after defining all mid-points, the whole bundle of admissible trajectories could be built. The observation stream scheduled a GlueTree experiment. The inputs to this experiment include DOCK, the start and end points, and the global transformation rule, i.e., the algorithm for generating the shortest trajectories.
5.5.2 The Construction Stream
The construction stream started the GlueTree experiment to construct the final admissible tree. The construction stream chose the first mid-point in DOCK, however, it is not important which point is chosen due to the global interchangeability rule. When the stream picked a mid-point, two trees representing the two bundles of shortest trajectories were created using the appropriate algorithm, each tree representing either the To-Mid or From-Mid aggregate. The construction stream then began building the Mid-aggregate by taking the To-Mid tree and gluing every leaf on this tree to the root of the From-Mid tree, see Figure 5.
40


From-Mid
Figure 5. Visual simulation of the GlueTree thought experiment.
While monitoring the work of the construction stream, the expression stream generated the algorithm associated with building the final tree. The expression stream initiated a shell named “Construct one branch.” The first instruction aggregate added to this shell was “Choose (Next mid-point, DOCK).” As explained above, the instruction aggregates needed to build the DOCK set are defined in the “Find Mid” shell. When the mid-point is chosen, the construction stream called the algorithm for generating the shortest trajectories to build two bundles of shortest trajectories (one from the start-point to the current mid-point and the other from the current mid-point to the end-point). The expression stream saved into the “Construct one branch” shell the following instruction aggregates: “To-Mid := Call generate-
41


shortest-trajectories(start, mid-point)”, “From-Mid := Call generate-shortest-trajectories(mid-point, end).”
The gluing point is the Mid-tile. The construction stream declared a new alignment rule, where the Mid-tiles of the two tree mosaics must be aligned. Then, the stream glued the aggregates together by making copies of the From-Mid aggregate and overlaying the root of each copy of the From-Mid tree to the leaves of the To-Mid tree. The result of this gluing is a Mid-aggregate, which represented the bundle of admissible trajectories passing through the chosen mid-point, shown in Figure 6.
Figure 6. Building a Mid-aggregate from the To-Mid and From-Mid aggregates.
Consequently, the expression stream added more instruction aggregates to the “Construct one branch” shell based on the new actions of the construction stream. The expression stream first tagged the action of multiplying the From-Mid aggregate, “Copy (From-Mid, new-From-Mid)” The new-From-Mid was glued to the current leaf of the To-Mid in such a way that the two mid-points were merged. Therefore, the expression stream reasoned that the two trees must be overlaid so that the two mid-points are aligned. The expression stream generated new instruction aggregates “Align (new-From-Mid, To-Mid, alignment rules)” and “Attach (new-From-Mid, To-Mid, complementarity rules).”
42


The construction stream then chose another mid-point from DOCK and glued the second To-Mid tree to the second From-Mid tree to form the next Mid-aggregate. The stream repeated these steps for all mid-points. Once the construction stream had visited all of the mid-points of DOCK, a complete set of Mid-aggregates was generated.
The expression stream reasoned that every time a new mid-point is chosen from DOCK, the “Construct one branch” shell would run again. Therefore, the expression stream added a loop that repeats the shell until all the mid-points in DOCK are visited.
Finally, the construction stream reasoned that the construction of the final admissible tree from the Mid-aggregates requires the merging of those aggregates. This is necessary due to the fact that admissible trajectories stored in different Mid-aggregates may have common parts beginning from the start-point. These parts may include the whole trajectory. Merging the aggregates would eliminate such duplication. This would terminate the GlueTree experiment. The construction stream reasoned that the scanning would be done using the Ghost that would move following the movement rule2, which starts to move from the tree root mosaic and pass through all branches of the tree and mark any redundant branch. The expression stream saved the final instruction aggregates, “Scan (admissible-tree, movement rules)” and “Remove (redundant branches)” Figure 7 shows the visual representation of the final tree mosaic of the admissible trajectories.
Finally, the expression stream combined all the shells defined while monitoring the work of the discovery streams. Combining these shells led to the construction of the final algorithmic mosaic for generating the admissible trajectories, see Figure 8.
2 Movement rules are part of the transformation rules that guide the movement of one mosaic inside visual streams.
43


Figure 7. The visual representation of the mosaic of the tree of the admissible trajectories.
Figure 8. The final algorithmic mosaic corresponded to the construction of the admissible
trajectories tree.
44


5.6.0
Results
According to analysis of the construction of the algorithm for admissible trajectories, mosaic reasoning guided the AD to this discovery. Specifically, the AD applied mosaic reasoning to build the admissible tree mosaic. Defining tiles, aggregates, and matching rules helped the AD to focus the construction. A new set of matching rules appeared in this discovery, the reading and the alignment rules. The readings rules were applied by the AD to read the forward and backward distances, find the sum, and to compare it with the length of the admissible trajectory. The alignment rules were needed to align the From-Mid aggregate.
Via "erasing the particulars" approach, the AD found the scope set of the mid-points.
Once the mid-points were constructed, visual streams started assembling the mosaic components and building the final admissible tree mosaic. The algorithm for generating the shortest trajectories was utilized by the AD as a standard procedure in the construction of the output mosaic. The algorithm for shortest trajectories was used several times to generate bundles of those trajectories from start- to mid-point and from mid- to end-point. By having the tiles, Start, Mid and End, the Matching rules, and the bundle of shortest trajectories, the AD was able to generate a complete admissible trajectory tree mosaic and, this way, construct a complete algorithm for generating admissible trajectories of degree 2.
For the first time, the thought experiments that led to the discovery of the admissible trajectories algorithm were identified. The “hidden” role of mosaic reasoning and its components in this discovery was revealed. Also, all the communication thought experiments run by expression stream to translate internal streams into a written algorithmic mosaic were defined in this Chapter.
45


CHAPTER VI
REVISITING THE DISCOVERY OF GENETIC CODE
6.1.0 Overview
In this chapter, the AD was applied to revisit the discovery of the genetic code. This application of the AD required the investigation of existing mosaics to reveal the rules needed to build a new mosaic out of existing ones. This led to the conclusion that mosaic reasoning is a major component of the AD and may be used to identify the unknown structure of any real world or informational object. Also, mosaic reasoning could be applied by the AD to reveal the rules essential in building one mosaic out of multiple existing ones.
Information mosaics are single- or multi-dimensional tables that appeared for the first time in this discovery. In order to build any information mosaic, the AD must apply the transformation rules to convert a known mosaic into entirely new information one. While investigating the genetic code problem, the AD declared the reading rules, which were needed to read one mosaic and build another one based on that reading. The AD builds the one-dimensional table mosaic as a string of aggregates and the n-dimensional table mosaic as n strings of aggregates. The aggregates of the information mosaic are built out of letter tiles and a visual stream must read these letter tiles according to the reading rules. The aggregates of the input mosaic are read and, based on that reading, are turned into new aggregates in the new information mosaic.
In this discovery, three information mosaics, required during the process of protein synthesis, were identified. Also, two transformation rule phases were declared: transcription and translation. Each phase represents a complete process that is part of the protein synthesis. During transcription, the mRNA mosaic is transformed from the DNA mosaic. The DNA mosaic is built out of a single string of aggregates. Each aggregate is a group of letter tiles. A, T, C or G. The
46


DNA letter tiles are transformed into the corresponding mRNA letter tiles, A, U, C and G, respectively, to form a new string of mRNA aggregates.
During translation, the genetic sequence in the mRNA mosaic is read, according to the reading rules, and forms the peptide information mosaic, the output mosaic. The rest of this chapter discusses the AD trials undertaken to find the correct reading rules to identify the transformation algorithm. Such an algorithm controls the reading of the mRNA to turn it into the peptide mosaic.
6.2.0 Motivation
After applying the AD to the discovery of admissible trajectories, the motivation was to find another discovery from a different field, i.e., one not LG. The genetic code problem was chosen because it is an information problem like LG. Two information problems, from two different fields, required investigating the similarities in the behavior of the AD while solving the two problems, i.e. finding a general structure of the AD. Also, our research goal was to utilize the AD components identified during previous analysis, as well as to reveal any new components leading to the correct solution of the genetic code problem.
6.3.0 The Gamow Thought Experiments (The Diamond Code)
After finding the correct structure of the DNA mosaic [10], the AD triggered a new investigation of the genetic code problem. The AD questioned the way amino acids are aligned to form the protein peptide mosaic based on the genetic code in the DNA mosaic. In the following sections, the major thought experiments manifested by the AD to find the correct solution to the genetic problem are demonstrated, [7],
47


6.3.1 Building the Input Mosaic
In the preliminary stage, the observation stream started by identifying the input mosaic for decoding the genetic information. This stream received the initial inputs, which stated that genetic information is located in DNA. The observation stream also accepted the Watson and Crick schematic representation of the DNA double helix (Figure 9, [37]). It started to erase the particulars by removing all the chemical clutter from the DNA structure since it does not affect the actual genetic message. It described the input mosaic as an abstract DNA mosaic that mimics the schematic representation of the double helix, [7],
Then the observation stream started to generate the construction set. It defined four "oval shape letter tiles,” which correspond to the base tiles of DNA (A, T, G and C). The oval shape was chosen due to the influence of the actual locations of bases in the double helix see Figure 9. The observation stream considered the DNA mosaic as two long strings of symbols, utilizing four letters, since DNA is comprised of two strands. The stream assumed that each string consists of words or groups of letters, i.e., the aggregates. The sequence of tiles inside the aggregates in the DNA mosaic strand is exactly the same as the sequence of base tiles in the corresponding strand of the actual DNA structure.
48


Figure 9. The schematic representation of the double helix, [10],
The observation stream analyzed the structure of DNA to identify the rules needed to build the informational representation of DNA. The stream looked horizontally (i.e., orthogonally across the DNA axes) at the structure of DNA. The stream observed that the DNA strands are twisted around a virtual cylinder. Therefore, it reasoned that the two strands of DNA must be twisted around a cylindrical shape representing the basic structure of the DNA mosaic. The stream further concluded that the strands must twist around this basic structure following the same transformation algorithm identified in the DNA, this mimicking the same twisting of the strands as in the actual structure of the DNA, [10], and [7],
The observation stream saved this visual model of the DNA mosaic as the schematic representation. When building the DNA mosaic, the construction stream referenced this mosaic as the visual model of construction. The observation stream also saved, into the construction set, all the components required to construct the DNA mosaic, which include: the basic structure, the tiles and aggregates, and the transformation algorithm for constructing the aggregates and
49


inserting them along the structure of the DNA mosaic. The observation stream considered an oval aggregate with the proper base tile as a generator. The construction stream could generate every new oval aggregate by shifting the most recently built aggregate along the current DNA strand and twisting the strand around the cylindrical base. After identifying the visual model of DNA and the construction set, the observation stream arranged a thought experiment to build the DNA mosaic and sent the experiment and all needed inputs to the construction stream.
The construction stream built the DNA mosaic by reading one by one the tiles of a DNA strand. Every time the construction stream read a DNA tile, it placed its corresponding letter tile into an oval generator to build one input aggregate. The stream then aligned the input aggregates next to each other along the corresponding strand in the DNA information mosaic. The construction stream twisted that strand along the surface of the cylinder according to the transformation algorithm. When the construction stream finished reading the first strand, it constructed the second strand of the DNA information mosaic following the same steps. While placing the two strands along the virtual basic structure, the construction stream kept in mind the correct position of these strands in the DNA. In other words, the construction stream placed the two DNA strands in the abstract structure following the exact order of the actual DNA structure. Figure 10 shows a part of the abstract DNA mosaic built by the construction stream; in this figure, the letter tiles A, T, G, and C are assigned the numbers 1, 2, 3 and 4, respectively, [64],
Figure 10. Part of the input mosaic as shown in [64],
50


6.3.2 Generating the “Construct Input Mosaic” Shell
While the observation stream was identifying the rules needed to construct the input mosaic through internal thought experiments, the expression stream was working in parallel to create a communication thought experiment to build the shell corresponding to the construction of the DNA mosaic. The expression stream first monitored the observation stream’s internal thought experiments to identify the basic structure of DNA. It tagged the actions and generated the first instruction aggregate: “Build (Basic DNA structure, construction set).” In the construction set, the expression stream saved the cylindrical shape as the structure any construction stream must follow when reconstructing the DNA.
The expression stream also monitored the actions of the observation stream, while it was identifying the base tiles of DNA, and translated them into a symbolic mosaic that contained all possible letter tiles of the DNA information mosaic. This mosaic is a one-dimensional representation for the base tile set, see Table 3.
Table 3. One-dimensional letter tiles mosaic based on the DNA base tiles in Figure 9.
information DNA” shell. The expression stream saved in this shell the first instruction aggregate based on the construction stream’s actions. This instruction aggregate is: “Read (DNA strand,
51


reading rules)” The reading rule, for reading the DNA strand identified by the observation stream and converted by the expression stream into a string of symbols, is: “Read the DNA strand one tile at a time.” The expression stream then marked the converting action from the base tile in the DNA into the corresponding letter tile following the letter tiles symbolic mosaic, Table 3. The expression stream added the instruction aggregate: “Convert (base tile, letter tile, letter tiles table).” The expression stream then added to the instruction aggregate: “Generate (input aggregate, complementarity rules, construction set).” Generating the input aggregate required, according to the construction set, the use of a generator to construct an oval shape aggregate and place the letter tile inside it. The next instruction aggregate saved by the expression stream was: “Align (input aggregate, basic structure, DNA transformation rules).” Every time the construction stream read a base tile, it aligned that tile. This alignment required the twisting of the current strand based on the DNA transformation algorithm. The last instruction aggregate, in the current shell, identified by the expression stream was: “Twist (DNA strand, DNA transformation rules).” The expression reasoned that the same instruction aggregates would be repeated for the length of the DNA strand and also would be repeated for the next DNA strand. The final shell corresponding to construction of the input mosaic can be seen in Figure 11.
52


i— Construct Input mosaic i
L:
Choose (Unread DNA strand, Set of DNA mosaic)
N:
Read (DNA strand, reading rules)
Convert (base tile, letter tile, letter tiles table)
Generate (input aggregate, complementarity rules, construction set)
Align (input aggregate, Basic structure, DNA transformation rules)
Twist (DNA strand, DNA transformation rules)
Repeat (DNA strand length, N)
Repeat (1, L)
Figure 11. “Construct Input mosaic” shell generated by the expression stream.
6.3.3 Identifying the Output Mosaic
After identifying the input mosaic, the observation stream recognized the protein peptide chain as the output mosaic resulting from the transformation of the input mosaic. The observation stream reasoned that the output mosaic would be an information mosaic. Removing all chemical clutter from the amino acids in the peptide chain generated the amino aggregate of the output mosaic. The peptide information mosaic, the output mosaic, is simply a sequence of amino aggregates that are linked together. Each amino aggregate included one letter tile that represented the name of the corresponding amino acid. The observation stream must find and save these amino letter tiles in an amino set that was translated simultaneously, by the expression stream, into an amino table.
The observation stream noted the existence of a hole or cavity between every two base pairs in the DNA mosaic. The stream started a thought experiment to investigate the nature of these cavities by looking at the original schematic representation of DNA. The stream counted twenty different shapes as a result of this thought experiment, [42], Comparing the number of the different cavities and the known number of amino aggregates that build any peptide mosaic, the
53


observation stream found that these were equal and exactly twenty. The observation stream concluded that the peptide mosaic is synthesized inside the double helix, i.e., for each cavity of the DNA mosaic there is a corresponding amino aggregate. To produce the final peptide mosaic, amino aggregates must be aligned in their proper position. Once the genetic code is fully read and translated, the resulting amino aggregates are linked together to create the peptide mosaic. The matching rules for linking amino aggregates together are simple. By considering the peptide mosaic as one sentence and the amino aggregates as words of this sentence, all amino aggregates must be aligned next to each other as words are in any sentence, [42], and [33],
While the observation stream was defining the rules needed to construct the peptide mosaic, the expression stream was building the “Build the peptide” shell based on the actions of the observation stream. This shell was identified while investigating the protein mosaic. This will be explained in detail in Section 7.5.0.
6.3.4 Generating the “Construct Output Mosaic” Shell
The expression stream again monitored the observation stream while identifying the rules needed to build the output mosaic. From the actions of the observation stream, the expression stream defined the “Construct Output mosaic” shell corresponding to the current internal thought experiment. The expression stream added the first instruction aggregate, “Call (amino aggregate, amino set, code set)”, that calls for the correct amino aggregate based on the code aggregate in the current DNA cavity. This amino aggregate was placed inside the cavity, thus the expression stream inserted the instruction aggregate, “Insert (amino aggregate, current cavity, matching rules).” The amino aggregates are then linked to each other according to the protein matching rules and the expression stream added the corresponding instruction aggregate to “Construct
54


Output mosaic” shell. All of these steps were repeated until the entire code is read. Figure 12 shows the final shell for building the output mosaic.
i— Construct Output mosaic i
L:
Call (amino aggregate, amino set, code set)
Insert (amino aggregate, current cavity, matching rules)
Link (amino aggregate, protein matching rules)
Repeat (Length of the input mosaic, L)
Figure 12. “Construct Output mosaic” shell generated by the expression stream.
6.3.5 Defining the Intermediate Mosaic
After defining the DNA and the peptide mosaics, i.e., the input and output mosaics respectively, the observation stream started to investigate the question of how to place the correct amino aggregate inside the DNA cavities. The stream reasoned that the four different letter tiles in the DNA mosaic must code for the twenty different amino aggregates in the peptide mosaic. The stream concluded that this problem must be converted from a chemical problem into a symbolic cryptanalytic problem. There was a four letter code sequence, (A, T, G and C), that needed to be translated into a twenty letter sequence assumed to represent the names of different amino acid aggregates in the peptide mosaic. The observation stream reasoned that a mapping between the DNA and the peptide mosaics must be found. This mapping required that one or more input tiles would be used to generate one code aggregate, which then would be interpreted into one letter tile. This letter tile would be transformed into a specific amino aggregate to be added to the peptide mosaic, [42],
The observation stream then moved to investigate the rules needed to build the code aggregate. The observation stream reduced the problem to finding the answer to the question: how to convert a four-letter string into a twenty letter one? By converting the problem from a chemical into an information-coding problem, the observation stream reasoned that the input
55


tiles might be read in groups of two at a time. Every different reading would represent one code aggregate that code for one amino aggregate. However, when the stream counted all different combinations can be generated from reading two letter tiles at once, it counted only sixteen, i.e., 4x4 possible code aggregates that could be constructed out of reading the letter tiles in multiples of 2. Knowing that the number of amino aggregates should be at least twenty, the observation stream eliminated this multiple reading of input tiles and determined to increase this number.
The observation stream suggested reading letter tiles in triples, i.e., three letter tiles at the time. Again, the stream counted all possible combinations for code aggregates that could be generated from the triple reading of the DNA letter tiles. The observation stream found 64 = 4 x 4x4 different combinations which is much more than the needed code aggregates to code for twenty amino aggregates, [37],
The observation stream reasoned that the amino aggregates can be built between the DNA strands’ mosaic, i.e., in the hole between two DNA base paired tiles, whether between (A and T) or (C and G). The observation stream visualized the location of the amino aggregate to be surrounded from the left and right with those DNA base paired tiles and from top and bottom with any combination of base tiles. This results in the amino aggregate being inside a virtual diamond shape. The observation stream identified this virtual diamond to be a code aggregate that translates the surrounding letter tiles into the inside amino aggregate. This would create a long sequence of code aggregate between the DNA strands, shown in Figure 13. The observation stream declared this long sequence to be the intermediate mosaic that is the link between the input mosaic and the output mosaic.
56


Figure 13. The code mosaic sits between the cavities of the input mosaic as shown in [42], [7],
The observation stream began a thought experiment to identify the correct tiles and aggregates to build the intermediate mosaic. The observation stream assumed that the intermediate mosaic is constructed out of a sequence of code aggregates. One code aggregate of the intermediate mosaic is constructed out of a diamond body tile, at every comer of this diamond body tile is a code tile (1, 2, 3, or 4) that corresponds to one of the of the DNA letter tiles (A, T, U, or G). Every three successive input aggregates are used to build one code aggregate. The observation stream moved forward to find the transformation rules needed to convert the three input aggregates into a single code aggregate.
6.3.6 Assigning the Reading Rules
The observation stream established this thought experiment to identify the correct reading rules needed to construct the intermediate mosaic. The stream investigated the reading rules that the construction stream must follow in reading the DNA mosaic and building the intermediate mosaic. The observation stream followed the first definition requiring that three successive input aggregate be transformed into one code aggregate. Based on the incoming inputs regarding the protein mosaic, the observation stream reasoned that most of known protein peptide chain structures are symmetrical. Therefore, the observation stream postulated that the code aggregate is also symmetrical, i.e., it can be rotated to any direction and still code for the same amino aggregate. For example, a code aggregate that is constructed out of AAC input tiles would be the same as the code aggregates constructed from the ACA and CAA input tiles and all of the three
57


code aggregates would call for the same amino aggregate. The observation stream concluded that this reading rule groups all combinations of the code aggregates from the sixty-four distinct code aggregates into twenty group of code aggregates, [3], Since the number of the code aggregates, sixty-four, is much more than the number of the amino aggregates, twenty, and since more than one code aggregate is mapped to one amino aggregate, the observation stream reasoned that the mapping between input and output mosaics would be many-to-one.
The observation stream also found that the 3D spaces between the base aggregates in the DNA mosaic and the amino aggregates in the protein mosaic are similar. That similarity led the observation stream to confirm to the many-to-one mapping between the input and the output mosaics. Thus, the stream concluded the degeneracy of the genetic code. Degenerated code means that more than one code aggregate corresponded to the same amino aggregate. Having the degenerated code allowed the observation stream to group the sixty-four different code aggregates into twenty groups, every group includes multiple code aggregates that code for one of the twenty amino aggregates, Figure 14, [3], [42], and [64],
Figure 14. Four symmetrical code aggregates could be generated from the genetic code 123 by
the construction stream, [7],
In parallel, the expression stream established a new shell to save all the reading rules identified by the observation stream to control the reading of the input mosaic. The expression stream saved the first reading rule that the genetic code is degenerated. The expression stream initiated two symbolic mosaics to be completed later. The first table, the code table, would contain twenty groups of code aggregates. Each group contains one or more code aggregates.
58


The second table, the amino table, would include twenty different amino aggregates. The expression stream assigned a mapping between the groups in the code set with the amino aggregates in the amino set. Note that both of the sets did not include any specific members. The expression stream waited for these two sets to be identified later on by the observation stream.
The observation stream also established another reading rule that the reading of the code could be performed with no definite direction. This rule was based on the symmetrical nature of the code aggregates identified earlier by the observation stream. Following this definition, the expression stream saved a new reading rule stating that the DNA mosaic could be read from right to left or from left to right in the reading rules shell, [3], The expression stream included the description of this new reading rule in the “Reading rules” shell.
Based on the 3-D space similarities between the base aggregates in the input mosaic and the amino aggregates in the output mosaic, the observation stream reasoned that every base aggregate would be mapped to one amino aggregate. This one-to-one mapping would give the best mesh between the two mosaics, [33], That would maximize the the density of information storage. Therefore, the stream established a new reading rule which stated that the input mosaic must be read in an overlapped manner. Overlapped reading means to read a base tile sequence in the DNA mosaic, for example 1, 3, 3, 2, 1..., as following, the first reading is 1, 3, 3, the second 3, 3, 2, the third 3, 2, 1 and so on. Every triple reading of the input sequence is used to generate one code aggregate in the intermediate mosaic, [3], Again, the expression stream saved the description of the new reading rules in the “Reading rules” shell.
59


6.3.7 Identifying the Matching Rules
6.3.7.1 Local Matching Rules
The observation stream moved forward to identify the local matching rules needed to construct the intermediate mosaic. The observation stream found that three letter tiles from the input mosaic could be used to construct a code aggregate with four corners, and each corner would include a code tile corresponding to these letter tiles. As defined before, the left and right comers of the code aggregate diamond would be facing two base pairs in the input mosaic. Thus, the observation stream concluded that the left and right comers would be attached to complemented code tiles. Therefore, during construction, one of these corners would be assigned based on the genetic information in the input mosaic, while the other corner would be determined using the DNA complementarity base pair matching rules. The observation stream declared a new local complementarity matching rule which stated that the placement of three comers of the diamond would be determined based on the reading of the three input tiles, while the placement of the fourth letter tile would be based on the three letter tiles in the code aggregate. Also, the observation stream noted that the code aggregates are symmetric, which led the stream to introduce a new local interchangeability rule. This rule allows the symmetric code aggregates to be replaceable in the intermediate mosaic. Replacing one code aggregate with asymmetric one changes the final picture of the intermediate mosaic but the whole mosaic will stand. An example of four symmetric code aggregates can be seen in Figure 14.
At every step, while the observation stream visually defines one of the local matching rules, the expression stream was translating it into a written description inside the “Local matching rules of the intermediate mosaic” shell. This description will be used as an input when the AD runs again in future thought experiments. Of course, any incoming input to the AD must first be
60


translated into a visual movie by the impression stream through communication thought experiments before the AD can run it.
6.3.7.2 Global Matching Rules
The observation stream initiated a thought experiment to define the global matching rules. From the reading rule stating that the genetic code is overlapping, the observation stream reasoned that one code aggregate would partially determine the construction of the next code aggregate. This conclusion was due to the fact that two consecutive code aggregates share at least two letter tiles. Also, the placement of one code aggregate in one cavity would force the next code aggregate to be placed in the next cavity. The placement of any amino aggregate inside its code aggregate must be in a stereospecific fashion. Once the reading of the genetic code is completed, the amino aggregates would be linked to each other following the protein matching rules to form the final peptide mosaic, [42], and [64],
Again, the expression stream translated the actions of the observation stream and generated a written description of the intermediate mosaic’s global matching rules. The expression stream saved all of these global matching rules in the “Global matching rules of the intermediate mosaic” shell that will be used later by future runs of the AD.
6.3.8 Defining the Code and Amino Sets
The observation stream defined the input mosaic, the output mosaic, the intermediate mosaic, the reading rules, and the matching rules to construct the intermediate mosaic. The observation stream found that the next step of this discovery is to complete the code and the amino sets, and to correctly map between them. The code and the amino sets are types of scope sets that expedite the run of visual streams while working on the genetic code problem, which eventually reduces the path to find the correct solution of the problem. The stream started the thought experiment to
61


complete the code set with all the code aggregates. The stream triggered the construction stream to start filling the code set based on the rules identified by the observation stream to build any code aggregate.
The construction stream received the inputs from the observation stream and launched the thought experiment to find all combinations of code aggregates. The construction stream created all possible diamond aggregates using the following approach. First, the construction stream generated the body tile of the code aggregate with a diamond shape. Then, the stream added the code tiles to the left and right corners of the body tile keeping into mind that these code tiles must be base-paired tiles. For example, if the constructed stream added the code tile «3» in the right comer, it would automatically add its complement tile, «4», in the left comer. Then, the construction stream would add all possible combination of code tiles in the top and bottom comers «1»«1», «2»«2», «3»«3», «4»«4», «1»«2», «1»«3», «1»«4», «2»«3», «2»«4» and «3»«4», thus generating ten independent code aggregates. The construction stream repeated the same steps by adding the other two complement code tiles «1» and «2» to generate another ten distinct code aggregates. Then, the stream rotated every code aggregate and applied the interchangeability rules to group all the symmetric code aggregates, that code for the same amino aggregate. As was expected, the construction stream built twenty different groups, each group representing one amino aggregate, [42], and [64],
While the construction stream was building the code aggregates, the expression stream was drawing a representation of every code aggregate in the code table. The result was a symbolic mosaic that includes all twenty groups of code aggregates, as shown in Table 4.
62


Table 4. The code mosaic generated by the expression stream as Shown in [42], [7],
amino aggregates in the amino table. Even though not all names of amino aggregates were confirmed to be part of the protein mosaic, the observation stream still added them into the amino set. Then, the observation stream, having a virtual model of an existing DNA information mosaic, triggered a thought experiment to construct the intermediate mosaic from the genetic code in the DNA mosaic. The observation stream passed that experiment with all associated inputs to the construction stream.
6.3.9 Constructing the Intermediate Mosaic
The construction stream started a thought experiment to build the intermediate mosaic by reading the genetic code in the virtual input model. For a better explanation, assume that the
63


genetic code of the DNA mosaic is 123123... . The construction stream started to read the genetic sequence following the reading rules. It read the first three letter tiles. With no restriction in direction, the construction stream chose to read the code from left to right. Thus, the first successive input aggregates the construction stream read were 123. The stream first generated the body diamond tile of the code aggregate. It then added the first letter tile «1» at the top corner of the body tile. Following the local interchangeability rules, the construction stream could add the first letter tile «1» arbitrarily in any corner of the body tile. Next, the construction stream read the input aggregate and placed the corresponding letter tile «2» to the left of the tile «1». The next letter tile to be added must be complementary to the «2» tile, thus the construction stream added «2'» to the right of the tile «1» and in front of the «2» tile. In the last corner of the body tile, the construction stream placed the complement of the last input aggregate. The letter tile in the empty comer is determined by the complementarity-matching rule. Reading the input aggregate «3», the construction stream added «3'» to the empty corner of the body tile. Once the first code aggregate is built, the stream inserted it into its corresponding cavity, as shown in Figure 14, [42], and [64],
Then, using overlapped reading, the construction stream retrieved two input aggregates and read the next three successive input aggregates, 231, and repeated the steps to produce the second code aggregate. After creating the second aggregate, the stream inserted it into the cavity to the right of the previous one. The construction stream continued until the whole genetic sequence was read, see Figure 15.
64


Figure 15. Part of the intermediate mosaic constructed form the genetic sequence “123123”, [7],
The Gamow’s thought experiments were the first thought experiments to state the genetic problem as an information problem instead of a biochemical problem. During this run of the AD, the observation stream established the basis that any following AD run should follow to find the correct solution. To decipher the genetic code, the AD must first correctly identify the input and output mosaics. Then, the AD, manifesting visual streams, must discover the transformation rules needed to transform the input mosaic into the output one. These transformation rules must include the correct definition of the reading rules that control the reading of the input mosaic. This reading is needed to build the intermediate mosaic. The intermediate mosaic was defined, for some time, as a long sequence of code aggregates; each aggregate is built out of at least three letter tiles. Every letter tile in the code mosaic has a corresponding letter tile in the input mosaic. In addition to the definitions of the input, output, and intermediate mosaics and the transformation rules, the AD must build the code and amino sets. The code set must include all code aggregates and the amino set must contain all amino aggregates needed to built the intermedia mosaic and the peptide mosaic, respectively.
6.4.0 The Crick Thought Experiments
6.4.1 Validating the Diamond Code
In an effort to validate the diamond code, the AD initiated the observation stream. From the impression stream, the observation stream received the structure of the insulin peptide mosaic. Having this structure, the observation stream began a thought experiment to verify the correctness of the diamond code, [34], Insulin is the peptide protein mosaic that is constructed
65


out of fifty-one amino aggregates, [41], The observation stream sent the structure of insulin along with the results of Gamow’s experiment for evaluation to the validation stream.
The validation stream chose two different insulin mosaics whose structures were fully known: insulin B and B-corticotropin. Then, it initiated the construction stream to apply the diamond code to reconstruct the insulin mosaic. While studying the insulin B mosaic, the construction found the aggregate sequence «Leu.Tyr.Leu». The construction stream also found the sequence «Ser.Tyr.Ser» in B-corticotropin. The construction stream reasoned that these two sequences are of type «xyx». The «xyx» sequence is a peptide mosaic containing three consecutive amino aggregates, where the first and last aggregates are the same and the middle one is different. The stream began to apply the diamond code to create an intermediate mosaic that codes for any «xyx» sequence, see [7], and [34] for details.
In order to illustrate the steps followed by the construction stream, consider that «x» is represented by the code «111» and «112» represents «y». Since the reading rules require the overlapped reading of the diamond code, the code sequence of «xy» must be «1112». Figure 16 (left and middle) shows the code aggregates that correspond to the amino aggregates «x» and «y», respectively. The third amino aggregate in the sequence «xyx», «x», must be coded from the input sequence «111» or «1'1' 1'», since both code sequences generate the same code aggregate. However, following the overlapping reading requirement, the next reading is «12n», where «n» is any letter tile. This reading would generate amino aggregates different than «x», see Figure 16 (right). The construction stream failed in all attempts to apply the diamond code to generate the intermediate mosaic that codes for the sequence «xyx». This failure led the validation stream to invalidate the diamond code, [7], and [34],
66


Figure 16. The code aggregates generated from the input sequence 1112n.
6.4.2 Redefining the Input Mosaic
Once the diamond code was eliminated, the AD started a new cycle to look for a solution to the genetic code problem. The observation stream initiated a thought experiment to find the correct input mosaic, which includes the genetic information. In the preliminary stage, the observation stream gathered any known inputs about the genetic code. The first input showed that the sequence of amino acids is determined by genes. The constructed sequence of amino aggregates represents the protein peptide mosaic. The stream received the assumption that the DNA backbone’s base sequence carries the genetic code. The observation stream continued receiving new inputs about the cell nucleus’ components. Cells of human beings include DNA, RNA, and cytoplasm. The stream collected information about the RNA molecule. It found that RNAs are single-strand structured molecules that are synthesized from one strand of the DNA in the nucleus of the cell. The cytoplasm of a cell is the fluid inside the cell that surrounds its contents. The observation stream also received information that proteins are synthesized in the cytoplasm and that the DNA is located only in the cell’s nucleus. This led the observation stream to reason that a copy of the DNA must be created in the nucleus and sent to the cytoplasm. Analyzing these inputs, the observation stream concluded that the RNA mosaic contains the genetic information, by copying it from the DNA, to construct the protein, [7], and [45],
Further, the observation stream divided protein synthesis into two phases: transcription and translation. During transcription, the DNA mosaic is read to construct the RNA. Since the length of RNA is much shorter than that of DNA, the observation stream concluded that the genetic
67


information inside the DNA mosaic codes for more than one protein mosaic, and that each segment of the code corresponding to one protein mosaic would be copied into one RNA mosaic. The stream reasoned that the resulting RNA mosaic from the transcription phase would be used later as an input mosaic to the translation phase in order to create one protein mosaic, [7], [50], and [37],
Having a new definition of the input mosaic, the observation stream ran a thought experiment to find the correct structure of this new mosaic. The stream had inputs stating that the RNA mosaics are built out of four different RNA base tiles, A, U, G and C. The stream reasoned that during transcription, the RNA mosaic is constructed by reading the base tiles of one DNA strand and transformed them into their corresponding base tiles in the RNA mosaic. Each one of the DNA base tiles, A, T, G and C, represents one of the RNA base tiles as follows, A to A, T to U,
G to G and C to C , as shown in Figure 17. The observation stream defined the input mosaic to be an abstract copy of the RNA mosaic after removing all chemical clutter. The stream defined every base tile in the RNA mosaic as a letter tile in the input mosaic (A, U, C, and G). These letter tiles are linked together as a long sequence that, if read correctly, provides guidance to the engine of protein synthesis, while building the output mosaic, [7], [45], [50], and [37],
68


Figure 17. Constructing the RNA mosaic from a single strand of the DNA mosaic.
The observation stream also studied the nature of the output mosaic. With no new inputs that contradicted Gamow's definition of the output mosaic, the observation stream kept the same definition of the output mosaic described in Section 6.3.3.
6.4.3 Evaluating the Reading Rules
Next, the observation stream investigated the reading rules defined during Gamow's experiment, [7], The stream did not find any new inputs that invalidated the triple reading. Therefore, the observation stream accepted the triple reading of the genetic code as defined in Section 6.3.6.
The observation stream also investigated the overlapped reading. In Gamow's experiment, this reading method was concluded to be based on the 3D space similarities between the protein mosaic’s aggregates and the DNA mosaic’s aggregates. The observation stream excluded this conclusion based on the new assigning of the input mosaic as the RNA mosaic instead of the DNA mosaic. However, the observation stream reasoned that even though the input mosaic of the translation phase is the RNA mosaic, this mosaic could still be mapped one-to-one to its
69


corresponding genetic portion of the DNA mosaic. That led the observation stream to consider the DNA mosaic as the initial step of several transformation steps in building the peptide mosaic. Having no new evidence that contradicted overlapped reading and to maximize the density of information storage, the observation stream accepted the overlapped reading of the input mosaic, see [7], and [33] for details.
Finally, the observation stream examined the bidirectional code reading. The stream searched for any known input mosaic that could generate two different sequences of the peptide mosaic. These two sequences must be reversed, such as «Thr.Pro.Lys.Ala» and «Ala.Lys.Pro.Thr», which could be a result of reading the genetic code in both directions. The observation stream reasoned that bidirectional reading means that the two resulting peptide mosaics should be distinguishable by nature, but no valid inputs supported this hypothesis, [7], and [34], This drove the observation stream to investigate the way letter tiles could be grouped into code aggregates. To find the answer, the observation stream looked to the x-ray pictures of the backbone of DNA. These pictures showed no regularity in the DNA backbone to conclude how the letter tiles are grouped together. The stream also realized that the triple reading of the code would create various possibilities for an incorrect reading of the code.
At the end of this thought experiment, the observation stream found lab evidence proving that the code could only be read in one direction. This evidence guided the observation stream to reject the bidirectional reading of the code and hypothesis to conclude that the genetic code must be read from one direction only, [7], [34], and [35],
Once the bidirectional reading was invalidated, the observation stream discarded the symmetric feature of the code aggregates resulting from the bidirectional reading rule. From the conclusions of the Gamow experiment, the code aggregate is symmetric because it can be placed
70


into the input cavity from any side. However, the results of the other thought experiments, as inputs to the observation stream, found that the DNA structure has bonds that stick out perpendicularly to its axis and that the DNA structure’s surface has knobs on it. The stream found different aspects of the DNA’s two sides. It concluded that placing code aggregates inside the cavity of the DNA would not be possible, [7], and [34],
6.4.4 Rebuilding the Amino Set
Next, the observation stream started a thought experiment to define the twenty amino aggregates within the amino set. The stream received the amino set that was initiated in Gamow's experiment. Since Gamow's amino set was assumed with no solid evidence, the observation stream reasoned that the correct definition of the amino set is critical to accurately deciphering the genetic code.
Moreover, even before the code is deciphered, knowledge of the amino set would accelerate constructing the correct mapping between the amino aggregates and code aggregates.
Essentially, defining the correct amino set would minimize the time needed to crack the genetic code, [7], and [34],
In order to build the correct amino set, the observation stream utilized current information about amino acids in protein. Inputs showed that the same amino acids are used to construct proteins in most living creatures. Also, the DNA base tiles in all organisms are built out of the same set of bases. These similarities guided the stream to conclude that the genetic code is universal, which means that the genetic code in most living organisms would be deciphered following the same transformation of DNA code aggregates into amino aggregates. The stream identified the amino set by investigating several known protein chains. The stream eliminated all amino acids never found in any protein chain, such as Norvaline and Hydroxy Glutamic. It also
71


excluded the amino acids that are seen in small peptides. In contrast, the stream added all of the amino acids that are rarely seen in the protein chain, such as Asparagine and Glutamine. Using all known inputs about amino acids, the observation stream saved, in the amino set, twenty amino aggregates, [7], [33], and [34], This set was translated by the expression stream into the amino table, which was used as an input for other thought experiments and was eventually confirmed to be correct. The Crick thought experiments did not provide any description of a coded algorithm that would be used to code the RNA mosaic into a peptide mosaic. However, these thought experiments resulted in defining the amino set that was proved correct in later thought experiments.
6.5.0 Modifying the Reading Rules
By receiving new inputs that forced the AD to modify the reading rules, the AD triggered the observation stream to start a thought experiment to investigate the correctness of the genetic code reading rule. The stream, in its preliminary stage, retrieved all the conclusions coming from Crick’s thought experiments and used the same definition of the input and output mosaics.
These thought experiments concluded that the code should be read in triples, overlapped, and unidirectional.
The observation stream looked at new lab inputs that changed the genetic structure of virus RNA by treating it with nitrous acid. The modified RNA is then read to construct the peptide mosaic. The result of this construction is a new virus that has a similar genetic sequence to the original virus, but differs in only one amino acid. Having these results, the observation stream questioned the overlapped reading of the genetic code. If the code overlapped, modifying one base tile in the RNA sequence would affect at least two amino aggregates in the output mosaic,
72


but only one amino aggregate was affected. Therefore, the observation stream concluded that the reading of the genetic code must be non-overlapped, [7], and [36],
Once the observation stream eliminated overlapped reading, it reasoned that the mapping between the input and output mosaics is one-to-one. There was no need to construct any intermediate mosaic between the input and output mosaics. When building the intermediate mosaic from the RNA mosaic using non-overlapping reading, the resulting intermediate mosaic would be identical to the RNA mosaic and, therefore, not necessary. At the end of this thought experiment, the observation stream saved, into the construction set, the new reading rules for the genetic code. These rules stated that the reading of the genetic code must be triple, non-overlapped, and unidirectional. In parallel, the expression stream saved the reading rules into a shell that included a description of these rules. This shell will be used as input for any future thought experiments.
6.6.0 Deciphering the First Triple
The Crick thought experiment identified the input mosaic to be the RNA mosaic. However, there are many types of RNA. Therefore, a thought experiment was constructed to determine if all RNA types are used during protein synthesis and what their roles were in the process. The observation stream received lab inputs that observed an RNA mosaic labeled with an amino aggregate, tRNA. After labeling, tRNA participates in protein synthesis by carrying amino aggregates to be added to the peptide mosaic, [7], and [56],
The observation stream reasoned that tRNA must act as an adapter mosaic. This mosaic carries the amino aggregates during translation. Since tRNA is an RNA mosaic, and all RNA mosaics are constructed by replicating the base tiles in the tRNA, the observation stream identified the basic structure of the tRNA to be the same as the one defined in the Crick’s
73


thought experiments. However, the final structure of tRNA, as seen in x-rays, wasn’t known at that time, therefore, the observation stream concluded that a visual model of the tRNA mosaic could not be built. The stream found that the role of the tRNA is auxiliary: it is only needed for carrying amino aggregates during translation. The construction stream could visualize any arbitrary structure of tRNA and link it to an amino aggregate, and it would still participate in the translation process, [7], and [58],
6.6.1 Defining the Input Mosaic
There was no proof of the input mosaic for translation. Thus, the AD started its run to find the correct definition of the input mosaic. It triggered the observation stream to begin a series of thought experiments to solve the genetic code problem. The observation stream first initiated a thought experiment to define the input mosaic. The observation stream accepted the results of the two lab experiments that tested the nature of the input mosaic. Both experiments used a lab extract containing the mosaics involved in translation, i.e., all different types of the tRNA mosaics and the amino aggregates. For the first lab experiment, the DNA mosaic was added to the extract to be the input mosaic. In the other lab experiment, the input mosaic added to the extract was a special type of RNA, messenger RNA (mRNA). The results showed the existence of a peptide mosaic only in the extract containing the mRNA mosaic. With these results, the observation stream concluded that the input mosaic for translation is the information mosaic that is constructed from the mRNA mosaic, [7], [55], and [56],
In order to build a visual model of the mRNA information mosaic, the observation stream declared that this mosaic is an abstract copy of the mRNA mosaic. Since the mRNA mosaic is another type of an RNA mosaic, the stream defined the tiles of the mRNA to be the four RNA base tiles (A, U, C, or G). Each of these base tiles would be converted into the corresponding
74


letter tiles in the mRNA information mosaic, after removing all the chemical clutter. The reading rules of the genetic code were confirmed to be triple, non-overlapped, uni-directional reading. The observation stream knew that every three consecutive letter tiles would be grouped together to build one code aggregate. Every code aggregate in the mRNA mosaic would code for one specific amino aggregate. The length of the tRNA mosaic would be the same length of the gene area in the DNA mosaic. The gene area is the sequence of base tiles in the DNA mosaic that code for a single protein mosaic.
6.6.2 Generating the “Construct information mRNA” Shell
Following the new definition of the input mosaic, the expression stream modified the input mosaic construction shell. The expression stream tagged the actions of the construction stream while it was building the basic structure of the mRNA. It generated the first instruction aggregate: “Build (Basic mRNA structure, construction set).” The expression stream also saved the set of the mRNA letter tiles as declared by the observation stream in a one-dimensional symbolic mosaic, as shown in Table 5.
Table 5. One-dimensional mosaic of the mRNA letter tiles converted from the DNA base tiles.
DNA Base tile mRNA letter tile
(A
#
â–  C
G
The expression stream saved the steps followed by the construction stream in the “Construct information mRNA” shell. The first instruction aggregate saved in this shell is: “Read (DNA strand, reading rules).” The reading rule identified by the observation stream and translated by the expression stream for reading the DNA strand is “Read one DNA strand tile by tile.” The
75


expression stream then marked the conversion from the base tile in the DNA to the corresponding letter tile by following the letter tiles mosaic, in Table 5. The expression stream created the instruction aggregate: “Convert (DNA base tile, mRNA letter tile, letter tiles mosaic)” and saved it in the current shell. Next, the instruction aggregate: “Add (letter tile, mRNA, matching rules)” was also saved to the shell. The expression stream reasoned that the same instruction aggregates would be repeated starting from the instruction aggregate Read until all base tiles of the current gene area in the DNA strand are scanned. The final “Construct information mRNA” shell can be seen in Figure 18.
Construct information mRNA Choose (gene, Set of DNA mosaic) L: Read (DNA strand, reading rules) Convert (base tile, letter tile, letter tiles table) Add (letter tile, mRNA, matching rules) Repeat (gene length, L)

Figure 18. “Construct information mRNA” shell generated by the expression stream.
6.6.3 Translation Components
After defining the input mosaic, the observation stream identified the components that engaged in translation: the mRNA input mosaic, the peptide output mosaic, the tRNA adapter mosaic, the code and amino sets, and the reading rules. In order to generate the peptide mosaic, the construction stream read one code aggregate in the mRNA, tile by tile, and called the tRNA that carries the corresponding amino aggregate. The incoming amino aggregate would be added to the growing peptide mosaic. The construction stream must read the input mosaic in one direction only. Moreover, previous thought experiments showed that the reading of the code is non-overlapped, i.e., triple after triple. Since the lab results showed that mRNA, tRNA,
76


ribosome, and amino are the mosaics needed during translation, the observation stream reasoned that all of the components involved in translation were identified. It concluded that the next step in deciphering the genetic code is to map between the code aggregates of the mRNA and the amino aggregates in the peptide mosaic, i.e, to find the correct mapping between the code and the amino sets, [7], [55], and [36],
6.6.4 The Mapping Thought Experiments
In order to map correctly between the code and amino sets, the observation stream reasoned that mappings between a known structure of mRNA and peptide mosaics must be found. Since the mappings between the two mosaics are one-to-one, if an mRNA with known code aggregates was transformed into a peptide mosaic with known amino aggregates, the meaning of every code aggregate in the mRNA would be known, i.e., the amino aggregate that each code aggregate represents. To apply this reasoning, the observation stream called for twenty lab experiments to be performed. These lab experiments dealt with a cell-free extract that contains a poly-U mRNA mosaic, the twenty amino aggregates in the amino set, and twenty tRNA mosaics (one for every different amino acid). The stream found that, having an mRNA mosaic with the same code sequence, e.g. UUUUU..., would result in a peptide mosaic of all of the same amino aggregates.
In the lab, in order to see the resulting peptide mosaic, its amino aggregates must be radioactive. Since the peptide mosaic is constructed out of a series of the same amino aggregate and there are twenty possible amino aggregates to construct the peptide mosaic, twenty lab experiments were prepared. Each of them included only one radioactive amino acid, which differed in every lab experiment. The stream concluded that the extract having the radioactive amino aggregate, which corresponded to the code aggregate, would have a radioactive peptide mosaic.
77


To visualize those lab experiments, the observation stream triggered a thought experiment to build the poly-U mRNA mosaic and passed the experiment to the construction stream. Using the construction rules for building the input mosaic, Section 6.6.1, the construction stream built the input mosaic of poly-U, as seen in Figure 19.

uuu uuu uuu uuu uuu
t_J t i i t ‘ t
Code ag â– rcgatcs
Figure 19. An mRNA mosaic of Poly-U code aggregates, [7],
In one of the above lab experiments, which included radioactive phenylalanine, the resulting peptide mosaic was also radioactive. The observation stream initiated the construction stream to build a poly-phenylalanine mosaic based on the construction rules for building an information peptide mosaic. These construction rules came from previous thought experiments. A protein peptide mosaic is the abstract representation of the protein mosaic built out of amino aggregates that are linked to each other in a sequential line, [7], [45], [50], and [37],
After building the input and output mosaic, the construction stream mapped the code aggregates in the mRNA to the amino aggregates in the peptide mosaic. The observation stream found that every code aggregate of «UUU» was mapped to one amino aggregate «phenylalanine», thus it concluded that the code aggregate «UUU» represents the amino aggregate «phenylalanine». Figure 20 shows the mapping between the Poly-U mRNA mosaic and the resulting peptide mosaic after translation, [7], [55], and [56], Following the same technique, the observation stream concluded that «AAA», «CCC» and «GGG» code aggregates represent the amino aggregates «lysine», «proline» and «glycine», respectively.
78


Poly-U mRNA mosaic
uuu UUU uuu UUU UUU
4 y k 4 t \ k 4 f y k 4 f \ k 4 t \ k t
Phe. Phe. Phe. Phe. Phe.
Peptide mosaic
Figure 20. The mapping between the code and amino aggregates, [7],
By the end of this AD run, not all code aggregates were mapped to their corresponding amino aggregates. However, the AD identified the process rules needed to map between the code and amino sets. First, the AD must identify the structure of the mRNA mosaic. Then, it observes the results from the lab experiments to find the correct structure of the resulting peptide mosaic. After identifying the peptide mosaic, the construction stream must map the code aggregates to the respective amino aggregates.
By following the above process rules, subsequent thought experiments managed to map other code aggregates to their amino aggregates. After many rounds, the AD deciphered all sixty-four code aggregates. The AD also identified two special sets of code aggregates, the initiation set and the termination set. Both sets are part of the scope sets that help in eliminating the run time of the future thought experiments. The initiation set includes all start code aggregates that trigger the building of the peptide mosaic. The initiation set included only one start-code aggregate, «AUG». The termination set contains three end-code aggregates, «UAA», «UAG» or «UGA», that signal the termination of the translation process, [7], [55], and [56],
Every time the discovery visual streams decipher one code, the expression stream saved the results into a “Genetic code” symbolic mosaic. This mosaic includes the meaning of every code aggregate and was built as a result of multiple thought experiments. The output of the expression stream was a symbolic mosaic that included the full genetic code, Table 6.
79


Table 6. The final genetic code mosaic3, [37],
First letter tile Second etter tile 4 Third letter tile
4 U C A G 4
UUU^ Phe UCU^ Ser UAU—» Tyr UGU^ Tyr u
u UUC^ Phe UCC^ Ser UAC^ Tyr UGC-> Tyr c
UUA—> Leu UCA—> Ser UAA-> Stop UGA-> Stop A
UUG-> Leu UCG-> Ser UAG-> Stop UGG^ Stop G
CUU^ Leu CCU^ Pro CAU^ His CGU^ Arg U
c CUC^ Leu CCC^ Pro CAC-> His CGC^ Arg C
CUA—> Leu CCA-> Pro CAA^ Gin CGA^ Arg A
CUG-> Leu CCG-> Pro CAG^ Gin CGG^ Arg G
AUU—» lie ACU—» Thr AAU—» Asn AGU^ Ser U
A AUC—» lie ACC-> Thr AAC^ Asn AGC^ Ser C
AUA^ He ACA^ Thr AAA—> Lys AGA^ Arg A
AUG^ Met ACG^ Thr AAG-> Lys AGG^ Arg G
GUU-> Val GCU-> Ala GAU-> Asp GGU^ Gly U
ri GUC-> Val GCC^ Ala GAC^ Asp GGC^ Gly C
u GUA^ Val GCA^ Ala GAA^ Glu GGA^ Gly A
GUG^ Val GCG^ Ala GAG^ Glu GGG^ Gly G
6.7.0 Results
Several thought experiments involved in cracking the genetic code were identified. In each of them, a major role of mosaic reasoning was identified in finding the solution to this problem. Tiles and aggregates, local and global matching rules, information mosaics and an unstructured environment led the AD towards the correct solution. New components of mosaic reasoning, with an essential role in the discovery, were revealed that could be seen in any discovery of the same nature. A new type of mosaic, the information mosaic, was declared. Information mosaics are critical to deciphering the genetic code, because the genetic code is a special type of information. In addition, four different scope sets were identified: the initiation, termination, 3
3 In this table, the genetic code consists of letter tiles; three of these letter tiles represent a code aggregate that is mapped to one amino aggregate. For example, the UUU code aggregate is mapped to the Phe. amino aggregate.
80


code, and amino sets. Each of these sets was essential in cracking the genetic code and reducing the time required to find the correct solution.
Analysis also revealed new components of the transformation rules, the reading rules and the process rules. Reading rules will be used in any discovery that requires the reading of an information mosaic. These rules control the reading of one information mosaic to construct another mosaic, based on the read information. The process rules were identified to control the actions which must be applied by the AD to map the code and amino sets. Following these rules allowed the deciphering of every code aggregate in the code set.
Two new types of transformation rules were identified. These new rules are involved in the two phases of the protein synthesis process: transcription and translation. The reading algorithm was also revealed and applied during the translation to read the mRNA mosaic and convert it into the peptide mosaic. In the next chapter, the work on translation is continued with the goal of defining the full process of translation.
81


CHAPTER VII
REVISITING THE DISCOVERY OF PROTEIN TRANSLATION
7.1.0 Overview
As mentioned in Section 4.4.0, the translation process is the second phase of protein synthesis. During translation, the genetic code in the mRNA is read and transformed into a sequence of amino aggregates. These aggregates are attached to each other to form the preliminary structure of the protein mosaic, i.e., the peptide mosaic.
The AD identified the fact that different mosaics participate in the translation. They are the tRNA, mRNA, and ribosome mosaics. The AD also recognized the output of translation as the peptide mosaic. The construction of each was emphasized by mosaic reasoning, where the tiles, aggregates, and the matching rules are declared and saved in the construction set.
Based on incoming inputs, the AD divided the translation process into three smaller processes: initiation, elongation, and termination. These three processes run sequentially starting with the initiation process, where the ribosome links to the mRNA. Once the ribosome is attached to the mRNA, it will scan the mRNA to read the genetic code. According to the genetic reading, the ribosome will compare the current mRNA’s codon (code aggregate) to all the codons belonging to the initiation set. If the ribosome finds a match, then it will end the initiation and enter the elongation process.
During elongation, the ribosome continues to read the mRNA’s codon and calls for the tRNA that carries the corresponding amino aggregate. Every incoming amino aggregate will then be added to the growing peptide mosaic. Elongation ends when the ribosome reads a termination codon. Once this codon is found, the translation process will end by releasing, or detaching, all of the mosaics, used by the process, including the new peptide mosaic.
82


Full Text

PAGE 1

FINDING THE COMPONENTS OF THE PRIMARY LANGUAGE FOCUSING ON THE ALGORITHM OF DISCOVERY b y NESREEN ALHARBI B.S., King Abdulaziz University Jeddah Saudi Arabia, 2001 M.S., University of Colorado Denver USA, 2009 A thesis submitted to the Faculty of the Graduate School of the University of Colorado in partial fulfillment of the requirements for the degree of Doctor of Philosophy Computer Science and Information Systems Program 201 8

PAGE 2

! ii This thesis for the Doctor of Philosophy degree by Nesreen Alharbi h as been approved for the Computer Science and Information Systems Program b y Gita Alaghband, Chair Boris Stilman, Advisor Tom Altman Jahangir Karimi Ashis Biswas Date May 12, 2018

PAGE 3

! iii Alharbi, Nesreen. (Ph.D., Computer Science and Information Systems Program ) Finding the Components of the Primary Language Focusing on the Algorithm of Discovery Thesis dire cted by Professor Boris Stilman ABSTRACT In the late nineteen fifties, Professor John von Neumann suggested that the human brain uses an internal l anguage for mental calculation. He named that language the Primary Language. He suggested that the Primary Language differs from all the human (Secondary) languages used for communication. W e consider Primary Science to be the science done with the Primary Language, that is differs from the familiar, conventional science. The ultimate goal of this research is to reveal the nature of the Primary Language. To accomplish that, it is important to find those algorithms, used by the human brain, based directly on this language. The results of previous research revealed the existence of at least two ancient algorithms critical for the development of human intelligence. It is our assumption that in absence of the secondary languages, during times long past, those algorithms would have directly utilized the Primary Language. These two algorithm s are Linguistic Geometry (LG) the algorithm for optimizing warfight ing strategies, and t he Algorithm of Discovery (AD) the algorit hm for inventing new algorithms . The main hy pothesis of this research is that the Primary Language is the "language" of visual streams (mental "movies"). Another hypothesis states that the AD is a universal algorithm used for making discoveries. It suggests that the AD is based on multiple thought e xperiments, which manifest themselves via visual streams. It appears that visual streams are the only interface to the AD. The AD operates with three classes of visual streams: observation, construction, and validation. These visual streams can run concurr ently and exchange information. Each stream may initiate additional thought experiments, program

PAGE 4

! iv them, and then execute them. Visual streams are used by the AD to construct new algorithms and, in this way, make discoveries. The current objective is to furt her investigate the AD by applying it to various discoveries, from different domains, such as Computer Science, Molecular Biology, etc., in order to reveal its inherent details. During this research, it was found that some components of the AD are utilized for every discovery , while others may be utilized for specific discoveries only. Once the major elements of the AD are revealed, generalizing them will lead to a complete understanding of the AD. A comprehensive understanding of the AD and its components will lead to the implementation of the generalized AD, i.e., the final goal of making discoveries on demand. The expectation is that the implemented AD will have a profound impact on all branches of science, including Computer Science, and, in particular , Artificial Intelligence (AI). The form and content of this abstract is approved. I recommend its publication. Approved: Boris Stilman

PAGE 5

! v I Dedicate This Thesis to My Husband, Yousef My Parents, Ghuzaie l and Mohammed My Daughters, Maria, Lin, and Dana Without your support and love I could not have done it.

PAGE 6

! vi ACKNOWLEDGMENTS First and f oremost, I would li ke to thank my advisor, Professor Boris Stilman, to whom I am so grateful for the amount of time that he spent guiding and supporting me through my research. I thank him for his patience , enthusiasm, motivation, and guidance throughout this entire process . I would never have been able to finish my dissertation without his continuous advising and immense knowledge. M y sinc ere thanks to my committee members Professor Gita Alaghband, Professor Tom Altman, Professor Jahangir Karimi, and Professor Ashis Biswas for their encouragement and valuable feedback . You all have made my defense an enjoyable event. I would like to especia lly thank Pr ofessor Gita Alaghband , the chair of my committee , from whom I have been receiving extensive guidance in both personal and professional aspects since the beginning of my PhD studies. Thank you Professor Alaghband , nothing would have been achiev ed without your support. I would like to ex press my extreme thanks to Professor Christopher Miller , Department of Integrative Biology at University of Colorado at Denver , for all the time and effort he spent review ing the accuracy of the analysis of the ge netic code and the protein translation in this research. Also, I would like to thank Professor Altman , from the bottom of my heart , for reviewing my proposal and providing me with valuabl e feedback that helped improve my work. In addition, I would like to than k Mrs. Hanna Altman for helping me check the accuracy of the biology information in my proposal , and I would like to thank Diane Yoha, Sara h Mandos , and all the staff of the C omputer S cience department for their assistance and support.

PAGE 7

! vii In addition, m y appreciation and gratitude to the government of Saudi Arabia for granting me a full scholarship that enabled me to pu rsue my PhD studies and to attend this great university . Finally, and most importantly, my sincere thanks to my amazing family . My husband Yousef who had to live far away from us in order to give me the opportunity to study . My daughters Maria, Lin, and Dana who picked up the slack at home and helped each other, becoming independent little ladies an d allowing mom to study. Big thanks to my fa ther Mohammed who left home to come with me and he lp ed with the upbringing of my daughters while they were young. Another big thanks to my mother Ghuzaiel for traveling all the way from Saudi Arabia to the United stated for my sake. She spent four months t aking care of my little daughter so that I would have enough time to prepare for the preliminary exam ; her presence was such a huge relief. Their continuous love and support for my studies is the reason why this research is completed.

PAGE 8

! viii TABLE OF CONTENTS CHAPTER I. INTRODUCTION ................................ ................................ ................................ ....................... 1 1.1.0 Overview ................................ ................................ ................................ .......................... 1 1.2.0 Method of Study ................................ ................................ ................................ .............. 2 1.3.0 Motivation ................................ ................................ ................................ ........................ 3 1.4.0 Objectives ................................ ................................ ................................ ........................ 3 1.5.0 Work Domain ................................ ................................ ................................ ................... 4 1.6.0 Challenges ................................ ................................ ................................ ........................ 5 1.7.0 Contribution ................................ ................................ ................................ ..................... 6 1.8.0 Dissertation Structure ................................ ................................ ................................ ....... 8 II. THE PRIMARY LANGUAGE ................................ ................................ ................................ 10 2.1.0 Overview ................................ ................................ ................................ ........................ 10 2.2.0 The Primary Language (PL) ................................ ................................ .......................... 10 2.3.0 Visual Streams and Thought Experiments ................................ ................................ ..... 11 2.4.0 Types of Visual Streams ................................ ................................ ................................ 13 2.4.1 Discovery Streams and the Algorithm of Discovery (AD) ................................ ....... 14 2.5.0 Proximity and Mosaic reasoning ................................ ................................ ................... 16 III. COMMUNICATION STREA MS AND ALGORITHMIC MOSAICS ................................ . 19 3.1.0 Overview ................................ ................................ ................................ ........................ 19 3.2.0 Motivation ................................ ................................ ................................ ...................... 19 3.3.0 Communication Streams ................................ ................................ ................................ 20 3.4.0 Algorithmic and Symbolic Mosaic s ................................ ................................ ............... 22

PAGE 9

! ix 3.5.0 Results ................................ ................................ ................................ ............................ 24 IV. INTRODUCTION TO THE INVESTIGATED DISCOVERIES ................................ .......... 26 4.1.0 Introduction to Linguistic Geometry ................................ ................................ ............. 26 4.1.1 Trajectorie s ................................ ................................ ................................ ................ 26 4.1.2 Introduction to the Central Dogma ................................ ................................ ............ 28 4.2.0 DNA and RNA ................................ ................................ ................................ ............... 29 4.3.0 The Genetic Code ................................ ................................ ................................ .......... 30 4.4.0 Protein Synthesis ................................ ................................ ................................ ............ 31 V. REVISITING THE ALGORITHM OF ADMISSIBLE TRAJECTORIES ............................. 33 5.1.0 Overview ................................ ................................ ................................ ........................ 33 5.2.0 Motivation ................................ ................................ ................................ ...................... 34 5.3.0 FindMid Experiment ................................ ................................ ................................ ...... 34 5.3.1 The Observation Stream ................................ ................................ ............................ 34 5.3.2 The Construction Stream and the Expression Stream ................................ ............... 36 5.4.0 Define the Mosaic's Aggregates and Matching Rules ................................ ................... 38 5.5.0 GlueTr ee Experiment ................................ ................................ ................................ ..... 40 5.5.1 The Observation Stream ................................ ................................ ............................ 40 5.5.2 The Construction Stream ................................ ................................ ........................... 40 5.6.0 Results ................................ ................................ ................................ ............................ 45 VI. REVISITING THE DISCOVERY OF GENET IC CODE ................................ ..................... 46 6.1.0 Overview ................................ ................................ ................................ ........................ 46 6.2.0 Motivation ................................ ................................ ................................ ...................... 47 6.3.0 The Gamow Thought Experiments (The Diamond Code) ................................ ............. 47

PAGE 10

! x 6.3.1 Building the Input Mosa ic ................................ ................................ ......................... 48 6.3.2 Generating the "Construct Input Mosaic" Shell ................................ ........................ 51 6.3.3 Identifying the Output Mosaic ................................ ................................ .................. 53 6.3.4 Generating the "Construct Output Mosaic" Shell ................................ ..................... 54 6.3.5 Defining the Intermediate Mosaic ................................ ................................ ............. 55 6.3.6 Assigning the Reading Rules ................................ ................................ .................... 57 6.3.7 Identifying the Matching Rules ................................ ................................ ................. 60 6.3.8 Defining the Code and Amino Sets ................................ ................................ ........... 61 6.3.9 Constructing the Intermediate Mosaic ................................ ................................ ...... 63 6.4.0 The Crick Thought Experiments ................................ ................................ .................... 65 6.4.1 Validating the Diamond Code ................................ ................................ ................... 65 6.4.2 Redefining the Input Mosaic ................................ ................................ ..................... 67 6.4.3 Evaluating the Reading Rules ................................ ................................ ................... 69 6.4.4 Rebuilding the Amino Set ................................ ................................ ......................... 71 6.5.0 Modifying the Reading Rules ................................ ................................ ........................ 72 6.6.0 Deciphering the First Triple ................................ ................................ ........................... 73 6.6.1 Defining the Input Mosaic ................................ ................................ ......................... 74 6.6.2 Generating the "Construct information mRNA" Shell ................................ .............. 75 6.6.3 Translation Compo nents ................................ ................................ ............................ 76 6.6.4 The Mapping Thought Experiments ................................ ................................ .......... 77 6.7.0 Results ................................ ................................ ................................ ............................ 80 VII. REVISITING THE DISCOVERY OF PROTEIN TRANSLATION ................................ ... 82 7.1.0 Overvie w ................................ ................................ ................................ ........................ 82

PAGE 11

! xi 7.2.0 Motivation ................................ ................................ ................................ ...................... 83 7.3.0 Applying the AD on the Discovery of tRNA Activation ................................ ............... 83 7.3.1 The Adapter Hypothesis Thought Experiments ................................ ........................ 84 7.3.2 Identifying the Process of Activating Amino Acids ................................ ................. 85 7.3.3 Identifying the Process Rules to Link Amino Acids to tRNA ................................ .. 95 7.4.0 Finding the Rules to Attach tRNA and mRNA ................................ ............................. 99 7.4.1 Defining Matching rules to link mRNA and tRNA Mosaics ................................ .. 100 7.4.2 Discovering Matching Rules ................................ ................................ ................... 101 7.5.0 The Watson Model ................................ ................................ ................................ ....... 111 7.5.1 Defining the Struct ure of the Ribosome ................................ ................................ .. 117 7.5.2 Identify the Matching Rules between the mRNA and the Ribosome ..................... 118 7.5.3 Finding the Plugging Sites for tRNA into the Ribosome Mosaic ........................... 119 7.5.4 Constructing the Mechanism of Elongation ................................ ............................ 120 7.6.0 The Watson extension Model ................................ ................................ ...................... 128 7.6.1 Validating the Nierhaus model ................................ ................................ ................ 131 7.6.2 Declaring the Matching Rules fo r Occupying the E Site ................................ ........ 133 7.6.3 Reconstructing the Translation Process ................................ ................................ .. 136 7.7.0 The Alternative Model: Rebuilding the Elongation Model ................................ ......... 143 7.7.1 Preparing the Const ruction Set for Elongation: the Observation Stream ................ 146 7.7.2 Reconstructing Elongation ................................ ................................ ...................... 148 7.7.3 Verifying the New Three site Model: The Validation Stream ................................ 156 7.8.0 T he Hybrid Model of Elongation ................................ ................................ ................. 157 7.8.1 Declaring the First Process Rule: The Observation Stream ................................ ......... 159

PAGE 12

! xii 7.8.2 The Observation Stream: Declaring the Second Process Rule ................................ .... 161 7.8.3 The Observation Stream: Declaring More Process Rules ................................ ............ 163 7.8.4 Building the Model: The Construction Stream ................................ ............................ 165 7.9.0 Results ................................ ................................ ................................ .......................... 173 VIII. CONTRIBUTIONS AND FUTU RE WORK ................................ ................................ .... 178 8.1.0 Overview ................................ ................................ ................................ ...................... 178 8.2.0 Contributions ................................ ................................ ................................ ................ 178 8.2.1 New Matching Rules ................................ ................................ ............................... 178 8.2.2 Investigation of New Streams ................................ ................................ ................. 181 8.2.3 New Types of Mosaics ................................ ................................ ............................ 181 8.2.4 New Transformation Phases of Protein Synthesis ................................ .................. 184 8.2.5 New Scope Sets ................................ ................................ ................................ ....... 185 8.2.6 Identifying a New Stage of the Observation Stream ................................ ............... 185 8.2.7 Identifying New Thought Experiments ................................ ................................ ... 186 8.3.0 Towards Automation of the AD ................................ ................................ .................. 188 8.4.0 Future Work ................................ ................................ ................................ ................. 191 BIBLI OGRAPHY ................................ ................................ ................................ ....................... 193

PAGE 13

! xiii LIST OF TABLES TABLE 1. Different types of internal streams. ................................ ................................ ........................... 14 2. Several instruction aggregates introduced during this research. ................................ ............... 23 3. One dimensional letter tiles mosaic based on the DNA base tiles in Figure 9. ........................ 51 4. The code mosaic generated by the expression stream as Shown in [42], [7]. .......................... 63 5. One dimensional mosaic of the mRNA letter tiles converted from the DNA base tiles. ......... 75 6. The final genetic code mosaic, [37]. ................................ ................................ ......................... 80 7. The RNA complementarity matching mosaic between the first two tiles in the mRNA and tRNA. ................................ ................................ ................................ ................................ ... 103 8. The wobble complementarity matching mosaic between th e third tile in the mRNA and tR 103 9. The updated wobble complementarity matching mosaic. ................................ ....................... 106 10. The update d wobble complementarity matching mosaic. ................................ ..................... 107 11. The final wobble complementarity matching mosaic. ................................ .......................... 110

PAGE 14

! xiv LIST OF FIGURES FIGURE 1. The communication streams pass information betw een the internal streams and the outer world and vice versa. ................................ ................................ ................................ ....................... 21 2. Two different trajectories between points x i and x j . ................................ ................................ . 27 3. Visual representation of the initial input model. ................................ ................................ ....... 36 4. The visual model after being morphed by the construction stream. ................................ ......... 37 5. Visual simulation of the GlueTree thought experiment. ................................ ........................... 41 6. Building a Mid aggregate from the To Mid and From Mid aggregates. ................................ .. 42 7. The visual representation of the mosaic of the tree of the admissible trajectories. .................. 44 8. The final algorithmic mosaic corresponded to the construction of the admissible tra jectories tree. ................................ ................................ ................................ ................................ ........ 44 9. The schematic representation of the double helix, [10]. ................................ ........................... 49 10. Part of the input mosaic as shown in [64]. ................................ ................................ .............. 50 11. "Construct Input mosaic" shell generated by the expression stream. ................................ ..... 53 12. "Construct Output mosaic" shell generated by the expression stream. ................................ .. 55 13. The code mosaic sits between the cavities of the input mosaic as shown in [42] , [7]. ........... 57 14. Four symmetrical code aggregates could be generated from the genetic code 123 by the construction stream, [7]. ................................ ................................ ................................ ........ 58 15. Part of the intermediate mosaic constructed form the genetic seq uence "123123", [7]. ........ 65 16. The code aggregates generated from the input sequence 1112n. ................................ ............ 67 17. Constructing the RNA mosaic from a single strand of the DNA mosaic. .............................. 69 18. "Construct information mRNA" shell generated by the expression stream. .......................... 76 19. An mRNA mosaic of Poly U code aggregates, [7]. ................................ ................................ 78 20. The mapping between the code and amino aggregates, [7]. ................................ ................... 79 21. The input aggregate: Acetate. ................................ ................................ ................................ . 87

PAGE 15

! xv 22. The input aggregate: Acetyl CoA. ................................ ................................ .......................... 87 23. CO NH peptide bond in peptide mosaic retrieved from a previous thought experiment. ...... 88 24. General representation of amino aggregates. ................................ ................................ .......... 90 25. The ATP aggregate. ................................ ................................ ................................ ................ 90 26. First scenario CO binds with Ad.P. ................................ ................................ ........................ 90 27. Second scena rio CO binds with Ad.P~P. ................................ ................................ ................ 91 28. Third scenario CO binds with P~P. ................................ ................................ ........................ 91 29. Fourth scenario CO binds with P. ................................ ................................ ........................... 91 30. The shell corresponded to constructing the ATP aggre gate. ................................ .................. 92 31. The shell corresponded to constructing the amino aggregate. ................................ ................ 93 32. The shell generated to construct the Active amino aggregate. ................................ ............... 93 33. Active amino aggre gate carried by the enzyme backbone as visualized by the construction stream. ................................ ................................ ................................ ................................ .... 94 34. The final shell to construct the active amino aggregate. ................................ ......................... 95 35. The first step of the translation process: cons tructing the amino tRNA. ................................ 98 36. "Construct amino tRNA" shell. ................................ ................................ .............................. 99 37. The third tile's matching rule is unknown, [6]. ................................ ................................ .... 102 38. The construction stream failed to build t he G A aggregate, [6]. ................................ .......... 105 39. The construction stream succeeded in building the U C aggregate, [6]. .............................. 106 40. The G U aggregate, [6]. ................................ ................................ ................................ ........ 107 41. The I A aggregate, [6]. ................................ ................................ ................................ ......... 107 42. The physical position of the bonds for the pair U C. ................................ ............................ 108 43. The physical position of the bonds violates the physical rule: too close bonds. .................. 109 44. The physi cal position of the bonds obeys the physical rule. ................................ ................. 110 45. The complementarity rules that control the binding of mRNA and tRNA, [6]. ................... 110 46. The tRNA mosaic before folding as represented by the co nstruction stream. ...................... 112

PAGE 16

! xvi 47. The tRNA mosaic after folding as represented by the construction stream. ........................ 113 48. "Construct tRNA body" shell. ................................ ................................ .............................. 113 49. The modified "Constr uct tRNA body" shell. ................................ ................................ ....... 114 50. Three types of tRNA mosaics built by the construction stream: (a) amino tRNA, (b) peptide tRNA, and (c) deacylated tRNA. ................................ ................................ ........................ 114 51. "Construct amino tRNA" shell. ................................ ................................ ............................ 115 52. "Construct peptide tRNA" shell. ................................ ................................ .......................... 115 53. Constructing the mRNA mosaic. ................................ ................................ .......................... 116 54. The translation process as defined by previous thought experiments, [46]. ......................... 117 55. A general shape of the 50 s and 30 s mosaics. ................................ ................................ ..... 121 56. Constructing the ribosome mosaic. ................................ ................................ ....................... 121 57. The shell corresponded to the construction of the ribosome mosaic. ................................ ... 1 22 58. The mRNA enters the ribosome mosaic. ................................ ................................ .............. 122 59. The starting state in the mechanism of elongation identified by the construction stream. ... 123 60. The amino tRNA enters in to the ribosomal A site. ................................ .............................. 124 61. The amino tRNA occupies the A site. ................................ ................................ .................. 125 62. The peptide formation. ................................ ................................ ................................ .......... 125 63. The tRNA leaves the ribosome. ................................ ................................ ............................ 126 64. The peptide tRNA translocates into the P site. ................................ ................................ ..... 126 65. The next amino tRNA enters the ribosome. ................................ ................................ ......... 127 66. The "Elongation" shell for the Watson model. ................................ ................................ ..... 127 67. The mRNA mosaic. ................................ ................................ ................................ .............. 129 68. Three Different Types of tRNA Mosaics ................................ ................................ .............. 129 69. The modified "Construct tRNA body" shell. ................................ ................................ ....... 130 70. The ribosome mosaic built by the constructi on stream and tagged by the expression stream. ................................ ................................ ................................ ................................ ............. 135

PAGE 17

! xvii 71. The shell corresponded to the construction of the ribosome mosaic. ................................ ... 135 72. The mRNA mosaic takes a U turn shape inside the ribosome. ................................ ............ 136 73. The "Link the ribosome to the mRNA" shell generated by the expression stream. ............. 136 74. Different snapshots of the initiation process by the construction stream: (1) The current codon is not an init iation codon the construction stream will read the next codon, (2) The current codon is an initiation codon, (2) call the initiator tRNA the fMet tRNA, and (4) place the fMet tRNA in the P site and create the codon anticodon aggregate. .................. 138 75. The "Initiate" shell corresponded to the process in Figure 74. ................................ ............. 138 76. One cycle of the elongation process as visualized by the construction stream. ................... 140 77. The "Elongation" shel l corresponded to the process in Figure 76. ................................ ....... 140 78. The first scenario for a new version of the "Elongation" shell. ................................ ............ 141 79. The second scenario for a new version of the "Elongation" shell. ................................ ....... 142 80. Reading a termination codon. ................................ ................................ ............................... 142 81. Terminating protein synthesis. ................................ ................................ .............................. 143 82. The "Termination" shell. ................................ ................................ ................................ ...... 143 83. The new structure of the r ibosome mosaic. ................................ ................................ .......... 149 84. The "Construct ribosome" shell generated by the expression stream. ................................ .. 149 85. Three different types of tRNA mosaics as visualized by the construction stream. .............. 150 86. The mRNA enters the ribosome. ................................ ................................ .......................... 151 87. The starting cycle of elongation. ................................ ................................ ........................... 152 88. The amino tRNA is ready to accept the peptide mosaic. ................................ ...................... 153 89. Pep tidyltransferase. ................................ ................................ ................................ ............... 153 90. The tRNA mosaics translocation. ................................ ................................ ......................... 154 91. The new amino tRNA is ready to enter the elongation cycle and the deacylated tRNA is ready to move out of the cycle. ................................ ................................ ............................ 154 92. The beginning of a new elongation cycle. ................................ ................................ ............ 155

PAGE 18

! xviii 93. The algorithmic mosaic for the alternative elongation model built by the expression stream. . 155 94. The three classical st ates the tRNA mosaics can move to during elongation. ...................... 160 95. The P/E state where deacylated tRNA occupied the P site of the 30 s mosaic and the E site of the 50 s mosaic at the same time. ................................ ................................ .................... 161 96. The A/E state where peptide tRNA occupied the A site of the 30 s mosaic and the P site of the 50 s mosaic simultaneously. ................................ ................................ .......................... 162 97. The ribosome mosaic built by the construction stream. ................................ ....................... 166 98. The initial scene to start elongation. ................................ ................................ ..................... 166 99. The EF Tu GTP amino tRNA is in the A/T state. ................................ ............................... 167 100. The algorithmic mosaic explaining the process in Figure 99. ................................ ............ 167 Figu re 101. The amino tRNA moves from the A/T state to the A/A state. ................................ 168 102. The elongation shell after translating the operations from Figure 101. .............................. 169 103. The peptide mosaic links to the amino agg regates and the deacylated tRNA and the peptide tRNA moves to the states P/E and A/P respectively. ................................ .......................... 169 104. The expression stream continues building the elongation shell by watching the visual movie run by the construction str eam. ................................ ................................ ............................ 170 105. The deacylated tRNA and the peptide tRNA move completely to the E site and the P site, respectively. ................................ ................................ ................................ ......................... 171 106. "Elongation" Shell ................................ ................................ ................................ .............. 171 107. The previous d eacylated tRNA must leave the ribosome before the new deacylated tRNA can enter the E site. ................................ ................................ ................................ ............. 172 108. The final "Elongation" shell. ................................ ................................ .............................. 172 109. The final algorithmic mosaic for the translation process . ................................ ................... 177 110. The stages of protein synthesis. ................................ ................................ .......................... 184 111. The components of the AD utilized in each case study. ................................ ..................... 188

PAGE 19

! 1 CHAPTER I INTRODUCTION 1.1.0 Overview In the late 1950's, P rofessor John von Neumann [47] proposed the existence of an internal language used by the human brain for mental calculation. He named the brain's language the Primary Language (PL). He assumed that the PL differs from other ex ternal languages humans use for communication, which he defined as secondary languages. The secondary languages include the natural languages and the languages of science, such as mathematics, computer science, etc. Von Neumann suggested that the PL is use d for thought. At the beginning of his research into the PL, Professor Stilman's originally hypothesized that the PL included all major algorithms essential for humanity's development and survival. During his investigations, Stilman discovered two algorith ms that are critical to the development of human intelligence and revealed their relationship to the PL. They are Linguistic Geometry (LG), the algorithm for optimizing warfighting and the Algorithm of Discovery (AD), the algorithm for inventing new algori thms, [27] , and [16] . Stilman concluded that the AD is based on multiple thought experiments, Section 2.3.0 , which manifest themselves via visual streams (mental movies) . It appears that, in the human brain, visual streams are the only interface to the AD. The AD operates with three classes of visual streams: observation, construction, and validation. Another class of visual streams, the communication streams, along with the AD, runs to translate the work of the AD's streams. All classes of visual streams can run concurrently and exchange information. Each stream may initiate additional thought experiments, program , and then execute them, [27] .

PAGE 20

! 2 After extensive investigation, Stilman revised his original hypothesis, [27] , and [16] . The PL is not a collection of major algorithms; it is the Language of Visual Streams (LVS). As predicted by v on Neumann, the PL is the base for all external languages. Stilman assumed that the PL is the engine that generates visual streams. This research is based on the main hypothesis that the AD is used for inventing, or discovering, all other algorithms. The AD is based directly on the PL, the Language of Visual Streams. The visual streams are used by the AD to construct new algorithms and, in this way, make discoveries. 1.2.0 Method of Study The method of study can be divided into two areas: case studies and softwa re implementation. To reveal the nature of the AD, it was applied (as currently understood) to many case studies from two classes of discoveries: the field of LG, focusing on the admissible trajectories algorithm, and the field of Molecular Biology/Biochem istry, focusing on the discoveries of the genetic code and the protein synthesis. By applying the AD and its components, the various stages the AD followed on its way to making these discoveries were defined . When the AD uncovers all stages of any innovati on, the expression stream, a subset of the communication streams, generates the algorithm that guides the discovery. Then, the generated algorithm can be implemented as a computer program. The basic assumption is that if this program finds the correct solu tion to a class of diverse problems, then the discovered algorithm is correct. Eventually , the goal is to automate future discoveries, i.e., produce discoveries on demand.

PAGE 21

! 3 1.3.0 Motivation Every discovery starts with an idea, which motivates the discoverer. The spark, that motivated the work in this field, was the idea of having an algorithm that generates other algorithms , i.e., the algorithm that make s discoveries. The main question to be answered is "How, by following the work of others, can the entire nature of the AD be reveal ed and its components be found?" It was possible to answer this question, after reading the works of Stilman a nd various other papers related to this area of study . If computer programs could produce discoveries routinely as an output, a huge leap would be made by the humanity. Achieving the goal of making discoveries on demand would be the final outcome of this work . This research started by working in the LG field, finding out that the AD led to the discovery of the algorithm for genera ting admissible trajectories. During this research, the AD is assumed to be universal, which means that any discovery will happen if the correct inputs, rules , and procedures are identified. Following this approach, it was found that the major innovations were made in stages employing various thought experiments and visual streams. During the thought experiments, the visual streams' engine may drop one algorithm and introduce another one according to the analysis phase. However, the AD that controls those s treams is the same; it makes the discovery if the inputs and analysis are correct. Accurate analysis will generate new algorithms, which are the essence of any discovery. 1.4.0 Objectives The objective of this research is to apply the AD to various discoveries t o reveal its inherent details. In so doing, several components of the AD were investigated. In all of the case studies in this research, the AD was applied to two different fields of science, LG and

PAGE 22

! 4 Molecular Biology/Biochemistry. Once the elements of the AD in any discovery are revealed, those elements can be generalized. Then, it is possible to implement and simulate the algorithm for that discovery. By executing the implementation of the algorithm, the AD's reasoning components can be further analyzed an d verified. In every case study, usually there are short and long term goals. The short term goal is to better understand the AD by deeply studying its applications , while applying it to selected case studies. The long term goal is to automate future disc overies, i.e., develop a program for making them on demand. 1.5.0 Work Domain The research's work domain is the PL; the PL domain includes investigating all the different types of visual streams employed by the AD to make any discovery. As mentioned in Section 1.2.0 , the case studies are from two different fields of science, the field of LG and the field of Molecular Biology/Biochemistry. When employing the AD, it is required to revisit all attempts to further refine t he algorithm responsible for a specific discovery, including any known failed attempts. The failed attempts reveal as much information about the nature of the AD as do the successful ones. The AD analyzes the reason(s) for its unsuccessful execution, then it either returns to a specific waypoint or re initializes itself to try to find new inputs and procedures guiding the construction in a different direction in an attempt to make the correct discovery. During the analysis phase, the most difficult step was finding a clear description of all attempts made by the discoverers. For that reason, only discoveries with clear descriptions were considered as candidates for the AD's application and selected as case studies for this research. The first case study is t he application of the AD to the discovery of the algorithm

PAGE 23

! 5 for generating admissible trajectories in LG. Unfortunately, not all experiments were recorded, but by discussing the se discoveries with Stilman it was possible to apply the AD and obtain good resu lts. The second case study is the discovery of the genetic code. Many scientists were involved in cracking the genetic code. Multiple written descriptions of failed experiments to decipher the code were found and studied. After finding a clear description of several recorded trials, the process of discovery was broken into smaller problems. The AD was applied to those smaller problems to reveal the components of the algorithm that guided the discovery of deciphering the genetic code. The third case study i s the discovery of the algorithm of translation. Again, multiple scientists participated in revealing the process of translation, the last step of protein synthesis. Only major experiments were examined in an effort to discover the algorithm used in protei n synthesis. The descriptions of several trials were followed, including failed ones, and the problem was broken into several smaller problems. By applying the AD to these smaller problems, the components of the AD that led to the discovery of the algorith m of translation were successfully revealed. 1.6.0 Challenges During this research, one goal was to study the PL by applying the AD to a non computer science field. While analyzing the case studies in the Molecular Biology and Biochemistry fields, the main conce pts of these fields had to be studied to better understand the reasoning that led to the conclusions of those case studies. Also, due to the nature of the discoveries in this field, many papers with different conclusions were found. Therefore, it was impor tant to select only those papers that eventually led to the discovery of the algorithm of translation.

PAGE 24

! 6 The most important part, of this area of research, was to determine which papers were instrumental in following the work of the AD. Otherwise, there was a risk of analyzing many papers only to discover those papers did not add any value to this research and , thus , wasted time. After applying the AD on a series of thought experiments and revealing more components of the AD or finding a pattern of visual st reams, the newly discovered components had to be applied to previous work. First , it was necessary to understand the nature of the new components, and second , to see if the new components were applicable to all discoveries or associated with discoveries of a special nature. This recursion significantly increased the research time because the AD had to be reapplied to previous work every time a new component of the AD was discovered. This , in turn, could lead to new findings, which must also be applied to p revious work. 1.7.0 Contribution While applying the AD to the genetic code and translation discoveries, new components of the AD were revealed. These components were found to be applicable to any similar discovery. New transformation rules that played a major r ole in revealing the genetic code and the translation algorithms were introduced. Movement rules were identified; these are part of the transformation rules guiding the movement of one mosaic (constructed object ) , while copying another to create a new mosa ic. Another subset of the transformation rules are the reading rules, which identify the reading of information from one mosaic to construct another mosaic, based on that reading. Along with the movements rules come the alignment rules that define how to a lign two mosaics , if necessary.

PAGE 25

! 7 The process rules are another subset of the transformation rules and appeared initially during the analysis of the translation discovery. The process rules must be identified in any discovery re quiring multiple mosaic actio ns, whether in sequence or in parallel. Such rules are responsible for creating the order of the actions for every mosaic inside the thought experiment. The identification of the correct process rules is the basis of defining the correct algorithm for any discovery. While working on the protein synthesis problem, two transformation phases were identified: transcription and translation. The transcription phase transforms the DNA mosaic into the RNA mosaic. The translation phase transforms an RNA mosaic into an amino acid mosaic. Multiple scope sets were declared while analyzing many thought experiments. The observation stream defines scope sets to focus the run of visual streams while working on any problem and reducing the time needed to reach a solution. P hysical rules were identified while working on the translation problem. The physical rules are part of the environmental rules that represent the laws of the real world environment. Also, a new type of visual stream, the communication stream , was studied a nd applied to all case studies within this research. An expression stream is a subset of the communication streams , and is responsible for generating the communication thought experiments. Identified for the first time, these streams translate the interna l thought experiments into a secondary language understood by the outer world. The PL applies the internal streams to develop the internal thought experiments. The different types of visual streams are explained in Section 2.4.0 .

PAGE 26

! 8 Several n ew types of data storage mosaics were identified. The first type is the information mosaic (one dimensional and two dimensional). Information mosaics are created via a sequence of transformation phases by the internal streams. Other new types of mosaics are the algorithmic and symbolic mosaics that include information to be passed to the outer world. The algorithmic and symbolic mosaics are constructed via the communication thought experiments by the expression stream. 1.8.0 Dissertation Structure T he chapters of this dissertation are organized as follows: • Chapter II : Review of the literature discussing the PL, thought experiments, the AD, including a brief description of all its streams and components, and LG and its components. • Chapter III : Explan ation of the communication streams and the algorithmic mosaics. • Chapter IV : Review of the literature discussing the discover ies of the admissible trajectories, the genetic c ode , and the protein synthesis . • Chapter V : The first case study is the application of the AD to the rediscovery of the algorithm for generating admissible trajectories in LG. • Chapter VI : The second case study is the application of the AD to the rediscovery of the genetic code. • Chapter VII : The third case study is the application of the AD to the rediscovery of the translation algorithm. • Chapters V , VI , and VII contain the essential parts of this research. In every case study, the original descriptions of the discoveries are analyzed, the AD is

PAGE 27

! 9 applied, and the components of the AD guidi ng those discoveries are presented . • Chapter VIII : Contributions and Future Work. It includes contributions of this research in the fields of Artificial Intelligence and Computer Science.

PAGE 28

! 10 CHAPTER II THE PRIMARY LANGUAGE 2.1.0 Overview This chapter includes a P L literature review. As mentioned in Section 1.1.0 , von Neumann introduced the PL as the language the human brain uses for mental calculation. While investigating PL, two of its major algorithms essential for the development of humanity were revealed, LG and the AD. According to the hypothesis of this research, the PL is the Language of Visual Streams (LVS) that manifests itself via various means including invention of new algorithms. 2.2.0 The Primary Language (PL) The two types of languages that the human brain uses were suggested by von Neumann in 1957, [47] . He introduced the concept that the PL is used by the brain for computation and mental calculation . Von Neumann suggested that the PL was develop ed "beneath" the secondary languages, i.e. , the external languages . When von Neumann declared the existence of the PL, he argued that the nature of the PL was unknown. He writes, "It is only proper to realize that [human] language is largely a historical a ccident. The basic human languages are traditionally transmitted to us in various forms, but their very multiplicity proves that there is nothing absolute and necessary about them. Just as languages like Greek or Sanskrit are historical facts and not absol ute logical necessities, it is only reasonable to assume that logic and mathematics are similarly historical, accidental forms of expression. They may have essential variants, i.e., they may exist in other forms than the ones to which we are accustomed É T he language here involved may well correspond to a short code in the sense described earlier, rather than to a complete code: when we talk mathematics, we may be

PAGE 29

! 11 discussing a secondary language, built on the primary language truly used by the central nervo us system [47] . " For more than 60 years, the nature of the PL was still unknown. Following von Neumann's hypothesis, the investigation of the PL started, [27] [6] , with the assumption that the PL is the basis of all algorithms crucial for the development of human intelligence. According to Stilman [16] , the PL appeared long before the time when the external languages were developed. During ou r investigation, it was assumed that the PL generates mental movies called visual streams that interact with each other. It was suggested that the PL is the Langua ge of Visual Streams . Visual streams were introduced as imaginary animated movies rather than sets of strings of symbols (called languages in Mathematics and Computer Science) . In particular, visual streams are generated by the PL to find solutions for many problems by focusing them in proper direction, and reasoning about them. The PL uses those visual streams to interact with each other . The PL is hypothesized to be the language utilized for the development of all algorithms discovered by humans; it is the foundation for all external , i.e., spoken languages. 2.3.0 Visual Streams and Thought Experiments Thought experiments are the methods created by an investigator's imagination to visualize several scenarios to analyze a problem , "solve it," and "see" the results of different solutions. Thought experiments provide visual reasoning about the laws of nat ure in different fields. Hobbes and Locke introduced the state of nature using philosophical thought experiments, [52] . Also, thought experiments can be used to examine other hypotheses without actually testing these hypothese s. For example, Galileo created a thought experiment to refute Aristotle's theory about the free fall of bodies. Without dropping a single object, Galileo concluded that bodies with unequal mass fall at the same speed, [60] , an d [1] .

PAGE 30

! 12 The possibility of thought experiments shows that the human brain can mix a variety of animated models reflecting reality and its natural (or artificial) laws. By studying the brain's cortex, scientists have successfully detected some of those mental images. They found that human mind converts any shape or symbol into a mental image. Such images have an analogical nature because, according to the studies, the visual areas of the brain sketch the contours of such imagined objects , [60] . It appears that for any species to survive, it is very important for said species to understand the laws of nature. Humans ca n represent those laws in their brain s by simulating them in the form of mental images . Neuropsychologists found that people carry with them mental universes that simulate the laws of the world around them, [60] . Moreover, other studies proposed that humans also carry the laws of major human relations, including the laws of warfighting. Those laws manifest themselves in several ways; for example, the sensorimotor system of the human brain can understand kinematics when predicting an object's trajectory. Humans use those laws routinely and sub consciously while vis ualizing an object in a mental movie or predicting its trajectory on a map, [1] , and [61] . The visual streams may simulate different artificial worlds; each of which is ruled by artificial laws of n ature and populated with animated entities, which represent real or simulated objects. The visual streams are constructed and run by the human brain to observe realistic or artificial events in those mentally constructed worlds. According to this research' s hypothesis, every human invention is made by employing visual streams , [27 ] . By observing the effects of the physical laws in the environment around it, the brain simulates them and constructs the animated mental events accor dingly. The brain has the power to alter the mental worlds by redefining their components according to the problem

PAGE 31

! 13 statement. In this research, a visual stream is assumed to be a mental movie that is played in the human brain to show, at its conclusion, a solution of a staged problem. Stilman proposed that it is possible to model and eventually implement such mental worlds and visual streams by employing conventional algorithms. Stilman further proposed that the PL is the LVS, [8] , [17] , [9] , and [28] . Visual streams can trigger additional thought experiments and schedule them to be run by other visual streams. The objective, structure, and out come of the created thought experiments should become known to the triggered visual stream. Moreover, the visual streams that schedule additional thought experiments (to be exe cuted later) are usually are not aware of the algorithms to be executed in those experiments. The structure of these algorithms will be developed and, thus, will become known during the runs of the associated thought experiments, [23] , and [10] . 2.4.0 Types of Visual Streams Visual s treams can be broken into groups according to their application. Most common , there are visual streams that interact internally only with other visual streams. They have no interaction with the outer world. This type of visual streams is called the intern al stream . A different type of visual stream that interact with the outer world is the communication stream , [27] . The visual streams can be classified by the information the streams process. Mundane streams deal with ordinary information while scientific streams deal with scientific information. Moreover, the visual streams used by the AD can be broken into three groups based on their function: observation, construction, and validation streams. Visual streams

PAGE 32

! 14 may also be clas sified based on the type of reasoning they utilize. So far, two major types of reasoning have been revealed: proximity reasoning and mosaic reasoning. Additionally, visual streams can be classified according to their programming structures (parallel, seq uential, or nested). Parallel streams run concurrently, while sequential streams run consecutively. A typical example of sequential streams is how they are utilized by the AD: observation to construction to validation streams. The nested streams call each other in no specific order. Yet another type of classification is related to the stream's common theme, which unifies a group of streams with a common set of constraints that limit possible directions of the streams' morphing. Table 1 represents the different types of the identified internal streams . Table 1 . Different t ypes of i nternal s treams . Classification Criteria The AD Streams Information Reasoning Programming Structure Common Theme Observation, co nstruction, and validation Mundane and scientific Proximity and mosaic Parallel, sequential, and nested Any type of internal streams has similar set of constrains 2.4.1 Discovery Streams and the Algorithm of Discovery (AD) In this research, it is assumed that the AD is the algorithm for inventing new algorithms; in other words, it is the algorithm that makes discoveries. Based on multiple thought experiments involved in the discoveries in LG and Molecular Biology/Biochemistry, in [16] through [59] , it is argued that the AD does not utilize a tree based search to solve a problem . Instead, it is theorized that the PL generates visual streams to build a mental movie, that is played in the human brain, and wh ich concludes with a solution to the staged problem. The

PAGE 33

! 15 PL will use those visual streams as inputs to the AD. An input visual stream is usually built from multiple imaginary instances of the object under investigation. The first visual streams the AD rece ives are usually a visual replication of natural or imaginary entities. The AD then runs this visual stream, several times, to understand these objects, find their structure, and the rules of their construction, [16] . Stilman categorized visual streams based on their nature or the work for which they are responsible, [27] . He proposed that streams could be divided into groups based on their purpose. He identified three classes of discovery visual st reams the AD can execute: observation, construction and validation, [12] . These discovery streams can be sequential, parallel or nested. At the beginning of every AD run, the AD first initiates the observation stream. In this r esearch, it is proposed that the observation stream has two stages, the preliminary stage, and the analysis stage. The main job of the preliminary observation stream is to collect the required inputs from a stream of inputs including all known data about t he problem. The analysis stage of the observation stream requires it to investigate and build at least two tools: the construction set and the visual model, usually by "erasing the particulars." , [24] . Erasing the particulars m eans to remove any unnecessary specifics of the objects inside the visual stream by executing the visual stream several times in attempt to reduce multiple instances of the object to the one or two simple instances , considering the outputs from previous th ought experiments as well as incoming inputs from the impression stream . By doing so, t he observation stream modifies objects on hand to "observe" their real structure and "find" the rules that control it. Usually, in order to discover the structure of an object or a process, the observation stream has to construct it. Once the observation stream reveals the

PAGE 34

! 16 rules that direct the construction of an object , it builds a visual prototype of this object and saves the rules in the construction set. The observati on stream then triggers another visual stream, the construction stream, to start building the object, or staging the process, and, eventually, reveals the algorithm that guided the construction. The validation stream tests the correctness of the construct ed object, or staged process, and compares it to the visual model identified by the observation stream. The validation stream also examines the rules for building the object, or the staging process, by comparing them to any new data coming from the outer w orld. If a contradiction is found, the validation stream will reject the results of that construction and call the observation stream to rebuild the construction set or modify the visual model. In such cases, the construction stream must be called again to reconstruct the object or restage the process. These multiple executions are needed for the AD to correctly solve the problem, [16] . In many cases, the AD may utilize a virtual moving entity, the Ghost, to move in the artifici al world to investigate other objects in this artificial world. The observation stream employs the Ghost to "walk" and "observe" the structure of an object or to find the rules that control objects in the artificial world. The construction stream can also use the Ghost to build an object by assembling its parts and placing these parts in their correct positions. The validation stream triggers the Ghost to "scan" through the artificial world and "test" the rules and objects identified by previous visual stre ams in order to "find" any contradictions that might invalidate these rules or objects, [16] . 2.5.0 Proximity and Mosaic reasoning The AD sometimes needs to apply special types of reasoning to find solutions for certain categories of problems. Some problems require optimization components. In this case, the

PAGE 35

! 17 AD tries to find the solution by running a sequence of thought experiments that run via proximity reasoning, [23] . In many cases, to find a solution, the AD has to run visual streams that have the ability to assemble objects of different shapes and types, i.e., run under the influence of mosaic reasoning , [16] . In [10] , Stilman introduced mosaic reasoning to describe the reasoning that helps visual streams build a mosaic picture of any object out of small colorful tiles. The major components of mosaic reasoning, re vealed to date, are tiles, aggregates, and matching rules. A tile is the smallest unit of the mosaic. In order to construct any object correctly, tiles must be accu rately inserted in the mosaic. An i ncorrect insertion of any tile will ruin the final struct ure of the mosaic leading to an incorrect solution of the problem. Two or more tiles linked together form one aggregate. In general, mosaic reasoning runs via multiple visual streams using tiles and aggregates. The observation stream starts mosaic reasoni ng by identifying the tiles and aggregates that will be used to build the final mosaic. This identification process might take several executions and morphing of the visual streams. Once the tiles and aggregates are identified, the observation stream must investigate the matching rules. The matching rules determine how tiles and aggregates link together. For the mosaic of processes, those rules control how different objects move within the artificial world. These rules can affect the construction of any mo saic locally or globally. The global matching rules are the global complementarity rule, the environmental rules, and the transformation rules, while the local matching rules are the local complementarity rule and the interchangeability rule.

PAGE 36

! 18 Global comple mentarity rules control the insertion of one aggregate into the mosaic according to the placement of its adjacent aggregate. Further, visual streams can use global transformation rules to construct new mosaics from one or more existing mosaics. Environment al rules represent all the rules of the real enviro nment. The environmental rules must be obeyed while building any mosaic. These rules include physical constraints, chemical rules, etc. Transformation rules include the movement rules that guide the stre ams to move one or more mosaics. As a result of this movement, a new mosaic is generated. The process rules are also part of the transformation rules. They identify the order of the mosaic's actions within the thought experiment. For example, if there exi sts three different mosaics that must interact together to build another mosaic, the process rules control which mosaic moves first , while the movement rules guide the motion of a mosaic. On the other hand, local complementarity rules control the placemen t of a tile based on the adjacent tiles . Interchangeability rules allow visual streams to remove one aggregate from the mosaic and replace it with an interchangeable one that has the same general structure. Interchangeable aggregates that share the same f eatures are called plug ins. This interchangeability changes the final picture of the mosaic , but the correctness of the final structure of the mosaic will not be affected.

PAGE 37

! 19 CHAPTER III COMMUNICATION STREAMS AND ALGORITHMIC MOSAICS 3.1.0 Overview While studyi ng the PL, several types of visual streams were intro duced. As mentioned in Chapter II , it is assumed that the PL is the engine that generates visual streams to visualize and investigate any problem. During this research, visual streams were divided into t wo major groups based on their interaction with the outer world. The internal streams are the streams interact internally, i.e., only with each other. An example of an internal stream is a discovery stream, which could be an observation, construction, or a validation stream. On the other hand, the communication streams are the streams that operate between the internal streams and the outer world. Communication streams transform the information from the world around us to visual inputs to be processed by the internal streams and vice versa. The subset of the communication streams that converts the outer world information into the internal streams is the impression streams. The subset of the communication streams that converts information of the internal stre ams to secondary languages, i.e. human languages, is called the expression streams. This chapter includes a detailed description of the communication streams and the associated results of this research. 3.2.0 Motivation After several investigations of the PL, a new class of visual streams was discovered, the communication streams, [27] . These streams were never applied to any of the previous work on the PL. Thus, it was important to investigate the nature of the commination streams to reveal additional structures of the PL , if any.

PAGE 38

! 20 3.3.0 Communication Streams The main purpose of the communication streams is to exchange dynamic pictorial information between the internal streams of one entity such as a human or a robot and the outer world, i.e ., other entities (humans or robots), and the nature. These streams utilize a symbolic language t hat other entities understand. In addition to symbols represented as letters, digits, scientific symbols, sounds, etc., this language could use drawings, whic h are snapshots of the internal streams. Another group of communication streams passes symbolic and pictorial information back to the internal streams. Communication streams can be divided into groups based on their functionality. One group is the set exp ression streams. The main function of these is to convert information from internal streams into a form acceptable for passing to the outer world. Extensive case studies in [16] [6] revealed two ty pes of expression streams, pictorial and symbolic, [25] . The pictorial expression stream translates an internal visual stream into a sequence of snapshots of this internal stream. A typical example of the pictorial stream is th e set of illustrations that scientists draw to explain their ideas. On the other hand, the symbolic expression stream translates the internal visual streams by "watching" the morphing of this stream. This "watching" also includes "tagging" or naming the im portant objects, actions, and events to create the symbolic shell. Then, employing a set of rules, i.e., a grammar [27] , the symbolic expression stream converts this symbolic shell into a string of symbols in a speech, an algor ithm, a scientific theory, etc. In such a case, the tags are utilized as a set of terminal symbols for the grammar. The symbolic expression stream also has the ability to add captions to the illustrations generated by the pictorial visual streams. In parti cular, the

PAGE 39

! 21 output of the expression stream could serve as a solution for some problem to be presented to the scientific community. The other subset of the communication streams is the set of impression streams. While the expression stream links the interna l streams with the symbolic languages, the impression stream works in the opposite direction. Impression stream converts the external languages into visual objects that interact based on the description provided by the symbolic language. The impression st ream converts the symbolic strings of an external language into the animated visual objects. All visual streams operate in the form of thought experiments, [16] . In that sense, the thought experiments generated by the communica tion visual streams provide a link between the PL and the outer world. The internal visual streams operate employing internal thought experiments , while the communication streams, both expression and impression streams, utilize communication thought experi ments. These thought experiments work in mutually opposing directions, as shown in Figure 1 . Figure 1 . The communication streams pass information between the internal streams and the outer world and vice versa .

PAGE 40

! 22 According to the hypothesis in Section 1.1.0 , the expression of an idea is the transformation of the internal visual streams into symbolic language via the expression stream. As already mentioned, a symbo lic language can be verbal, written, code based, e.g . , Morse code , etc. It can also be a technical or natural language. Technical languages usually include specific symbols and rules developed within a given branch of science. In some cases, it is very di fficult to develop proper communication streams, especially those related to science. However, when those streams have been developed, such as those for natural languages, they could run in real time or even faster. This is what happens in the human brain. Indeed, humans usually do not have gaps between their thoughts and the expression of them (unless they are not familiar with the language they are speaking). In the latter case, humans create the following chain of streams. First, they use an expression stream based on their native language. This stream is then converted into an internal stream of another entity, using its impression stream. This new internal stream is translated into another expression stream using the language familiar to another entity (and non familiar to the first one). In other words, one expression stream initiates another impression stream to create a new "different language" internal stream to be linked to another expression stream , thus creating a chain of internal and communicat ion streams. 3.4.0 Algorithmic and Symbolic Mosaics While working on the different problems in Chapters 5, 6, and 7, the expression stream translated the discovery streams into an algorithmic or a symbolic mosaic. The expression stream creates algorithmic mosaic s by attaching several instruction aggregates. Every such aggregate represents one of more actions made by other visual streams. The instruction aggregate consists of two information aggregates, the operation aggregate and the parameters

PAGE 41

! 23 aggregate. For exa mple, the instruction aggregate "Add (number1, number2)" tagged the action of the construction stream when it adds two numbers together. During this research we discovered several instruction aggregates introduced by the expression stream. Each aggregate r epresented an operation executed by the construction stream while building a part of a mosaic. Table 2 shows a general form of an instruction aggregate utilized by the expression stream , as revealed in Chapters V, VI , and VII . The set of various instruction aggregates is expected to increase based on the subsequent PL investigation s . Table 2 . Several instruction aggregates introduced during this research . Instruction Aggregate Meaning Overlay (first mosaic, second mosaic) Overlay two mosaics Add (number1, number2) Add two numbers Read (information mosaic, reading rules) Read information in one mosaic based on defined reading rules If condition then É Check a condition during construction Repeat (number of cycles , repeating from) Repeat several instruction aggregates. This operation aggregate takes two parameter aggregates: 1. Number of cycles : how many times this operation should be repeated. 2. Repeating from: the starting place to repeat Choose (mosaic/ tile/ aggregate, set) Take a tile, aggregate, mosaic from a set of tiles, aggregates, mosaics , respectively Copy (mosaic1, mosaic 2) Take an identical copy of one mosaic and save it into another Align (first mosaic, second mosaic, alignment rules) Align two mosaics according to a defined alignment rules Attach (first mosaic, second mosaic, Complementarity rule ) Link two mosaics together based on the complementarity rules Scan (mosaic, movements rules) Send the Ghost to walk through a whole mosaic based on the movement rules Remove (aggregates, tiles, or mosaics) Delete tiles, aggregates, or mosaics from the visual stream

PAGE 42

! 24 Instruction Aggregate Meaning Detach (mosaic, aggregate) Detach an aggregate from a mosaic Generate (aggregate, construction rule s, tiles set) Build an aggregate out of a set of tiles based on the construction rules Call algorithmic shell Call any part of the algorithmic mosaic Move (mosaic, current location, destination location, movement rules) Move one mosaic from the start loc ation to the end location according to the movement rules Erase (particulars, mosaic/aggregate, general shape) Erase the specifics from one mosaic and convert it to a general shape. The expression stream builds the symbolic mosaics to save the informat ion declared by an internal mosaic. During this research, many symbolic mosaics were generated by the expression stream, in the form of tables, to save the different matching rules announced by the discovery streams. In C hapter VI , the table that maps betw een the DNA base and letter tiles was constructed by the expression stream, shown in Table 3 . As the AD was applied to the protein synthesis problem more symbolic mosaics were identified, Chapters V, VI , and VII . 3.5.0 Results The algo rithmic mosaic is yet another type of information mosaic built to communicate written information. However, algorithmic mosaics are constructed by the expression stream using a secondary language. These mosaics are a direct translation of the internal stre ams in order to be understandable by the outer world. They can also be read by the impression stream and translated into a visual input to be passed to the internal streams. In the next chapters, we consider different algorithmic mosaics generated by the e xpression streams. Specifically, at the end of every discovery, a complete algorithmic mosaic representing a solution to the investigated problem is generated. The work on Table 2 cont'd

PAGE 43

! 25 communication streams and algorithmic mosaics will be continued in the future based on the results obtained in this research.

PAGE 44

! "# ! CHAPTER IV INTRODUCTION TO THE INVESTIGATED DISCOVERIES 4.1.0 Introduction to Linguistic Geometry Linguistic Geometry, LG, is another algorithm based directly on the PL, i.e., on visual streams, [28] , and [26] . Currently, it is understood that the development of LG was not an invention, but the discovery, or rediscovery, of an ancient algorithm developed by the human brain in an attempt to optimize warfighting, [29] , and [30] . LG was rediscovered, and further developed over the last 40 years, as an approach to solving opposing games, such as chess, in real time. This research, started in 1972 as a series of experiments for the analysis and modeling of the chess experience, to include previously developed strategies of advanced players or masters of the game. By generalizing this experience, a computer program, PIONEER, was developed, [28] . PIONEER solved a number of well known complex chess positions and endgames with only 100 variations included on the search tree. Studies following PIONEER's approach led to the general LG theory, a new type of game theory for solving Abstract Boa rd Games (ABG). The mathematical foundation of LG is the Hierarchy of Formal Languages. One of the major results in LG permits the solving of classes of games without any kind of search and simultaneously proving the optimality of the proposed solutions. T he analysis of this discovery played an important role in the development of the AD, [8] , [17] , [12] , and [13] [18] . 4.1.1 Trajectories Before LG generates optimal gaming strategies, the problem must be defined as an ABG. This requires the definition of all the ABG terms, e.g. , the definition of the abstract board, the players, the pieces, and the game rules.

PAGE 45

! "$ ! Assume that an ABG is defined including its reachabilities, which describe all the locations reachable from an arbitrary location in one step, for an arbitrary piece p, [28] . Trajectories for piece p from location x i to x j of length l , is the set of strings of symbols that represent all possible paths for piece p between x i and x j of the length l . Any trajectory of length l is the shortest trajectory if l is the minimum possible length among all the trajectories of piece p between x i and x j . It is possible to have more than one shortest trajectory between two locations. Admissible trajectories of degree k are those trajectories that can be divided into k shortest trajectories. Hence, any shortest trajectory is an admissible trajectory of de gree 1. In LG, the set of all t he trajectories of the length less than H (H is an integer and H ! 1) for the current state of an ABG is formally represented as the Language of Trajectories L t H , [11] , and [28] . Figure 2 shows two different trajectories that piece p can take between x i and x j , one of length 4 and the other of length 6. Following a series of previous works by Stilman and Aldossary to apply the AD on different discoveries in the field of LG , [8] [4] , the AD was applied to rediscover the admissible trajectories algorithm and revealed all the AD components led to this discovery, which is shown in Chapter V . Figure 2 . Two different trajectories between points x i and x j .

PAGE 46

! "% ! 4.1.2 Introduction to the Central Dogma DNA is a repository of genetic information for all living organisms, and since the discovery of its 3D structure scientists have been puzzled by the genetic co de, [3] . One of the areas that puzzled scientists is the generation of proteins; specifically, understanding how the genetic information is utilized for generating proteins. In 1958, Francis Crick hypothesized that the genetic information travels from DNA to another repository, RNA, and then it is translated into proteins, [35] . Today, this process is considered the Central Dogma of Molecular Biology. As the first step of applying the AD to rediscovering the solution of the two major problems in the Central Dogma, i.e. , the genetic code and protein synthesis, the AD broke them into a sequence of thought experiments (including successful and failed experiments). This sequence did not ap pear all at once, i.e., every experiment generated the next one until a so lution was found. The i nformation about the experiments that historically took place, was obtained from the following papers [3] [54] . In order to develop this sequence of thought experiments, the inputs and outputs of the original experiments must be considered, in order to include the information available at the time of the o riginal discovery. In terms of input, the actual components of DNA (the DNA bases) were known, but the meaning of the sequence of those bases (the genetic code) was not completely understood. For the output, some fragments of protein sequences were known, such as the amino sequence of insulin. The total number of amino acids, which are the fundamental components of proteins in all living organisms, was known as well. However, some of the amino acids that built the protein mosaic were still unknown at that time, [3] . As mention ed in Chapter II , the AD was expected to construct the sequence of thought experiments to reveal the algorithm for translation of genetic information into the structure of proteins. To construct this sequence, it had to utilize information about the compon ents of the

PAGE 47

! "& ! algorithm of translation known at the time of the original discovery. For example, the mechanism for DNA replication was hypothesized as follows: unzip the double helix and then duplicate each strand independently. Each strand should form a new copy of the DNA. This algorithm could also be applied for creating the RNA mosaic from a DNA template, [50] . On the other hand, the complete algorithm of translating DNA into protein was unknown. Even the most basic questions were yet to be answered. The correct method of reading the DNA sequence was still a mystery. It was not clear whether to read the genetic code on the two strands of the DNA simultaneously or only read one strand at a time . If the genetic c ode was read only from one DNA strand, then the dilemma was which one should be read. 4.2.0 DNA and RNA In order to find the essential algorithms that guide protein synthesis, the AD had to define the visual model required as the initial input. The AD reasoned t hat since the genetic information resides in the DNA, the initial input must be a complete 3D structure of the DNA , the double helix . Visual streams, especially the observation stream, had to morph this input to create the informational representation of t he DNA necessary to start transcription, the first step in protein synthesis that creates the RNA intermediate mosaic. In order t o do so, the AD retrieved the output resulting from the experiments that revealed the structure of DNA , [10] . Such information stated that the DNA mosaic consists of several types of tiles. They include four different nitrogenous bases, Adenine (A), Guanine (G), Thymine (T), and Cytosine (C). One of those nitrogenous bases is included in every nucleot ide aggregate, which also includes a sugar tile and a phosphate tile. Each strand of the double helix is a chain of nucleotide aggregates. The strands are linked together by hydrogen bonds via nitrogenous bases (inside the structure of the DNA), whereas th e outside structure of the double helix is formed by two sugar phosphate backbones.

PAGE 48

! '( ! Following the complementarity matching rule, the nitrogenous bases that link the DNA strands are paired together as follows, A with T and C with G. At the beginning of inve stigating protein synthesis, the existence of RNA was known, but its role in the process was not clear. The nucleotide bases of RNA matched well those in the DNA; A, G and C were exactly the same, while the base of U can be considered as a replacement for T. Inside the cells, special molecules copy one strand of the DNA to create the RNA molecule. While copying, every T nucleotide is transformed into a U nucleotide. During the discovery of protein synthesis, different kinds of RNA were identified. They incl ude mRNA and tRNA. The mRNA is the RNA that carries the genetic code are discussed in Chapters VI , and VII . The tRNA is the RNA are responsible for carrying different amino acids and moving them to be assembled next to each other to form the peptide mosaic , [50] . 4.3.0 The Genetic Code This section discusses the genetic information that was revealed during the application of the AD to the genetic code problem. At the end of revisiting genetic code discovery, the AD con cluded that the genetic alphabet consists of the four letter tiles. Every three letters would be connected to each other to form one code (codon) aggregate that signals synthesis of the one amino aggregate. The genetic alphabet contains 64 different code a ggregates, 61 of which represent the existing amino acids and three are the "Stop" aggregates that signal the termination of the current genetic reading, [50] . After discovering the DNA structure, the AD trigger ed thought experiments to decipher the genetic code as shown Chapter VI . A series of thought experiments was initiated to redefine this problem , i.e., to switch visual streams from one theme to another , see Sections 6.3.0 , 6.4.0 , 6.5.0 , and 6.6.0 . These experiments resulted in the conversion of a 3D chemical theme into an

PAGE 49

! ') ! informational one, leading to the AD's first attempt to crack the geneti c code. When this attempt failed, the AD defined and redefined the inputs and assumptions in order to decode each genetic code symbol, one by one, until the meaning of all of the genetic symbols was finally discovered, [42] , and [55] . 4.4.0 Protein Synthesis The AD preliminary analysis showed that the algorithm of protein synthesis consists of two main phases: transcription and translation. Transcription is the first stage needed to turn the DNA mosaic into an RNA mosaic. This algorithm converts the four letter tiles in the DNA mosaic, A, T, C and G into A, U, C and G, respectively, to form the letter tiles in the RNA mosaic. At the beginning of transcription, RNA is generat ed by a special RNA polymerase that binds to a specific region of DNA, which contains the genetic information. First, the polymerase unzips the two DNA strands (from each other) to start the transcription. This polymerase then reads the base tiles in the D NA and builds the RNA mosaic accordingly. It will continue to read through the DNA strand until it reaches a termination region that stops the transcription. Once transcription is terminated, the polymerase releases the DNA mosaic. The result of this proc ess is an RNA mosaic that carries the genetic information required to form one protein chain, [50] . During translation, the genetic code on the generated mRNA resulting from transcription is read to construct th e protein peptide mosaic. Many molecules work together to read the mRNA strand in order to transform it into a protein mosaic. In this research, the AD was applied to rediscover the translation algorithm. A number of thought experiments essential to re disc overing the process of activating the tRNA mosaic were defined, as shown in Section 7.3.0 . Then, the AD was applied to rediscover different models of the translation process, given in Sections 7.5.0 , 7.6.0 , 7.7.0 , and 7.8.0 .

PAGE 50

! '" ! By now it is known that during the translation, ribosome molecules are constructed and linked to the mRNA mosaic. Once a ribosome is attached to an mRNA mosaic, the ribosome starts reading the genet ic code of the mRNA in triples, i.e., three mRNA tiles at a time. Based on the mRNA reading, the ribosome will call for the corresponding tRNA mosaic , yet another type of the RNA molecule . Every tRNA mosaic is attached to a specific building block, i.e., the specific amino aggregate. The ribosome will continue reading the genetic code on the mRNA mosaic and calling for the correct tRNA mosaics. Every time a new tRNA mosaic enters the ribosome, the ribo some will link its attached amino aggregate to the growing peptide mosaic. The ribosome will stop reading when a "Stop" code aggregate is reached. Chapter VI describes the beginning of the research to apply the AD to redefine the rules that control the tra nslation algorithm. At the end of Chapter VI the rules that control the reading of the m RNA mosaic were rediscovered . Chapter VII continues the investigation to rediscover the entire mechanism of translation.

PAGE 51

! '' ! CHAPTER V REVISITING THE ALGORITHM OF ADMISSIB LE TRAJECTORIES 5.1.0 Overview In this chapter, the work of applying the AD to rediscovering the algorithm for generating admissible trajectories is described. For this purpose, all the thought experiments that led to this discovery were replayed by the AD , [5] . The role of mosaic reasoning was emphasized, for the first time, in this series of thought experiments. Specifically, all the components of mosaic reasoning that participated in constructing the algorithm for generating admiss ible trajectories of degree 2 between arbitrary locations on an abstract board were found. Also, the communication thought exper iments were identified. These thought experiments are run by the expression stream to translate the internal streams into a symb olic language. The final result of the expression stream is an algorithmic mosaic that describes the algorithm for discovering admissible trajectories, [5] , and [28] . In LG, admissible trajectories of degree 2 are all the trajectories that are constructed from two shortest trajectories. The AD assumed that the algorithm for generating the shortest trajectories has already been discovered, [8] . Via the impression stream, t he AD received the following inputs: the start and end points, the length of the desired trajectory (l) the type of reachability 1 , as well as the output from the algorithm for generating the shortest trajectories, if any exist. The reachability relation go verns the movement from one point to another. The output of the algorithm for generating admissible trajectories is a tree mosaic that represents all possible admissible trajectories from start to end, [11] . !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! !!!!!!!!!!!!!!!!!!!! ! ) ! *+,!-./,0!1+21!3451-4/!1+,!647,6,51!8-46!1+,!012-1!14!1+,!,59!:4;510< !

PAGE 52

! '= ! 5.2.0 Motivation Redisco vering the algorithm to generate the admissible trajectories was the next step, after applying the AD to rediscover the shortest trajectories algorithm, [4] . The goal was to emphasize the components of the AD that were found wh ile rediscovering the shortest trajectories and to find new components, if possible. Also, it was important to study the applicability of mosaic reasoning in this discovery. Further, a simple straightforward discovery such as admissible trajectories was ne eded to break down all the communication thought experiments that led to the identification of the admissible trajectories algorithmic mosaic. 5.3.0 FindMid Experiment 5.3.1 The Observation Stream In its preliminary stage, the observation stream received the definiti on that any admissible trajectory of degree 2 consists of a pair of shortest trajectories. From this definition, the observation stream concluded that there must exist a mid point between these shortest trajectories. It focused itself to save the algorithm for generating the shortest trajectories. It also saved the incoming admissible trajectory and its length, l . The observation stream, by analyzing the current inputs, found that all components of the admissible trajectory were already known. The only mis sing piece of information was the set of all mid points. The observation stream tasked itself to investigate the construction of this set. To schedule the required experiments, the observation stream modeled this problem as a backyard to be paved with til es and introduced an appropriate construction set. In this case, the bundle of admissible trajectories to be constructed represents a complete mosaic of tiles,

PAGE 53

! '> ! which is similar to the process of paving a backyard. The stream also assumed that adjacent til es along the trajectories would be linked by the reachability relation. It then introduced the set of trajectory tiles (T tiles). The observation stream declared special subsets of the T tile set: the mid point tiles (Mid tiles), the start tile (Start tile ) and the end tile (End tile). The stream reasoned that for every admissible trajectory of degree 2, a mid point is always the end point of one shortest trajectory and the start point of another one. Figur e 3 shows the initial vi sual model created by the observation stream. At the beginning, the start point (blue square) and end point (orange square) were marked on the ABG board. The observation stream waited for the length of the admissible trajectories, l . Once received, the for ward map (bottom left of a given block) and the backward map (top right of a given block) were retrieved. These maps include all the forward and backward distances for each square in Figur e 3 . Forward distances are those distanc es from the starting to the end point s . Backward distances are the reverse. The observation stream scheduled an experiment, FindMid, to study the nature of the mid points and, possibly, to find and store them as a set of all possible mid points (DOCK). Th e input of FindMid was a visual model that includes the start and end points and the construction set that contained : the length of the trajectory ( l) , the type of reachability relation, and the algorithm for generating the shortest trajectories.

PAGE 54

! '# ! Figur e 3 . Visual representation of the initial input model . 5.3.2 The Construction Stream and the Expression Stream The construction stream started the FindMid experiment after receiving the visual model and the construction set. As this stream searched for the mid points, it reasoned that every mid point is an end point of a given shortest trajectory. Therefore, the length from the start point to the midpoint must be equal to the length of the shortest trajectory. Similarly, this mid point wou ld be the start point of another shortest trajectory and its distance from the end point had to be equal to the length of the second trajectory. Consequently, the total length l of the admissible trajectory, which is equal to the sum of the lengths of the first and second components, must be equal to the sum of distances described above. The observation stream reasoned that it could find those mid points as the locations where the sum of forward and backward distances is equal to l . The reading rules needed to read the forward and backward distances, as well as the algorithm for calculating the sum, comes to the

PAGE 55

! '$ ! construction stream through its input via the impression stream, as a part of the algorithm for generating the shortest trajectories. The constructi on stream retrieved the forward and backward distance maps created originally for generating the set of shortest trajectories. The construction stream overlaid the two maps to create the sum map. It concluded that all the mid points must be the points wher e the sum equals to l . The construction stream saved this conclusion as the reading rule to mark the mid points. These points would be marked as members of DOCK. Figure 4 shows the visual model, after morphing by the construction stream. This model included all Mid tiles (green squares) that belong to the set DOCK after "erasing the particulars", i.e., eliminating unnecessary locations on the ABG board. Figure 4 . The visual model after being morphed by the construction stream . The expression stream monitored the actions of the construction stream and tagged the events needed to find the mid point. The stream generated a "Find Mid" shell that included all steps, or instruction aggregates, needed to build the DOCK set. Such a shell, once finalized, could run independently from the internal streams to identify all the mid points for

PAGE 56

! '% ! any incoming admissible trajectories. At the beginning, the expression stream saved the instruction aggregate "Call generate forwar d distance()." The expression stream then added to the current shell "Call generate backward distance()." The algorithms for building the forward distance and backward distance maps were inputs from the "shortest trajectories experiments." The expression s tream saved the action of overlaying the two maps. Based on that action, the expression stream added a new instruction aggregate "Overlay (forward distance, backward distance)." The last instruction aggregates added by the expression stream are " Add (forwa rd distance, backward distance) " , "Read ( current sum , Null) " , and "If current sum := l then Save current location in DOCK." The expression stream saved the "Find Mid" shell to be added later to the final algorithmic mosaic corresponding to the construction of the admissible trajectories. After defining the DOCK set, the visual streams, both observation and construction, focused on the investigation needed to solve the problem into an examination of the members of the DOCK set only. The DOCK set is a special type of scope set that the AD sometimes identifies to help focus the streams and minimize the time required to solve a problem. 5.4.0 Define the Mosaic's Aggregates and Matching Rules The observations stream then moved to the definition of the mosaic's aggrega tes and used the previous conclusion that for every admissible trajectory of degree 2, a mid point is always the end point of the first trajectory and the start point of the second one. The stream identified the From Mid and To Mid aggregates. The To Mid a ggregate leads from the Start tile to the Mid tile and the From Mid aggregate leads from the Mid tile to the End tile of an admissible trajectory. Each of these additional aggregates represents a bundle of shortest trajectories that either end at or start from the chosen mid point. However, they are

PAGE 57

! '& ! represented not just as lists of trajectories, but also as tree data structures, taking into account that many trajectories in a bundle have a substantial number of parts in common. The observation stream also r ecognized a third aggregate, the Mid aggregate. The two Mid tiles are used to build the Mid aggregate, the major component of the bundle of complete admissible trajectories associated with a specific mid point. This bundle is also represented as a tree of admissible trajectories. To construct the From Mid and To Mid aggregates out of a set of tiles, the observation stream identified the global transformation rule. This rule guided the construction stream while building the To Mid and From Mid aggregates. Th ese aggregates turned the pair, Start tile and Mid tile, into the To Mid aggregate, and the pair, Mid tile and End tile, into the From Mid aggregate. This global transformation rule is the algorithm for generating the shortest trajectories and is included as an input to the AD via the impression stream. Another transformation rule turned each triple, To Mid, Mid tile and From Mid, into the Mid aggregate of admissible trajectories. The algorithm which accomplishes this has yet to be discovered, see Section 5.5.0 . The observation stream then specified the global complementarity rule, which controlled the proper construction of the Mid aggregate out of the To Mid and From Mid aggregates. The stream found that the To Mid and From Mid aggregates could be merged, or "glued", together at a specific tile, the Mid tile. Next, t he observation stream started to assign interchangeability rules. Global interchangeability rules allowed the construction stream to take one Mid aggregate of the a dmissible trajectory tree mosaic and replace it with another one from the set DOCK. This new Mid aggregate had a different mid point, which was a Mid tile. This would certainly

PAGE 58

! =( ! change the picture, but the whole structure would stand. The observation strea m defined the Mid aggregates to be the global plug ins for the admissible tree mosaic. Plug ins are the interchangeable aggregates in any mosaic that can be replaced by another without affecting the final structure of the mosaic. The local interchangeabil ity rule permitted the construction stream to take any Mid tile and interchange it with any other Mid tile, which may lead to the construction of a different Mid aggregate. After passing through all the Mid tiles of DOCK, the stream generated a complete mosaic composed of the bundle of admissible trajectories. 5.5.0 GlueTree Experiment 5.5.1 The Observation Stream The observation stream concluded that after defining all mid points, the whole bundle of admissible trajectories could be built. The observation stream sch eduled a GlueTree experiment. The inputs to this experiment include DOCK, the start and end points, and the global transformation rule, i.e., the algorithm for generating the shortest trajectories. 5.5.2 The Construction Stream The construction stream started th e GlueTree experiment to construct the final admissible tree. The construction stream chose the first mid point in DOCK, however, it is not important which point is chosen due to the global interchangeability rule. When the stream picked a mid point, two t rees representing the two bundles of shortest trajectories were created using the appropriate algorithm, each tree representing either the To Mid or From Mid aggregate. The construction stream then began building the Mid aggregate by taking the To Mid tree and gluing every leaf on this tree to the root of the From Mid tree, see Figure 5 .

PAGE 59

! =) ! Figure 5 . Visual s imulation of the GlueTree t hought e xperiment . While monitoring the work of the construction stream, the expression stream generated the algorithm associated with building the final tree. The expression stream initiated a shell named "Construct one branch." The first instruction aggregate added to this shell was "Choose ( Next mid point, DOCK)." As explained above, the instruction aggregates needed to build the DOCK set are defined in the "Find Mid" shell. When the mid point is chosen, the construction stream called the algorithm for generating the shortest trajectories to build two bundles of shortest traject ories (one from the start point to the current mid point and the other from the current mid point to the end point). The expression stream saved into t he "Construct one branch" shell the following instruction aggregates: "To Mid := Call generate *4 ? @;9! ! A-46 ? @;9! !

PAGE 60

! =" ! shortest t rajectories(start, mid point)", "From Mid := Call generate shorte st trajectories(mid point, end) . " The gluing point is the Mid tile. The construction stream declared a new alignment rule , where the Mid tiles of the two tree mosaics must be aligned. Then, t he stream glued the aggregates together by making copies of the From Mid aggregate and overlaying the root of each copy of the From Mid tree to the leaves of the To Mid tree. The result of this gluing is a Mid aggregate, which represented the bundle of adm issible trajectories passing through the chosen mid point, shown in Figure 6 . Figure 6 . Building a Mid aggregate from the To Mid and From Mid aggregates . Consequently, the expression stream added more inst ruction aggregate s to the "Construct one branch" shell based on the new actions of the construction stream. The expression stream first tagged the action of multiplying the From Mid aggregate, "Copy (From Mid, new From Mid) . " The new From Mid was glued to the current leaf of the To Mid in such a way that the two mid points were merged. Therefore, the expression stream reasoned that the two trees must be overlaid so that the two mid points are aligned. The expression stream generated new instruction aggregat e s "Align (new From Mid, To Mid, alignment rules)" and "Attach (new From Mid, To Mid, complementarity rules)."

PAGE 61

! =' ! The construction stream then chose another mid point from DOCK and glued the second To Mid tree to the second From Mid tree to form the next Mid aggregate. The stream repeated these steps for all mid points. Once the construction stream had visited all of the mid points of DOCK, a complete set of Mid aggregates was generated. The expression stream reasoned that every time a new mid point is chose n from DOCK, the " Construct one branch" shell would run again. Therefore, the expression stream added a loop that repeats the shell until all the mid points in DOCK are visited. Finally, the construction stream reasoned that the construction of the final a dmissible tree from the Mid aggregates requires the merging of those aggregates. This is necessary due to the fact that admissible trajectories stored in different Mid aggregates may have common parts beginning from the start point. These parts may include the whole trajectory. Merging the aggregates would eliminate such duplication. This would terminate the GlueTree experiment. The construction stream reasoned that the scanning would be done using the Ghost that would move following the movement rule 2 , whi ch start s to move from the tree root mosaic and pass through all branches of the tree and mark any redundant branch. The expression stream saved t he final instruction aggregates, "Scan (admissible tree, movement rules)" and "Remove (redundant branches)." Figure 7 shows the visual representation of the final tree mosaic of the admissible trajectories. Finally, the expression stream combined all the shells defined while monitoring the work of the discovery streams. Combining these s hells led to the construction of the final algorithmic mosaic for generating the admissible trajectories, see Figure 8 . !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! !!!!!!!!!!!!!!!!!!!! ! 2 Movement rules are part of the transformation rules that guide the movement of one mosaic inside visual streams. ! !

PAGE 62

! == ! Figure 7 . The v isual r epresentation of the m osaic of the t ree of the a dmissible t raje ctories . Figure 8 . The f inal a lgorithmic m osaic corresponded to the construction of the a dmissible t rajectories t ree . @;9 ? 2BB-,B21, ! @;9 ? *;/, ! C12-1 ? *;/, ! D59 ? *;/, ! Call generate forward distance Call generate backward distance Find Mid Overlay (forward distance, backward distance) Add (forward distance, backward distance) L: Read ( current sum , Null) If current sum : = l then Save current location in DOCK Repeat ( Board length, L) Choose (Next mid point, Dock) To Mid : = Call generate shortest trajectories(start, mid point) Construct one branch From Mid : = Call generate shortest trajectorie s(mid point, end) N: Copy (From Mid , new From Mid) Align (new From Mid, To Mid, alignment rules ) Attach (new From Mid, To Mid , complementarity rules ) Repeat ( To Mid leaves, N) Repeat (number of elements in DOCK , L ) Scan (admissible tree, movements rules) Remove (redundant branches) L:

PAGE 63

! => ! 5.6.0 Results According to analysis of the construction of the algorithm for admissible trajectories, mosaic reasoni ng guided the AD to this discovery. Specifically, the AD applied mosaic reasoning to build the admissible tree mosaic. Defining tiles, aggregates, and matching rules helped the AD to focus the construction. A new set of matching rules appeared in this disc overy, the reading and the alignment rules. The readings rules were applied by the AD to read the forward and backward distances, find the sum, and to compare it with the length of the admissible trajectory. The alignment rules were needed to align the Fro m Mid aggregate. Via "erasing the particulars" approach, the AD found the scope set of the mid points. Once the mid points were constructed, visual streams started assembling the mosaic components and building the final admissible tree mosaic. The algorit hm for generating the shortest trajectories was utilized by the AD as a standard procedure in the construction of the output mosaic. The algorithm for shortest trajectories was used several times to generate bundles of those trajectories from start to mid point and from mid to end point. By having the tiles, Start, Mid and End, the Matching rules, and the bundle of shortest trajectories, the AD was able to generate a complete admissible trajectory tree mosaic and, this way, construct a complete algorithm for generating admissible trajectories of degree 2. For the first time, the thought experiments that led to the discovery of the admissible trajectories algorithm were identified. The "hidden" role of mosaic reasoning and its components in this discovery was revealed. Also, all the communication thought experiments run by expression stream to translate internal streams into a written algorithmic mosaic were defined in this Chapter .

PAGE 64

! =# ! CHAPTER VI REVISITING THE DISCOVERY OF GENETIC CODE 6.1.0 Overview In this cha pter, the AD was applied to revisit the discovery of the genetic code . This application of the AD required the investigation of existing mosaics to reveal the rules needed to build a new mosaic out of existing ones. This led to the conclusion that mosaic reasoning is a major component of the AD and may be used to identify the unknown structure of any real world or informational object. Also, mosaic reasoning could be applied by the AD to reveal the rules essential in building one mosaic out of multiple exi sting ones. Information mosaics are single or multi dimensional tables that appeared for the first time in this discovery. In order to build any information mosaic, the AD must apply the transformation rules to convert a known mosaic into entirely new in formation one. While investigating the genetic code problem, the AD declared the reading rules, which were needed to read one mosaic and build another one based on that reading. The AD builds the one dimensional table mosaic as a string of aggregates and t he n dimensional table mosaic as n strings of aggregates. The aggregates of the information mosaic are built out of letter tiles and a visual stream must read these letter tiles according to the reading rules. The aggregates of the input mosaic are read an d, based on that reading, are turned into new aggregates in the new information mosaic. In this discovery, three information mosaics, required during the process of protein synthesis, were identified. Also, two transformation rule phases were declared: tra nscription and translation. Each phase represents a complete process that is part of the protein synthesis. During transcription, the mRNA mosaic is transformed from the DNA mosaic. The DNA mosaic is built out of a single string of aggregates. Each aggreg ate is a group of letter tiles. A, T, C or G. The

PAGE 65

! =$ ! DNA letter tiles are transformed into the corresponding mRNA letter tiles, A, U, C and G, respectively, to form a new string of mRNA aggregates. During translation, the genetic sequence in the mRNA mosaic i s read, according to the reading rules, and forms the peptide information mosaic, the output mosaic. The rest of this chapter discusses the AD trials undertaken to find the correct reading rules to identify the transformation algorithm. Such an algorithm c ontrols the reading of the mRNA to turn it in to the peptide mosaic. 6.2.0 Motivation After applying the AD to the discovery of admissible trajectories, the motivation was to find another discovery from a different field, i.e. , one not LG. The genetic code proble m was chosen because it is an information problem like LG. Two information problems, from two different fields, required investigating the similarities in the behavior of the AD while solving the two problems, i.e. finding a general structure of the AD. Al so, our research goal was to utilize the AD components identified during previous analysis, as well as to reveal any new components leading to the correct solution of the genetic code problem. 6.3.0 The Gamow Thought Experiments (The Diamond Code) After finding the correct structure of the DNA mosaic [10] , the AD triggered a new investigation of the genetic code problem. The AD questioned the way amino acids are aligned to form the protein peptide mosaic based on the genetic code in the DNA mosaic. In the following sections, the major thought experiments manifested by the AD to find the correct solution to the genetic problem are demonstrated , [7] .

PAGE 66

! =% ! 6.3.1 Building the Input Mosaic In the preliminary stage, the o bservation stream started by identifying the input mosaic for decoding the genetic information. This stream received the initial inputs, which stated that genetic information is located in DNA. The observation stream also accepted the Watson and Crick sche matic representation of the DNA double helix ( Figure 9 , [37] ). It started to erase the particulars by removing all the chemical clutter from the DNA structure since it does not affect the actual genetic message. It described the input mosaic as an abstract DNA mosaic that mimics the schematic representation of the double helix , [7] . Then the observation stream started to generate the construction set. It def ined four "oval shape letter tiles," which correspond to the base tiles of DNA (A, T, G and C). The oval shape was chosen due to the influence of the actual locations of bases in the double helix see Figure 9 . The observation str eam considered the DNA mosaic as two long strings of symbols, utilizing four letters, since DNA is comprised of two strands. The stream assumed that each string consists of words or groups of letters, i.e., the aggregates. The sequence of tiles inside the aggregates in the DNA mosaic strand is exactly the same as the sequence of base tiles in the corresponding strand of the actual DNA structure.

PAGE 67

! =& ! Figure 9 . The s chematic r epresentation of the d ouble h elix , [10] . The observation stream analyzed the structure of DNA to identify the rules needed to build the informational representation of DNA. The stream looked horizontally (i.e. , orthogonally across the DNA axes ) at the structure of DNA. The stream observed that the DNA strands are twisted around a virtual cylinder. Therefore, it reasoned that the two strands of DNA must be twisted around a cylindrical shape representing the basic structure of the DNA mosaic. The stream further concluded that the strands mus t twist around this basic structure following the same transformatio n algorithm identified in the DNA, this mimicking the same twisting of the strands as in the actual structure of the DNA, [10] , and [7] . The observation stream saved this visual model of the DNA mosaic as the schematic representation. When building the DNA mosaic, the construction stream referenced this mosaic as the visual model of construction . The observation stream also saved, i nto the construction set, all the components required to construct the DNA mosaic, which include: the basic structure, the tiles and aggregates, and the transformation algorithm for constructing the aggregates and !

PAGE 68

! >( ! inserting them along the structure of the DNA mosaic. The observation stream considered an oval aggregate with the proper base tile as a generator. The construction stream could generate every new oval aggregate by shifting the most recently built aggregate along the current DNA strand and twistin g the strand around the cylindrical base. After identifying the visual model of DNA and the construction set, the observation stream arranged a thought experiment to build the DNA mosaic and sent the experiment and all needed inputs to the construction str eam. The construction stream built the DNA mosaic by reading one by one the tiles of a DNA strand. Every time the construction stream read a DNA tile, it placed its corresponding letter tile into an oval generator to build one input aggregate. The stream then aligned the input aggregates next to each other along the corresponding strand in the DNA information mosaic. The construction stream twisted that strand along the surface of the cylinder according to the transformation algorithm. When the constructio n stream finished reading the first strand, it constructed the second strand of the DNA information mosaic following the same steps. While placing the two strands along the virtual basic structure, the construction stream kept in mind the correct po sition of these strands in the DNA. In other words, the construction stream placed the two DNA strands in the abstract structure following the exact order of the actual DNA structure. Figure 10 shows a part of the abstract DNA mosaic bui lt by the construction stream; in this figure, the letter tiles A, T, G, and C are assigned the numbers 1, 2, 3 and 4, respectively, [64] . Figure 10 . P art of the i nput m osaic as s hown in [64] . ! ! 1 ! 3 ! 2 ! 1 ! 4 ! 4

PAGE 69

! >) ! 6.3.2 Generating the "Construct Input Mosaic" Shell While the observation stream was identifying the rules needed to construct the input mosaic through internal thought experiments, the expression stream was working in parallel to create a communication thought experiment to build the shell corresponding to the construction of the DNA mosaic. The expression stream first monitored the observation stream's internal thought experiments to identify the basic structure of DNA. It tagged the actio ns and generated the first instruction aggregate: "Build (Basic DNA structure, construction set) . " In the construction set, the expression stream saved the cylindrical shape as the structure any construction stream must follow when reconstructing the DNA. The expression stream also monitored the actions of the observation stream , while it was identifying the base tiles of DNA , and translated them into a symbolic mosaic that contained all possible letter tiles of the DNA information mosaic. This mosaic is a one dimensional representation for the base tile set, see Table 3 . Table 3 . One dimensional letter tiles mosaic based on the DNA base tiles in Figure 9 . Base tile letter tile Further, the expression stream watched the construction of the DNA information mosaic, tagged the steps followed by the construction stream, and saved them in the "Construct information DNA" shell. The expres sion stream saved in this shell the first instruction aggregate based on the construction stream's actions. This instruction aggregate is: "Read (DNA strand, ! E ! ! * ! ! F ! ! G !

PAGE 70

! >" ! reading rules)." The reading rule, for reading the DNA strand identified by the observation stream and converted by the expressio n stream into a string of symbols, is: " Read the DNA strand one tile at a time ." The expression stream then marked the converting action from the base tile in the DNA into the corresponding letter tile following the letter tiles symbolic mosaic, Table 3 . The expression stream added the instruction aggregate: "Convert (base tile, letter tile, letter tiles table)." The expression stream then added to the instruction aggregate: "Generate (input aggregate, complementarity rules, constr uction set)." Generating the input aggregate required, according to the construction set, the use of a generator to construct an oval shape aggregate and place the letter tile inside it. The next instruction aggregate saved by the expression stream was: "A lign (input aggregate, basic structure, DNA transformation rules)." Every time the construction stream read a base tile , it aligned that tile. This alignment required the twisting of the current strand based on the DNA transformation algorithm. The last i nstruction aggregate, in the current shell, identified by the expression stream was: "Twist (DNA strand, DNA transformation rules)." The expression reasoned that the same instruction aggregates would be repeated for the length of the DNA strand and also wo uld be repeated for the next DNA strand. The final shell corresponding to construction of the input mosaic can be seen in Figure 11 .

PAGE 71

! >' ! Figure 11 . "Construct Input mosaic" shell generated by the expression st ream . 6.3.3 Identifying the Output Mosaic After identifying the input mosaic, the observation stream recognized the protein peptide chain as the output mosaic resulting from the transformation of the input mosaic. The observation stream reasoned that the output mosaic would be an information mosaic. Removing all chemical clutter from the amino acids in the peptide chain generated the amino aggregate of the output mosaic. The peptide information mosaic, the output mosaic, is simply a sequence of amino aggregates t hat are linked together. Each amino aggregate included one letter tile that represented the name of the corresponding amino acid. The observation stream must find and save these amino letter tiles in an amino set that was translated simultaneously, by the expression stream, into an amino table. The observation stream noted the existence of a hole or cavity between every two base pairs in the DNA mosaic. The stream started a thought experiment to investigate the nature of these cavities by looking at the or iginal schematic representation of DNA. The stream counted twenty different shapes as a result of this thought experiment, [42] . Comparing the number of the different cavities and the known number of amino aggre gates that build any peptide mosaic, the Construct Input mosaic ! Align (input aggregate, Basic structure, DNA transformation rules) Generate (input aggregate, complementarity rules, construc tion set) Twist (DNA strand, DNA transformation rules) Choose (Unread DNA strand, Set of DNA mosaic) Convert (base tile, letter tile, letter tiles table) Read (DNA strand, reading rules) Repeat ( DNA strand length, N ) Repeat ( 1, L ) N: L:

PAGE 72

! >= ! ob servation stream found that these were equal and exactly twenty. The observation stream concluded that the peptide mosaic is synthesized inside the double helix, i.e. , for each cavity of the DNA mosaic there is a corresponding amino aggregate. To produce the final peptide mosaic, amino aggregates must be aligned in their proper position. Once the genetic code is fully read and translated, the resulting amino aggregates are linked together to create the peptide mosa ic. The matching rules for linking amino aggregates together are simple. By considering the peptide mosaic as one sentence and the amino aggregates as words of this sentence, all amino aggregates must be aligned next to each other as words are in any sente nce, [42] , and [33] . While the observation stream was defining the rules needed to construct the peptide mosaic, the expression stream was building the "Build the peptide" shell base d on the actions of the observation stream. This shell was identified while i nvestigating the protein mosaic. This will be explained in detail in Section 7.5.0 . 6.3.4 Generating the "Construct Output Mosaic" Shell The expression stre am again monitored the observation stream while identifying the rules needed to build the output mosaic. From the actions of the observation stream, the expression stream defined the "Construct Output mosaic" shell corresponding to the current internal tho ught experiment. The expression stream added the first instruction aggregate, "Call (amino aggregate, amino set, code set)", that calls for the correct amino aggregate based on the code aggregate in the current DNA cavity. This amino aggregate was placed i nside the cavity, thus the expression stream inserted the instruction aggregate, "Insert (amino aggregate, current cavity, matching rules)." The amino aggregates are then linked to each other according to the protein matching rules and the expression strea m added the corresponding instruction aggregate to "Construct

PAGE 73

! >> ! Output mosaic" shell . All of these steps were repeated until the entire code is read. Figure 12 shows the final shell for building the output mosaic. Figure 12 . "Construct Output mosaic" shell generated by the expression stream . 6.3.5 Defining the Intermediate Mosaic After defining the DNA and the peptide mosaics, i.e. , the input and output mosaics respectively, the observation stream started to inve stigate the question of how to place the correct amino aggregate inside the DNA cavities. The stream reasoned that the four different letter tiles in the DNA mosaic must code for the twenty different amino aggregates in the peptide mosaic. The stream concl uded that this problem must be converted from a chemical problem into a symbolic cryptanalytic problem. There was a four letter code sequence, (A, T, G and C), that needed to be translated into a twenty letter sequence assumed to represent the names of dif ferent amino acid aggregates in the peptide mosaic. The observation stream reasoned that a mapping between the DNA and the peptide mosaics must be found. This mapping required that one or more input tiles would be used to generate one code aggregate, which then would be interpreted into one letter tile. This letter tile would be transformed into a specific amino aggregate to be added to the peptide mosaic, [42] . The observation stream then moved to investigate the rules needed t o build the code aggregate. The observation stream reduced the problem to finding the answer to the question: how to convert a four letter string into a twenty letter one? By converting the problem from a chemical into an information coding problem, the ob servation stream reasoned that the input Construct Output mosaic ! Call ( amino aggregate, amino set, code set) Insert (amino aggregate, current cavity, matching rules) Link ( amino aggregate, protein matching rules ) Repeat (Length of the input mosaic, L ) L:

PAGE 74

! ># ! tiles might be read in groups of two at a time. Every different reading would represent one code aggregate that code for one amino aggregate. However, when the stream counted all different combinations can be genera ted from reading two letter tiles at once, it counted only sixteen, i.e. , 4 " 4 possible code aggregates that could be constructed out of reading the letter tiles in multiples of 2. Knowing that the number of amino aggregates should be at least twenty, the observation stream eliminated this multiple reading of input tiles and determined to increase this number. The observation stream suggested reading letter tiles in triples, i.e. , three letter tiles at the time. Again, the stream counted all possible combi nations for code aggregates that could be generated from the triple reading of the DNA letter tiles. The observation stream found 64 = 4 " 4 " 4 different combinations which is much more than the needed code aggregates to code for twenty amino aggregates, [37] . The observation stream reasoned that the amino aggregates can be built between the DNA strands' mosaic, i.e. , in the hole between two DNA base paired tiles, whether between (A and T) or (C and G). The observation stream v isualized the location of the amino aggregate to be surrounded from the left and right with those DNA base paired tiles and from top and bottom with any combination of base tiles. This results in the amino aggregate being inside a virtual diamond shape. T he observation stream identified this virtual diamond to be a code aggregate that translates the surrounding letter tiles into the inside amino aggregate. This would create a long sequence of code aggregate between the DNA strands, shown in Figure 13 . The observation stream declared this long sequence to be the intermediate mosaic that is the link between the input mosaic and the output mosaic.

PAGE 75

! >$ ! Figure 13 . The code mosaic sits between the cavities of the inp ut mosaic as shown in [42] , [7] . The observation stream began a thought experiment to identify the correct tiles and aggregates to build the intermediate mosaic. The observation stre am assumed that the intermediate mosaic is constructed out of a sequence of code aggregates. One code aggregate of the intermediate mosaic is constructed out of a diamond body tile, at every corner of this diamond body tile is a code tile (1, 2, 3, or 4) t hat corresponds to one of the of the DNA letter tiles (A, T, U, or G). Every three successive input aggregates are used to build one code aggregate. The observation stream moved forward to find the transformation rules needed to convert the three input agg regates into a single code aggregate. 6.3.6 Assigning the Reading Rules The observation stream established this thought experiment to identify the correct reading rules needed to construct the intermediate mosaic. The stream investigated the reading rules that t he construction stream must follow in reading the DNA mosaic and building the intermediate mosaic. The observation stream followed the first definition requiring that three successive input aggregate be transformed into one code aggregate. Based on the inc oming inputs regarding the protein mosaic, the observation stream reasoned that most of known protein peptide chain structures are symmetrical. Therefore, the observation stream postulated that the code aggregate is also symmetrical, i.e. , it can be rotate d to any direction and still code for the same amino aggregate. For example, a code aggregate that is constructed out of AAC input tiles would be the same as the code aggregates constructed from the ACA and CAA input tiles and all of the three

PAGE 76

! >% ! code aggrega tes would call for the same amino aggregate. The observation stream concluded that this reading rule groups all combinations of the code aggregates from the sixty four distinct code aggregates into twenty group of code aggregates, [3] . Since the number of the code aggregates, sixty four, is much more than the number of the amino aggregates, twenty, and since more than one code aggregate is mapped to one amino aggregate, the observation stream reasoned that the mapping between in put and output mosaics woul d be many to one. The observation stream also found that the 3D spaces between the base aggregates in the DNA mosaic and the amino aggregates in the protein mosaic are similar. That similarity led the observation stream to confi rm to the many to one mapping between the input and the output mosaics. Thus, the stream concluded the degeneracy of the genetic code. Degenerated code means that more than one code aggregate corresponded to the same amino aggregate. Having the degenerated code allowed the observation stream to group the sixty four different code aggregates into twenty groups, every group includes multiple code aggregates that code for one of the twenty amino aggregates, Figure 14 , [3] , [42] , and [64] . Figure 14 . Four symmetrical code aggregates could be generated from the genetic code 123 by the constructi on stream , [7] . In parallel, the expression stream established a new shell to save all the reading rules identified by the observation stream to control the reading of the input mosaic. The expression stream saved the first rea ding rule that the genetic code is degenerated. The expression stream initiated two symbolic mosaics to be completed later. The first table, the code table, would contain twenty groups of code aggregates. Each group contains one or more code aggregates. ! 1 ! 3` ! 2` ! 2 ! 3` ! 1 ! 2` ! 2 ! 3` ! 1 ! 2 ! 2` ! 1 ! 3` ! 2 ! 2`

PAGE 77

! >& ! Th e second table, the amino table, would include twenty different amino aggregates. The expression stream assigned a mapping between the groups in the code set with the amino agg regates in the amino set. Note that both of the sets did not include any specifi c members. The expression stream waited for these two sets to be identified later on by the observation stream. The observation stream also established another reading rule that the reading of the code could be performed with no definite direction. This r ule was based on the symmetrical nature of the code aggregates identified earlier by the observation stream. Following this definition, the expression stream saved a new reading rule stating that the DNA mosaic could be read from right to left or from left to right in the reading rules shell, [3] . The expression stream included the description of this new reading rule in the "Reading rules" shell. Based on the 3 D space similarities between the base aggregates in the input mosai c and the amino aggregates in the output mosaic, the observation stream reasoned that every base aggregate would be ma pped to one amino aggregate. This one to one mapping would give the best mesh between the two mosaics, [33] . That would maximize the the density of information storage. Therefore, the stream established a new reading rule which stated that the input mosaic must be read in an overlapped manner. Overlapped reading means to read a base tile sequence in the DNA mosaic, for example 1, 3, 3, 2, 1É, as following, the first reading is 1, 3, 3, the second 3, 3, 2, the third 3, 2, 1 and so on. Every triple reading of the input sequence is used to generate one code aggregate in the intermediate mosaic, [3] . Again, the expression stream saved the description of the new reading rules in the "Reading rules" shell.

PAGE 78

! #( ! 6.3.7 Identifying the Matching Rules 6.3.7.1 Local Matching Rules The observation stream moved forward to identify the local matching ru les needed to construct the intermediate mosaic. The observation stream found that three letter tiles from the input mosaic could be used to construct a code aggregate with four corners, and each corner would include a code tile corresponding to these lett er tiles. As defined before, the left and right corners of the code aggregate diamond would be facing two base pairs in the input mosaic. Thus, the observation stream concluded that the left and right corners would be attached to complemented code tiles. T herefore, during construction, one of these corners would be assigned based on the genetic information in the input mosaic, while the other corner would be determined using the DNA complementarity base pair matching rules. The observation stream declared a new local complementarity matching rule which stated that the placement of three corners of the diamond would be determined based on the reading of the three input tiles , while the placement of the fourth letter tile would be based on the three letter til es in the code aggregate. Also, the observation stream noted that the code aggregates are symmetric, which led the stream to introduce a new local interchangeability rule. This rule allows the symmetric code aggregates to be replaceable in the intermediate mosaic. Replacing one code aggregate with asymmetric one changes the final picture of the intermediate mosaic but the whole mosaic will stand. An example of four symmetric code aggregates can be seen in Figure 14 . At every step, while the observation stream visually defines one of the local matching rules, the expression stream was translating it into a written description inside the "Local matching rules of the intermediate mosaic" shell. This description will be used as an input when the AD runs again in future thought experiments. Of course, any incoming input to the AD must first be

PAGE 79

! #) ! translated into a visual movie by the impression stream through communication thought experiments before the AD can run it. 6.3.7.2 Global Matching Rules The observation stream initiated a thought exp eriment to define the global mat ching rules. From the reading rule stating that the genetic code is overlapping, the observation stream reasoned that one code aggregate would partially determine the constructio n of the next code aggregate. This conclusion was due to the fact that two consecutive code aggregates share at least two letter tiles. Also, the placement of one code aggregate in one cavity would force the next code aggregate to be placed in the next ca vity. The placement of any amino aggregate inside its code aggregate must be in a stereospecific fashion. Once the reading of the genetic code is completed , the amino aggregates would be linked to each other following the protein matching rules to form the final peptide mosaic, [42] , and [64] . Again, the expression stream translated the actions of the observation stream and generated a written description of the interm ediate mosaic's global matching rules. The expression stream saved all of these global matching rules in the "Global matching rules of the intermediate mosaic" shell that will be used later by future runs of the AD. 6.3.8 Defining the Code and Amino Sets The ob servation stream defined the input mosaic, the output mosaic, the intermediate mosaic, the reading rules , and the matching rules to construct the intermediate mosaic. The observation stream found that the next step of this discovery is to complete the code and the amino sets, and to correctly map between them. The code and the amino sets are types of scope sets that expedite the run of visual streams while working on the genetic code problem, which eventually reduces the path to find the correct solution of the problem. The stream started the thought experiment to

PAGE 80

! #" ! complete the code set with all the code aggregates. The stream triggered the construction stream to start filling the code set based on the rules identified by the observation stream to build any c ode aggregate. The construction stream received the inputs from the observation stream and launched the thought experiment to find all combination s of code aggregates. The construction stream created all possible diamond aggregates using the following appr oach. First, the construction stream generated the body tile of the code aggregate with a diamond shape. Then, the stream added the code tiles to the left and right corners of the body tile keeping into mind that these code tiles must be base paired tiles. For example, if the constructed stream added the code tile Ç 3 È in the right corner , it would automatically add its complement tile, Ç 4 È, in the left corner. Then, the construction stream would add all possible combination of cod e tiles in the top and bott om c o r ne r s Ç 1 ÈÇ 1 È , Ç 2 ÈÇ 2 È , Ç 3 ÈÇ 3 È , Ç 4 ÈÇ 4 È , Ç 1 ÈÇ 2 È , Ç 1 ÈÇ 3 È , Ç 1 ÈÇ 4 È , Ç 2 ÈÇ 3 È , Ç 2 ÈÇ 4 È and Ç 3 ÈÇ 4 È , thus generating ten independent code aggregates. The construction stream repeated the same steps by adding the other two complement code tiles Ç 1 È and Ç 2 È to gene rate another ten distinct code aggregates. Then, the stream rotated every code aggregate and applied the interchangeability rules to group all the symmetric code aggregates, that code for the same amino aggregate . As was expected, the construction stream b uilt twenty different groups, each group representing one amino aggregate, [42] , and [64] . While the construction stream was building the code aggregates, the express ion stream was drawing a representation of every code aggregate in the code table. The result was a symbolic mosaic that includes all twenty groups of code aggregates, as shown in Table 4 .

PAGE 81

! #' ! Table 4 . The c ode m osaic g enerated by the e xpression s tream as Shown in [42] , [7] . The observation stream collected the known inputs about the twenty amin o aggregates and saved them in the amino set. Consequently, the expression stream saved the names of those amin o aggregates in the amino table . Even though not all names of amino aggregates were confirmed to be part of the protein mosaic, the observation s tream still added them into the amino set. Then, the observation stream, having a virtual model of an existing DNA information mosaic, triggered a thought experiment to construct the intermediate mosaic from the genetic code in the DNA mosaic. The observat ion stream passed that experiment with all associated inputs to the construction stream. 6.3.9 Constructing the Intermediate Mosaic The construction stream started a thought experiment to build the intermediate mosaic by reading the genetic code in the virtual i nput model. For a better explanation, assume that the ! 3 ! 4 ! 1 ! 1 ! 3 ! 4 ! 2 ! 2 ! 3 ! 4 ! 3 ! 3 ! 3 ! 4 ! 4 ! 4 ! 3 ! 4 ! 2 ! 1 ! 3 ! 4 ! 3 ! 1 ! 3 ! 4 ! 4 ! 1 ! 3 ! 4 ! 3 ! 2 ! 3 ! 4 ! 4 ! 2 ! 3 ! 4 ! 4 ! 3 ! 1 ! 2 ! 1 ! 1 ! 1 ! 2 ! 2 ! 2 ! 1 ! 2 ! 3 ! 3 ! 1 ! 2 ! 4 ! 4 ! 1 ! 2 ! 3 ! 1 ! 1 ! 2 ! 4 ! 1 ! 1 ! 2 ! 2 ! 1 ! 1 ! 2 ! 3 ! 2 ! 1 ! 2 ! 4 ! 2 ! 1 ! 2 ! 4 ! 3

PAGE 82

! #= ! genetic code of the DNA mosaic is 123123... . The construction stream started to read the genetic sequence following the reading rules. It read the first three letter tiles. With no restriction in direc tion, the construction stream chose to read the code from left to right. Thus, the first successive input aggregates the construction stream read were 123. The stream first generated the body diamond tile of the code aggregate. It then added the first lett er tile Ç1È at the top corner of the body tile. Following the local interchangeability rules, the construction stream could add the first letter tile Ç1È arbitrarily in any corner of the body tile. Next, the construction stream read the input aggregate a nd placed the corresponding letter tile Ç2È to the left of the tile Ç1È. The next letter tile to be added must be complementary to the Ç2È tile, thus the construction stream added Ç2`È to the right of the tile Ç1È and in front of the Ç2È tile. In the last corner of the body tile, the construction stream placed the complement of the last input aggregate. The letter tile in the empty corner is determined by the complementarity matching rule. Reading the input aggregate Ç3È, the construction stream added Ç3`È to the empty corner of the body tile. Once the first code aggregate is built, the stream inserted it into its corresponding cavity, as shown in Figure 14 , [42] , and [64] . Then, using overlapped reading, the construction stream retrieved two input aggregates and read the next three successive input aggregates, 231 , and repeated the steps to produc e the second code aggregate. After creating t he second aggregate, the stream inserted it into the cavity to the right of the previous one. The construction stream continued until the whole genetic sequence was read, see Figure 15 .

PAGE 83

! #> ! Figure 15 . Part of the intermediate mosaic constructed form the genetic sequence "123123" , [7] . The Gamow's thought experiments were the first thought experiments to state the genetic problem as an information problem instead of a biochemical pro blem. During this run of the AD, the observation stream established the basis that any following AD run should follow to find the correct solution. To decipher the genetic code, the AD must first correctly identify the input and output mosaics. Then, the AD, manifesting visual streams, must discover the transformation rules needed to transform the input mosaic into the output one. These transformation rules must include the correct definition of the reading rules that control the reading of the input mosai c. This reading is needed to build the intermediate mosaic. The intermediate mosaic was defined, for some time, as a long sequence of code aggregates; each aggregate is built out of at least three letter tiles. Every letter tile in the code mosaic has a co rresponding letter tile in the input mosaic. In addition to the definitions of the input, output , and intermediate mosaics and the transformation rules, the AD must build the code and amino sets. The code set must include all code aggregates and the am ino set must contain all amino aggregates needed to built the intermedia mosaic and the peptide mosaic , respectively. 6.4.0 The Crick Thought Experiments 6.4.1 Validating the Diamond Code In an effort to validate the diamond code, the AD initiated the observation str eam. From the impression stream, the observation stream received the structure of the insulin peptide mosaic. Having this structure, the observation stream began a thought experiment to verify the correctness of the diamond code, [34] . Insulin is the peptide protein mosaic that is constructed ! 1 ! 3` ! 2` ! 2 ! 2 ! 1` ! 3` ! 3 ! ! 3 ! 2` ! 1` ! 1 ! 1 ! 3` ! 2` 2

PAGE 84

! ## ! out of fifty one amino aggregates, [41] . The observation stream sent the structure of insulin along with the results of Gamow' s experiment for evaluation to the validation stream. The validation stream chose two different insulin mosaics whose structures were fully known: insulin B and B corticotropin. Then, it initiated the construction stream to apply the diamond code to recon struct the insulin mosaic. While studying the insulin B mosaic, the construction found the aggregate sequence ÇLeu.Tyr.LeuÈ. The construction stream also found the sequence ÇSer.Tyr.SerÈ in B corticotropin. The construction stream reasoned that these two sequences are of type ÇxyxÈ. The ÇxyxÈ sequence is a peptide mosaic containing three consecutive amino aggregates, where the first and last aggregates are the same and the middle one is different. The stream began to apply the diamond code to create an int ermediate mosaic that codes for any ÇxyxÈ sequence, see [7] , and [34] for details . In order to illustrate the steps followed by the construction stream, consider that ÇxÈ is represen ted by the code Ç111È and Ç112È represents ÇyÈ. Since the reading rules require the overlapped reading of the diamond code, the code sequence of ÇxyÈ must be Ç1112È. Figu re 16 (left and middle) shows the code aggregates that corr espond to the amino aggregates ÇxÈ and ÇyÈ , respectively. The third amino aggregate in the sequence ÇxyxÈ, ÇxÈ, must be coded from the input sequence Ç111È or Ç1`1`1`È , since both code sequences generate the same code aggregate. However, following the ove rlapping reading requirement, the next reading is Ç12nÈ, where ÇnÈ is any letter tile. This reading would generate amino aggregates different than ÇxÈ, see Figu re 16 (right). The construction stream failed in all attempts to appl y the diamond code to generate the intermediate mosaic that codes for the sequence ÇxyxÈ. This failure led the validation stream to invalidate the diamond code, [7] , and [34] .

PAGE 85

! #$ ! Figu re 16 . The code aggregates generated from the input sequence 1112n . 6.4.2 Redefining the Input Mosaic Once the diamond code was eliminated, the AD started a new cycle to look for a solution to the genetic code problem. The observation stre am initiated a thought experiment to find the correct input mosaic , which includes the genetic information. In the preliminary stage, the observation stream gathered any known inputs about the genetic code. The first input showed that the sequence of amino acids is determined by genes. The constructed sequence of amino aggregates represents the protein peptide mosaic. The stream received the assumption that the DNA backbone's base sequence carries the genetic code. The observation stream continued receiving new inputs about the cell nucleus' components. Cells of human beings include DNA, RNA, and cytoplasm. The stream collected information about the RNA molecule . It found that RNAs are single strand structured molecules that are synthesized from one strand o f the DNA in the nucleus of the cell. The cytoplasm of a cell is the fluid inside the cell that surrounds its contents. The observation stream also received information that proteins are synthesized in the cytoplasm and that the DNA is located only in the cell's nucleus. This led the observation stream to reason that a copy of the DNA must be created in the nucleus and sent to the cytoplasm. Analyzing these inputs, the observation stream concluded that the RNA mosaic contains the genetic information, by co pying it from the DNA, to construct the protein, [7] , and [45] . Further, the observation stream divided protein synthesis into two phases: transcription and translation. During transcription, the DN A mosaic is read to construct the RNA. Since the length of RNA is much shorter than that of DNA, the observation stream concluded that the genetic ! 1 ! 1` ! 1` ! 1 ! 1 ! 2` ! 1` ! 1 ! ! 1 ! n` ! 2` ! 2

PAGE 86

! #% ! information inside the DNA mosaic codes for more than one protein mosaic, and that each segment of the code c orresponding to one protein mosaic would be copied into one RNA mosaic. The stream reasoned that the resulting RNA mosaic from the transcription phase would be used later as an input mosaic to the translation phase in order to create one protein mosaic, [7] , [50] , and [37] . Having a new definition of the input mosaic, the observation stream ran a thought experiment to find the correct struc ture of this new mosaic. The stream had inputs stating that the RNA mosaics are built out of four different RNA base tiles, A, U, G and C. The stream reasoned that during transcription, the RNA mosaic is constructed by reading the base tiles of one DNA str and and transformed them into their corresponding base tiles in the RNA mosaic. Each one of the DNA base tiles, A, T, G and C, represents one of the RNA base tiles as follows, A to A, T to U, G to G and C to C , as shown in Figure 17 . The observation stream defined the input mosaic to be an abstract copy of the RNA mosaic after removing all chemical clutter. The stream defined every base tile in the RNA mosaic as a letter tile in the input mosaic (A, U, C, and G). These letter tile s are linked together as a long sequence that , if read correctly, provides guidance to the engine of protein synthesis , while building the output mosaic, [7] , [45] , [50] , and [37] .

PAGE 87

! #& ! Figure 17 . Constructing the RNA mosaic from a single strand of the DNA mosaic . The observation stream also studied the nature of the output mosaic. With no new inputs that contradicted Gamow's definition of the output mosaic, the observation stream kept the same definition of the output mosaic described in Section 6.3.3 . 6.4.3 Evaluating the Reading Rules Next, the observation stream i nvestigated the reading rules defined during Gamow's experiment , [7] . The stream did not find any new inputs that invalidated the triple reading. Therefore, the observation stream accepted the triple reading of the genetic code as defined in Section 6.3.6 . The observation stream also investigated the overlapped reading. In Gamow's experiment, this reading method was concluded to be based on the 3D space similarities between the protein mosaic's aggr egates and the DNA mosaic's aggregates. The observation stream excluded this conclusion based on the new assigning of the input mosaic as the RNA mosaic instead of the DNA mosaic. However, the observation stream reasoned that even though the input mosaic o f the translation phase is the RNA mosaic, this mosaic could still be mapped one to one to its

PAGE 88

! $( ! corresponding genetic portion of the DNA mosaic. That led the observation stream to consider the DNA mosaic as the initial step of several transformation steps i n building the peptide mosaic. Having no new evidence that contradicted overlapped reading and to maximize the density of information storage, the observation stream accepted the overlapped reading of the input mosaic, see [7] , and [33] for details . Finally, the observation stream examined the bidirectional code reading. The stream searched for any known input mosaic that could generate two different sequences of the peptide mosaic. T hese two sequences must be reversed, such as ÇThr.Pro.Lys.AlaÈ and ÇAla.Lys.Pro.ThrÈ, which could be a result of reading the genetic code in both directions. The observation stream reasoned that bidirectional reading means that the two resulting peptide mo saics should be distinguishable by nature, but no valid inputs support ed this hypothesis , [7] , and [34] . This drove the observation stream to investigate the way letter tiles could b e grouped into code aggregates. To find the answer, the observation stream looked to the x ray pictures of the backbone of DNA. These pictures showed no regularity in the DNA backbone to conclude how the letter tiles are grouped together. The stream also r ealized that the triple reading of the code would create various possibilities for an incorrect reading of the code. At the end of this thought experiment, the observation stream found lab evidence proving that the code could only be read in one direction . This evidence guided the observation stream to reject the bidirectional reading of the code and hypothesis to conclude that the genetic code must be read from one direction only, [7] , [34] , and [35] . Once the bidirectional reading was invalidated, the observation stream discarded the symmetric feature of the code aggregates resulting from the bidirectional reading rule. From the conclus ions of the Gamow experiment, the code aggregate is symmetric because it can be placed

PAGE 89

! $) ! into the input cavity from any side. However, the results of the other thought experiments, as inputs to the observation stream, found that the DNA structure has bonds t hat stick out perpendicularly to its axis and that the DNA structure's surface has knobs on it. The stream found different aspects of the DNA's two sides. It concluded that placing code aggregates inside the cavity of the DNA would not be possible, [7] , and [34] . 6.4.4 Rebuilding the Amino Set Next , the observation stream started a thought experiment to define the twenty amino aggregates within the amino set. The stream received the amino s et that was initiated in Gamow's experiment. Since Gamow's amino set was assumed with no solid evidence, the observation stream reasoned that the correct definition of the amino set is critical to accurately deciphering the genetic code. Moreover, even bef ore the code is deciphered, knowledge of the amino set would accelerate constructing the correct mapping between the amino aggregates and code aggregates. Essentially, defining the c orrect amino set would minimize the time needed to crack the genetic code, [7] , and [34] . In order to build the correct amino set, the observation stream utilized current information about amino acids in protein. Inputs showed that the same amino acids are used to construct proteins in most living creatures. Also, the DNA base tiles in all organisms are built out of the same set of bases. These similarities guided the stream to conclude that the genetic code is universal, which means that the genetic code in most living organisms would be deciphered following the same transformation of DNA code aggregates into amino aggregates. The stream identified the amino set by investigating several known protein chains. The stream eliminated all amino acids never foun d in any protein chain, such as Norvaline and Hydroxy Glutamic. It also

PAGE 90

! $" ! excluded the amino acids that are seen in small peptides. In contrast, the stream added all of the amino acids that are rarely seen in the protein chain, such as Asparagine and Glutami ne. Using all known inputs about amino acids, the observation stream saved, in the amino set, twenty amino aggregates, [7] , [33] , and [34] . This set was translated by the expression stream into the amino table, which was used as an input for other thought experiments and was eventually confirmed to be correct. The Crick thought experiments did not provide any description of a coded algorith m that would be used to code the RNA mosaic into a peptide mosaic. However, these thought experiments resulted in defining the amino set that was proved correct in later thought experiments. 6.5.0 Modifying the Reading Rules By receiving new inputs that forced t he AD to modify the reading rules, the AD triggered the observation stream to start a thought experiment to investigate the correctness of the genetic code reading rule . The stream, in its preliminary stage, retrieved all the conclusions coming from Crick' s thought experiments and used the same definition of the input and output mosaics. These thought experiments concluded that the code should be read in triples, overlapped, and unidirectional. The observation stream looked at new lab inputs that changed the genetic structure of virus RNA by treating it with nitrous acid. The modified RNA is then read to construct the peptide mosaic. The result of this construction is a new virus that has a similar genetic sequence to the original virus , but differs in on ly one amino acid. Having these results, the observation stream questioned the overlapped reading of the genetic code. If the code overlapped, modifying one base tile in the RNA sequence would affect at least two amino aggregates in the output mosaic,

PAGE 91

! $' ! but only one amino aggregate was affected . Therefore, the observation stream concluded that the reading of the genetic code must be non overlapped, [7] , and [36] . Once the observation st ream eliminated overlapped reading, it reasoned that the mapping between the input and output mosaic s is one to one. There was no need to construct any intermediate mosaic between the input and output mosaics. When building the intermediate mosaic from the RNA mosaic using non overlapping reading, the resulting intermediate mosaic would be identical to the RNA mosaic and , therefore , not necessary. At the end of this thought experiment, the observation stream saved, into the construction set, the new reading rules for the genetic code. These rules stated that the reading of the genetic code must be triple, non overlapped, and unidirectional. In parallel, the expression stream saved the reading rules into a shell that include d a description of these rules. T hi s shell will be used as input for any future thought experiments. 6.6.0 Deciphering the First Triple The Crick thought experiment identified the input mosaic to be the RNA mosaic. However, there are many types of RNA. Therefore, a thought experiment was construc ted to determine if all RNA types are used dur ing protein synthesis and what their roles were in the process. The observation stream received lab inputs that observed an RNA mosaic labeled with an amino aggregate, tRNA. After labeling, tRNA participates in protein synthesis by carrying amino aggregates to be added to the peptide mosaic, [7] , and [56] . The observation stream reasoned that tRNA mus t act as an adapter mosaic. This mosaic carries the amino aggregates during translation. Since tRNA is an RNA mosaic, and all RNA mosaics are constructed by replicating the base tiles in the tRNA, the observation stream identified the basic structure of the tRNA to be the same as the one define d in the Crick's

PAGE 92

! $= ! thought experiments. However, the final structure of tRNA, as seen in x rays, wasn't known at that time, therefore, the observation stream concluded that a visual model of the tRNA mosaic could not be built. The stream found that the role of the tRNA is auxiliary: it is only needed for carrying amino aggregates during translation. The construction stream could visualize any arbitrary structure of tRNA and link it to an amino aggregate, and it would still participate in the translation proce ss, [7] , and [58] . 6.6.1 Defining the Input Mosaic There was no proof of the input mosaic for translation. Thus, the AD started its run to find the correct definition of the input mosaic. It triggered the observation stream to begin a series of thought experiments to solve the genetic code problem. The observation stream first initiated a thought experiment to define the input mosaic. The observation stream accepted the results of the two lab experiments that tested the nature of the input mosaic. Both experiments used a lab extract containing the mosaics involved in translation, i.e. , all different types of the tRNA mosaics and the amino aggregates. For the first lab experiment, the DNA mo saic was added to the extract to be the input mosaic. In the other lab experiment, the input mosaic added to the extract was a special type of RNA, messenger RNA (mRNA). The results showed the existence of a peptide mosaic only in the extract containing th e mRNA mosaic. With these results, the observation stream concluded that the input mosaic for translation is the information mosaic that is constructed from the mRNA mosaic, [7] , [55] , and [56] . In order to build a visual model of the mRNA information mosaic, the observation stream declared that this mosaic is an abstract copy of the mRNA mosaic. Since the mRNA mosaic is another type of an RNA mosaic, the stream defined the tiles of the mRNA to be the four RNA base tiles (A, U, C, or G). Each of these base tiles would be converted into the corresponding

PAGE 93

! $> ! letter tiles in the mRNA information mosaic, after removing all the chemical clutter. The reading rules o f the genetic code were confirmed to be triple, non overlapped, uni directional reading. The observation stream knew that every three consecutive letter tiles would be grouped together to build one code aggregate. Every code aggregate in the mRNA mosaic wo uld code for one specific amino aggregate. The length of the tRNA mosaic would be the same length of the gene area in the DNA mosaic. The gene area is the sequence of base tiles in the DNA mosaic that code for a single protein mosaic. 6.6.2 Generating the "Const ruct information mRNA" Shell Following the new definition of the input mosaic, the expression stream modified the input mosaic construction shell. The expression stream tagged the actions of the construction stream while it was building the basic structure of the mRNA. It generated the first instruction aggregate: " Build (Basic mRNA structure, construction set). " The expression stream also saved the set of the mRNA letter tiles as declared by the observation stream in a one dimensional symbolic mosaic, as shown in Table 5 . Table 5 . One dimensional mosaic of the mRNA letter tiles converted from the DNA base tiles . DNA Base tile mRNA letter tile The expression stream saved the steps follow ed by the construction stream in the "Construct information mRNA" shell. The first instruction aggregate saved in this shell is: " Read (DNA strand, reading rules). " The reading rule identified by the observation stream and translated by the expression stre am for reading the DNA strand is "Read one DNA strand tile by tile." The ! E ! ! H ! ! F ! ! G !

PAGE 94

! $# ! expression stream then marked the conversion from the base tile in the DNA to the corresponding letter tile by following the letter tiles mosaic, in Table 5 . The expression stream created the instruction aggregate: " Convert (DNA base tile, mRNA letter tile, letter tiles mosaic) " and saved it in the current shell. Next, the instruction aggregate: " Add (letter tile, mRNA, matching rules) " was also saved to the s hell. The expression stream reasoned that the same instruction aggregates would be repeated starting from the instruction aggregate Read until all base tiles of the current gene area in the DNA strand are scanned. The final "Construct information mRNA" she ll can be seen in Figure 18 . Figure 18 . "Construct information mRNA" s hell g enerated by the e xpression s tream . 6.6.3 Translation Components After defining the input mosaic, the observation stream identified the components that engaged in translation: the mRNA input mosaic, the peptide output mosaic, the tRNA adapter mosaic, the code and amino sets, and the reading rules. In order t o generate the peptide mosaic, the construction stream read one code aggregate in the mRNA, tile by tile, and called the tRNA that carries the corresponding amino aggregate. The incoming amino aggregate would be added to the growing peptide mosaic. The construction stream must read the input mosaic in one direction only. Moreover, previ ous thought experiments showed that the reading of the code is non overlapped, i.e., triple after triple. Since the lab results showed that mRNA, tRNA, Construct information mRNA Choose (gene, Set of DNA mosaic) Convert (base tile, letter tile, letter tiles table) Read (DNA strand, reading rules) Add ( letter tile, mRNA, matching rules) Repeat ( gene length, L ) L:

PAGE 95

! $$ ! ribosome, and amino are the mosaics needed during translation, the observation stream reasoned that all of the components involved in translation were identified. It concluded that the next step in deciphering the genetic code is to map between the code aggregates of the mRNA and the amino aggregates in the peptide mosaic, i.e, to find the correct mapping b etween the code and the amino sets, [7] , [55] , and [36] . 6.6.4 The Mapping Thought Experiments In order to map correctly between the code and amino sets, the observation stream reasoned that mapping s between a known structure of mRNA and peptide mosaics must be found. Since the mapping s between the two mosaics are one to one, if an mRNA with known code aggregates was transformed into a peptide mosaic with known amino aggregates, the meaning of every code aggregate in the mRNA would be known, i.e. , the amino aggregate that each code aggregate represents. To apply this reasoning, the observation stream called for twenty lab experiments to be performed. Th e se lab e xperiments dealt with a cell free extract that contains a poly U mRNA mosaic, the twenty amino aggregates in the amino set, and twenty tRNA mosaics (one for every different amino acid). The stream found that, having an mRNA mosaic with the same code sequen ce, e.g. UUUUU..., would result in a peptide mosaic of all of the same amino aggregates. In the lab, in order to see the resulting peptide mosaic, its amino aggregates must be radioactive. Since the peptide mosaic is constructed out of a series of the sa me amino aggregate and there are twenty possible amino aggregates to construct the peptide mosaic, twenty lab experiments were prepared. Each of them included only one radioactive amino acid , which differed in every lab experiment. The stream concluded tha t the extract having the radioactive amino aggregate, which corresponded to the code aggregate, would have a radioactive peptide mosaic.

PAGE 96

! $% ! To visualize those lab experiments, the observation stream triggered a thought experiment to build the poly U mRNA mos aic and passed the experiment to the construction stream. Using the construction rules for building the input mosaic, Section 6.6.1 , the construction stream built the input mosaic of poly U, as seen in Figure 19 . Figure 19 . An mRNA m osaic of Poly U c ode a ggregates , [7] . In one of the above lab experiments, which included radioactive phenylalanine, the resulting peptide mosaic was also radioactive. T he observation stream initiated the construction stream to build a poly phenylalanine mosaic based on the construction rules for building an information peptide mosaic. These construction rules c a me from previous thought experiments. A protein peptide mos aic is the abstract representation of the protein mosaic built out of amino aggregates that are linked to each other in a sequential line, [7] , [45] , [50] , and [37] . After building the input and output mosaic, the construction stream mapped the code aggregates in the mRNA to the amino aggregates in the peptide mosaic. The observation stream found that every code aggregate of ÇUUUÈ was mapped to one amino aggregate ÇphenylalanineÈ, thus it concluded that the code aggregate ÇUUUÈ represents the amino aggregate ÇphenylalanineÈ. Figure 20 shows the mapping between the Poly U mRNA mosaic and the resulting peptide mosaic after translation, [7] , [55] , and [56] . Following the same technique, the observation stream concluded that ÇAAAÈ, ÇCCCÈ and ÇGGGÈ code aggregates represent the amino aggregates ÇlysineÈ, ÇprolineÈ and ÇglycineÈ, respectively. ! UUU UUU UUU Code agg regates ! UUU UUU

PAGE 97

! $& ! Figure 20 . The m apping b etween the c ode and a mino a ggregates , [7] . By the end of this AD run, not a ll code aggregates were mapped to their corresponding amino aggregates. However, the AD identified the process rules needed to map between the code and amino sets. First, the AD must identify the structure of the mRNA mosaic. Then, it observes the results from the lab experiments to find the correct structure of the resulting peptide mosaic. After identifying the peptide mosaic, the construction stream must map the code aggregates to the respective amino aggregates. By following the above process rules, su bsequent thought experiments managed to map other code aggregates to their amino aggregates. After many rounds, the AD deciphered all sixty four code aggregates. The AD also identified two special sets of code aggregates, the initiation set and the termina tion set. Both sets are part of the scope sets that help in eliminating the run time of the future thought experiments. The initiation set includes all start code aggregates that trigger the building of the peptide mosaic. The initiation set included only one start code aggregate, ÇAUGÈ. The termination set contains three end code aggregates, ÇUAAÈ, ÇUAGÈ or ÇUGAÈ, that signal the termination of the translation process, [7] , [55] , and [56] . Every time the discovery visual streams decipher one code, the expression stream saved the results into a "Genetic code" symbolic mosaic. This mosaic includes the meaning of every code aggregate and was built as a result of multiple thought experiments. The output of the expression stream was a symbolic mosaic that included the full genetic code, Table 6 . ! UUU UUU UUU UUU UUU Peptide mosaic Poly U mRNA mosaic ! Phe. Phe. Phe. Phe. Phe.

PAGE 98

! %( ! Table 6 . The f inal g enetic c ode m osaic 3 , [37] . First letter tile ! Second letter tile ! Third letter tile ! U C A G U UUU " Phe UUC " Phe UUA " Leu UUG " Leu UCU " Ser UCC " Ser UCA " Ser UCG " Ser UAU " Tyr UAC " Tyr UAA " Stop UAG " Stop UGU " Tyr UGC " Tyr UGA " Stop UGG " Stop U C A G C CUU " Leu CUC " Leu CUA " Leu CUG " Leu CCU " Pro CCC " Pro CCA " Pro CCG " Pro CAU " His CAC " His CAA " Gln CAG " Gln CGU " Arg CGC " Arg CGA " Arg CGG " Arg U C A G A AUU " Ile AUC " Ile AUA " Ile AUG " Met ACU " Thr ACC " Thr ACA " Thr ACG " Thr AAU " Asn AAC " Asn AAA " Lys AAG " Lys AGU " Se r AGC " Ser AGA " Arg AGG " Arg U C A G G GUU " Val GUC " Val GUA " Val GUG " Val GCU " Ala GCC " Ala GCA " Ala GCG " Ala GAU " Asp GAC " Asp GAA " Glu GAG " Glu GGU " Gly GGC " Gly GGA " Gly GGG " Gly U C A G 6.7.0 Results S everal thought experiments involved in cracking the g enetic code were identified. In each of them, a major role of mosaic reasoning was identified in finding the solution to this problem. Tiles and aggregates, local and global matching rules, information mosaics and an unstructured environment led the AD tow ards the correct solution. New components of mosaic reasoning, with an essential role in the discovery, were revealed that could be seen in any discovery of the same nature. A new type of mosaic, the information mosaic, was declared. Information mosaics ar e critical to deciphering the genetic code , because the genetic code is a special type of information. In addition, four different scope sets were identified: the initiation, termination, !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! !!!!!!!!!!!!!!!!!!!! ! 3 In this ta ble, the genetic code consists of letter tiles; three of these letter tiles represent a code aggregate that is mapped to one amino aggregate. For example, the UUU code aggregate is mapped to the Phe. amino aggregate.

PAGE 99

! %) ! code, and amino sets. Each of these sets was essential in cracking t he genetic code and reducing the time required to find the correct solution. Analysis also revealed new components of the transformation rules, the reading rules and the process rules. Reading rules will be used in any discovery that requires the reading of an information mosaic. These rules control the reading of one information mosaic to construct another mosaic , based on the read information. The process rules were identified to control the actions which must be applied by the AD to map the code and ami no sets. Following these rules allowed the deciphering of every code aggregate in the code set. Two new types of transformation rules were identified. These new rules are involved in the two phases of the protein synthesis process: transcription and trans lation. The readi ng algorithm was also revealed and applied during the translation to read the mRNA mosaic and convert it into the peptide mosaic. In the next chapter, the work on translation is continued with the goal of defining the full process of trans lation.

PAGE 100

! %" ! CHAPTER VII REVISITING THE DISCOVERY OF PROTEIN TRANSLATION 7.1.0 Overview As mentioned in Section 4.4.0 , the translation process is the second phase of protein synthesis. During translation, the genetic code in the mRNA i s read and transformed into a sequence of amino aggregates. These aggregates are attached to each other to form the preliminary structure of the protein mosaic, i.e. , the peptide mosaic. The AD identified the fact that different mosaics participate in the translation. They are the tRNA, mRNA, and ribosome mosaics. The AD also recognized the output of translation as the peptide mosaic. The construction of each was emphasized by mosaic reasoning , where the tiles, aggregates, and the matching rules are declar ed and saved in the construction set. Based on incoming inputs, t he AD divided the translation process into three smaller processes: initiation, elongation, and termination. These three processes run sequentially starting with the initiation process, wher e the ribosome links to the mRNA. Once the ribosome is attached to the mRNA, it will scan the mRNA to read the genetic code. According to the genetic reading, the ribosome will compare the current mRNA's codon (code aggregate) to all the codons belonging t o the initiation set. If the ribosome finds a match, then it will end the initiation and enter the elongation process . During elongation, the ribosome continues to read the mRNA's codon and calls for the tRNA that carries the corresponding amino aggregate . Every incoming amino aggregate will then be added to the growing peptide mosaic. Elongation ends when the ribosome read s a termination codon. Once this codon is found, the translation process will end by releasing, or detaching, all of the mosaics, use d by the process, including the new peptide mosaic.

PAGE 101

! %' ! In parallel with the AD's discovery streams 4 , the expression stream runs to register and tag the actions of every discovery stream. The expression stream saves these tags into a symbolic shell that will be used to build part of an algorithmic mosaic. At the end of an AD round, the expression stream will group all the parts of the algorithm mosaic and the result is an algorithmic mosaic that describes the algorithm defined to solve the problem, i.e. , here, to describe translation process. In the next sections, the major thought experiments and all of the AD components that led to the discovery of protein translation and the associated algorithmic mosaic are described. 7.2.0 Motivation After revisiting the genetic code discovery, the next step was to analyze the thought experiments that identified the translation process. The main goal was to emphasize the components of the AD that were found in the previous chapters and discover new components that guide discoveri es of the same nature as the genetic code discovery. 7.3.0 Applying the AD on the Discovery of tRNA Activation By the end of this discovery, the AD generated the transformation algorithmic mosaic needed to activate a tRNA mosaic, i.e. , to transform the tRNA mosa ic into an amino tRNA mosaic. This step is the first step towards revealing the complete structure of the algorithmic mosaic that controls the translation process. In this section, three thought experiments done by two different scientists were followed. In Section 7.3.1 , the thought experiments performed by Crick, [34] , were analyzed. In Sections 7.3.2 and 7.3.3 , the AD was applied to the thought experiments based on the work of Hogland, [53] , who described in detail how he reached this discovery. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! !!!!!!!!!!!!!!!!!!!! ! 4 The observation, construction, and va lidation streams .

PAGE 102

! %= ! 7.3.1 The Adapter Hypothesis Thought Experiments The AD reasoned that during protein synthesis the amino aggregate must be aligned, according to the reading of the genetic code, in order to construct a protein peptide mosaic. The AD started this thought experiment to find the rules that must be followed to specify the sequence of amino acids based on the genetic code of the DNA mosaic, i.e. , to find the mechanism of protein translat ion. In order to find the relationship between the genetic code and the amino acid, the observation stream investigated the chemical structure of the amino aggregates and the nucleotide aggregates in the DNA or RNA mosaics. Nucleotide aggregates represent the sequence in the DNA, or RNA, that possibly might include the genetic code , [34] . During the preliminary stage, the observation stream collected all known inputs about the general structure of the amino aggregate in the pro tein mosaic and the nucleotide aggregate in the DNA mosaic. The observation stream then analyzed these inputs. It did not find any possible complementarity rule that can be applied to bind the amino aggregate and the nucleotide aggregate directly. Looking to the amino aggregate, the observation stream did not find any knobby surfaces that might be used by the DNA or RNA mosaics to distinguish between the different types of amino aggregates. In other words, nucleic acids, i.e. DNA and RNA, are built out of a string of nucleotides whose chemical structure helps the nucleic acids recognize each other , but they cannot distinguish the varied side chain of the amino acids. Moreover, the observation stream realized that amino aggregates have no charged group that could determine the position in which they would link to the nucleotides. The DNA and RNA structure only shows patterns of hydrogen bonds. Therefore, the observation stream excluded the possibility that the nucleotide sequence, in DNA or RNA, can recogniz e the amino aggregate

PAGE 103

! %> ! side chains, i.e., the chemical structure of both DNA and RNA cannot directly bind to amino acids to form the polypeptide mosaic. From the above, the observation stream concluded that there must be an intermediate mosaic, possibly a small one, that would have the correct chemical structure to be linked to a side of the proper amino aggregate. The other side of this mosaic should be linked to the nucleotide in the nucleic acids. This link is guided by a global complementarity rule. Th is rule requires that the hydrogen bonding surface of this mosaic be complementary to the hydrogen bond in the nucleotide aggregates inside the DNA or the RNA mosaic. The stream named this intermediate mosaic as the adapter mosaic. The observation stream r easoned that since there are twenty different kinds of amino acids , there could be, at least, twenty different kinds of special enzymes, one for each amino acid. Each enzyme would help in attaching the amino acid to a specific adaptor mosaic. Hence, t here are 20 adaptor molecules, one for each amino acid. In order t o verify the adapter theory, the validation stream started analyzing of all the known inputs about a small molecule found in living cells, however, none was found to fit the adapter hypothesis. Therefore, the validation stream postponed the validation of this theory while waiting for new inputs. 7.3.2 Identifying the Process of Activating Amino Acids The observation stream began this thought experiment by receiving the outputs of different thought exp eriments, but not the conclusions of the thought experiment in Section 7.3.1 . The observation stream analyzed all current information to identify the process in which the amino aggregates can be used to build the protein mosaic , i.e. , the translation of the protein. Initially , the stream found that the inputs dealing with the construction of the final product of protein

PAGE 104

! %# ! synthesis, the peptide chain, suggested that amino acids must be funneled with energy to be able to link to ea ch other. The chemical structure of the known amino acids shows that they do not have energy bonds, i.e., in their current state, they cannot have the energy needed to link to each other. In the analysis stage, the observation stream reasoned that this ene rgy must be donated to the amino acids by another molecule before starting the protein synthesis. Thus, the stream needed to reveal the transformation rules that control the process of converting the amino aggregate into an active amino aggregate, i.e. , an amino aggregate with the energy needed to link to another amino aggregate , [53] . In order to find the correct transformation rules, the observation stream f ocused itself by receiving any inputs, through the impression stream, that related to any energizing molecules. The observation stream received the output of the Lipmann experiments, [39] , which found that acetyl phosphate is produced in bacteria during the process of oxidizing carbohydrates. A cetyl phosphate can donate its acetate part to provide energy to non energized molecules. In living organisms, tissues constructing any molecule need energy. However, acetyl phosphate was confirmed as a source of energy only in bacteria. The result of the Lipmann experiments forced the AD to focus its streams on finding out if other molecules could be energized given that acetate can be energized by phosphate in bacteria. The observation stream followed many lab experiments that looked at molecules that can be linked to acetate. The stream received information about the acetyl coenzyme "acetyl CoA." In the lab, acetyl CoA was found to be a general form of the active acetate, "acetic acid", in all living tissues. The acetate aggregate in acetyl CoA is a struc ture possessing a form of constructing energy that is used in the process of making different molecules . Figure 21 shows the structure of the acetate aggregate retrieved from the construction stream of a previous thought experimen t , [53] .

PAGE 105

! %$ ! Figure 21 . The i nput a ggregate: A cetate . The acetyl coenzyme (acetyl CoA) aggregate is built out of two smaller parts, CoA S, and the acetic. Since the chemical structure of CoA is not importa nt for this discovery, the observation stream started erasing the particulars and identified the whole molecule as one tile, CoA. The lab results showed that CoA is first linked to a sulfur atom (S) to form the CoA S aggregate, thus the observation stream identified the first step in creating the acetyl CoA as the linkage between the CoA and S tiles. The observation stream triggered the construction stream to build the acetyl CoA. The construction stream found chemical inputs confirming that the CoA S aggr egate and acetate aggregate are linked together from the COOH aggregate of acetate and the (S) tile of the CoA S aggregate. This happens when the acetate aggregate loses its OH aggregate and the Co A S is bound together following the global complementarity rule. The result is the acetyl CoA aggregate , as seen in Figure 22 . Figure 22 . The i nput a ggregate: A cetyl CoA . Further lab inputs have shown that during the process of creating the acetyl CoA, energy from ATP (adenosine triphosphate) is required. ATP is the structure possessing the basic form of

PAGE 106

! %% ! chemical energy of cells and its chemical structure is Ad.P~P~P (~ denotes a high energy bond). The observation stream identified the ATP aggregate consisted of tw o tiles, Ad and P. Ad is connected to one phosphate tile, P, from its right side. Two more phosphate tiles is linked to the Ad.P aggregate through a (~) complementarity bond to construct the ATP aggregate. The observation stream saved the construction rul es to build the ATP in the construction set to be used later and continued looking at the incoming stream of inputs to find information related to the problem. The observation stream received other information, which explained the process of how the active acetate is linked to the acetate acceptor, called sulfanilamide. This information confirmed that the linkage occurs by first unplugging the "OH" aggregate from the acetate. Next , the acetate "CO" aggregate and the sulfanilamide "NH2" amino aggregate link together following the global complementarity rules, which results in a CO NH peptide bond similar to the amino acid peptide mosaic, see Figure 23 . Figure 23 . CO NH peptide bond in peptide mosaic retrieve d from a previous thought experiment . Other information obtained by the observation stream was related to the bacteria in lab experiments that use ATP to activate the pantothenic acid molecule and link it to # alanine, a special amino acid. Again, the linkage is a peptide bond identical to the one between protein amino acids shown in Figure 23 . The observation stream then created a general representation of the amino aggregates using the approach of erasing the particulars . This was accomplished by replacing different chemical

PAGE 107

! %& ! groups of amino acids with the symbol R , while keeping the NH 2 CH COOH group that is common to all of the amino aggregates. The amino aggregate is the generator that would eventually be used to build the general shape of the protein mosaic. In order t o transform the generator into a specific amino aggregate , the strea ms must use the amino set that includes 20 aggregates of different amino groups, one for each amino acid. Having all of the above inputs , the observation stream attempted to apply the principle of analogy of the two visual streams. It assumed that the link between amino acids is activated by accepting energy from the ATP. The observation stream reasoned that the activation of the amino aggregates could be similar to the bacteria stream , where the pantothenic acid links are activated by the formation of CO N H peptide bonds. The observation stream initiated the construction stream to build the two aggregates that should be linked, i.e., the ATP aggregate and the amino aggregate. The tiles and the local matching rules for both aggregates were already inputs fr om the observation stream and saved in the construction set. However, the construction stream had to define the global matching rules to control the linkage between the two aggregates. By comparing the nature of this discovery with the previous discoverie s, the construction stream assumed that this link would be between the carboxyl group of an amino aggregate (COOH) and the phosphate tile of ATP (P). This link would create a high energy bond (~) between the amino aggregate and the remaining part of the AT P aggregate. The chemical structure of ATP (Ad.P~P~P) has two (~) bonds and the reaction to COOH could have happened on any side of the bond. Thus, the construction stream triggered four thought experiments to construct the different scenarios of the react ions between the amino acid and ATP. The stream first built the amino and ATP aggregates, see Figure 24 and Figure 25 . The construction stream then built a visual representation for all possible scenarios to build the

PAGE 108

! &( ! active amino aggregate. Figure 26 Figure 29 show the four scenarios of the reaction investigated by the construction stream. Figure 24 . General r epresentation of a m ino a ggregates . Figure 25 . The ATP a ggregate . Figure 26 . First scenario CO binds with Ad.P .

PAGE 109

! &) ! Figure 27 . Second scenario CO binds with Ad.P~P . Figure 28 . Thir d scenario CO binds with P~P . Figure 29 . Fourth scenario CO binds with P .

PAGE 110

! &" ! The expression stream then followed the steps of the construction stream in creating the four scenarios. At every step, the expression stream generated an in struction aggregate associated with the observed scenario. To explain the creation of the algorithmic mosaic for the scenario shown in Figure 26 , the expression stream followed and tagged the building of the ATP aggregate and it c reated a shell named "Construct ATP", see Figu re 30 . While monitoring the process of building the amino aggregate, the expression stream tagged the tiles, aggregates and matching rules needed to build the aggregate. Again, the exp ression stream generated a shell associated with this process and named it "Construct amino." The process of building the amino aggregate starts by creating the COOH aggregate and then attaching it to the CH aggregate. All the local complementarity rules t o attach the tiles and aggregates were already saved in the construction set and were the inputs to this thought experime nt. The COOHCH aggregate would be attached to the NH2 aggregate. Finally, the R tile would be added as a symbol denoting different amin o groups. The result of this process is a visua l model of the amino aggregate, seen in Figure 24 , by the construction stream and a "Construct amino" shell by the expression stream, see Figure 3 1 . Figu re 30 . The shell corresponded to constructing the ATP aggregate . Attach (Ad, P, Complementarity rule) Attach (Ad.P, P, Complementarity rule) Construct ATP ! Attach (Ad.P~P, P, Comp lementarity rule)

PAGE 111

! &' ! Figure 3 1 . The shell corresponded to constructing the amino aggregate . The process then continued by following the chemical interaction between the ATP and the amino aggregate. The result was a new shell named "Construct active amino", see Figure 32 . The expression stream then grouped the three shells together and built the algorithmic mosaic corresponding to the process of activating the amino aggregates. The stream repeated these steps to build the three different algorithmic mosaics to explain the constructions of the scenarios in Figure 27 Figure 29 . Figure 32 . The shell generated to construct the Active amino aggregate . In order t o discover the correct transformation rules used to create the activated amino aggregate, the observation stream identified a lab experiment to verify which reaction , of the above four, is correct. The stream found that the energy attached to ATP before the reaction occurs was equivalent to the energy attached to the active amino, thus the stream reasoned, based on chemical principles, that the energy must easily go f rom one side of the formula to the other. This reasoning led the observation stream to choose the method of using activated phosphate atoms, letting them bind to the ATP, isolating the resulting radioactive ATP, engaging it to react with the amino acids, isolating the ATP again after the reaction, and testing its radioactivity to see how many phosphate atoms were lost. The observation stream then triggered the validation Generate (COOH, c omplementarity rule , construction set ) Attach (CH, COOH, Complementarity rule) Construct amino ! Generate (CH, c omplementarity rule , construction set ) Generate ( NH 2 , c omplementarity rule , construction set ) Attac h ( NH 2 , CH COOH , Complementarity rule) Attach ( NH 2 CH COOH , R, Complementarity rule) Detach (Ad.P, P~P) Construct Active amino ! Attach (Ad.P, NH 2 CH COOH , Complementarity rules)

PAGE 112

! &= ! stream to follow the actual experimental results. The validation stream, based on the lab data, confirmed that the process shown in Figure 26 is the correct activation of the amino aggregates. Finally, to discover how active amino acids move in the cell, the observation stream considered that more than one amino ag gregate exists and that all of them must follow the chemical rules of cells in living organisms. Those rules show that, in cells, all reactions are catalyzed by enzymes. Thus, the stream assumed that there are enzymes used during the activation process to activate the amino aggregates. One enzyme would carry both the ATP and a specific amino aggregate, thus the stream reasoned that there must be at least 20 different enzymes to activate the 20 different amino acids. The enzyme serves as the backbone of the mosaic by having special holes, one for amino aggregates and another one for the ATP aggregate. The resulting bigger aggregate is the new active amino aggregate attached to the enzyme, as shown in Figure 33 . The expression stream modified the "Construct Active amino" shell and included the step of creating a general shape of an enzyme with two holes. The expression stream then added the step of attaching the active amino to the new enzyme. The local complementarity rules control th e placement of the ATP and the amino aggregates in their correct place in the enzyme. The final "Construct Active amino" shell is shown in Figure 34 . Figure 33 . Active amino aggregate carried by the enzyme backbone as visualized by the construction stream.

PAGE 113

! &> ! Figure 34 . The final shell to construct the a ctive amino aggregate . 7.3.3 Identifying the Process Rules to Link Amino Acids to tRNA After the discovery of the transformation rules neede d to construct the active amino aggregate, the observation stream initiated a series of thought experiments to determine how the ribosomes correctly sort the amino aggregate based on the genetic code. Notice that at the time of this discovery, the mRNA mol ecule was not yet discovered. Also, the stream initially did not receive any of the outputs from the thought experiment in Section 7.3.1 . The observation stream began to identify the structure of the ribosome based on the curre nt information. It found that ribosomes are constructed out of RNA. The stream studied the nature of the RNA structure in general. Based on the RNA's general structure, the observations stream declared that RNA is constructed out of at least four different nucleotide aggregates. Each one of these nucleotides aggregates is built out of three tiles (a base, a sugar, and a phosphate tile). The structure of the sugar and the phosphate tiles are the same in every aggregate , while the base tile has at least four different structures (A, U, G, and C ) that differentiate one nucleotide aggregate from another. In order to build the preliminary structure of any RNA, the sugar tile of one nucleotide aggregate is linked to the phosphate tile of another nucleotide aggrega te to form a sugar phosphate bond following the global complementarity rules. The stream assumed that some of the RNA in ribosomes must have the same nucleotides sequence as the DNA and , Construct Active amino ! Attach (Ad.P, NH 2 CH COOH R , Complementarit y rules) Create (enzyme, Transformation rules) Attach (Ad.P CO CH NH 2 R , enzyme, Complementarity rules) Call Construct ATP Call Construct amino Detach (Ad.P, P~P)

PAGE 114

! &# ! therefore , acts as a template that carries the genetic code, which is used by the ribosome to assemble the amino aggregates. From previous lab experiments, the observation stream found that during protein synthesis, amino acids could bind to the RNA. These experiments found two different types of RNA that bind to the amino acids. The first type is the ribosomal RNA and the other is a small amount of soluble RNA not belonging to the ribosome. In study ing the role of each of these, the lab experiments were repeated by adding only the RNA ribosome in a protein synthesis extract and the soluble RNA only in another protein synthesis extract. The results found that the amount of soluble RNAs bound to the amino acids is much more than the amount of amino acids bound to the ribosome. In order t o confirm the latest conclusion about th e soluble RNA, the observation stream followed the results of a series of lab experiments that tested the role of soluble RNA. The observation stream noticed that, during the protein synthesis, when adding C 14 ATP, the result was an RNA fraction that is la beled with 14 C after it is donated by the ATP. Again, the stream noticed that while repeating the experiment by adding C 14 leucine instead of C 14 ATP, the resulting RNA fraction was labeled by the amino acid precursor. In both experiments, the observation stream had strong evidence that the amino acid covalently bound to the RNA. In order to confirm that the sources of RNA bound to the amino acids, the stream followed the lab experiment that excluded the ribosomal RNA from the lab extract. This lab experime nt confirmed that the C 14 leucine bound to this low molecular weight RNA (first called sRNA and later tRNA). The observation stream found that, during lab experiments, the fraction of attached amino to mRNA was the same as the fraction of active amino acid s. Therefore, the stream concluded that

PAGE 115

! &$ ! the ATP and amino aggregates first reacted to form the active amino aggregate that later binds to the tRNA after donating it from the enzyme that carried the active amino aggregate, see Figure 33 . Again, the choosing of the correct active amino aggregate to bind with the correct tRNA must be done using special enzymes that are responsible for this reaction. Thus, the stream concluded that there must be 20 different enzymes that are responsibl e for recognizing the specific type of tRNA and the specific active amino acid. This recognition might be achieved via the two holes in the enzyme that only fit one type of molecule (20 different shapes for each of the 20 different tRNAs and the 20 differe nt active amino aggregates). In order to identify the process rules, which are part of the transformation rules, that control the binding of the active amino aggregates to the tRNA, the observation stream started by identifying the tRNA mosaic to be a sequ ence of nucleotide aggregates that are linked to each other to form sugar phosphate bonds. The final structure of the tRNA mosaic was unknown at the time, but it was discovered later. The observation stream triggered the construction stream to build a visu al model of the tRNA. The construction stream generalized the shape of the tRNA ignoring all of its specifics and erasing all the particulars. The result is the general shape shown in Figure 35 .

PAGE 116

! &% ! Figure 35 . The first step of the translation process: constructing the amino tRNA . Next, the observation stream continued to identify the matching rules that control the merging of the active amino aggregate and the tRNA mosaic. The stream reasoned that the transfo rmation rules that activate amino aggregates would be applicable. The ATP aggregate is bound to the amino aggregate from the CO side. This CO side can also be chemically bound to the tRNA after the active amino aggregate releases the Ad.P aggregate. The o bservation stream triggered the construction stream to mimic the process of attaching the active amino to the tRNA to build the amino tRNA mosaic. The construction stream first built the active amino following the process rules given in Section 7.3.2 . Simultaneously, the expression stream retrieved the corresponding shells to activate the amino aggregates, see Figu re 30 , Figure 3 1 , and Figure 34 . The construction stream then built the amino tRNA mosaic by detaching the Ad.P from the active amino aggregate. Next, the active amino, using the energy from the detaching, was linked to the tRNA from the CO side. The resulting mosaic is an amino RNA mosaic wi th an activated amino aggregate attached. The process of building the amino tRNA is shown in Figure 35 . The expression stream also built the shell corresponding to this process, see Figure 36 .

PAGE 117

! && ! Figure 36 . "Construct amino tRNA" s hell . Once the observation stream received the hypothesis of the experiment in Section 7.3.1 , it started to identify the role of the activated tRNA mosaic in the mechanism of pro tein synthesis. The stream realized that the tRNA mosaic binds to the activated amino aggregate to construct the amino tRNA mosaic that travels toward the ribosome to be used in making the protein peptide mosaic. That assumption led to many lab experiments to follow the new amino RNA mosaic in ac tion and all results confirmed this mosaic as the intermediate step between activating the amino aggregate and the protein synthesis , [53] . 7.4.0 Finding the Rules to Attach tRNA and mRNA Duri ng the elongation stage of translation, the codon (code) aggregate in the mRNA attaches to the anticodon in the tRNA mosaic forming the codon anticodon aggregate. The anticodon is a special code aggregate in the tRNA that consists of three tiles complement ary to the tiles in the mRNA codon aggregate. The codon and anticodon should be linked, during translation, to form the mRNA tRNA mosaic, [6] , and [50] . In this section, the thought experiments that found the correct matching rules to construct the codon anticodon aggregate were identified by applying the AD on the discovery of the wobble theory in [38] . These rules are key to calling the correct amino aggregate based on the reading of the genetic code. Also, the communication thought experiments prepared by the expression stream to generate the algorithmic mosaic describing the process of attaching the mRNA mosaic to the tRNA mosaic are discussed , [6] . Call Construct Active amino Construct amino tRNA ! Detach (Ad.P, Active amino ) Create (tRNA, Transformation rules) Attach (Active amino, tRNA, Complementarity rules)

PAGE 118

! )(( ! 7.4.1 Defining Matching rules to link mRNA and tRNA Mosaics The AD launched the observation stream to study an important step of the algorithm of translation. To find the correct structure of this algorithm, the AD focused the visual streams to inve stigate the process of linking the mRNA mosaic to the tRNA mosaic , [38] . In the preliminary stages of the observation stream, the initial inputs corresponding to the problem in hand were collected. The observation stream retri eved the inputs about the mRNA mosaic declared by previous visual streams, Section 6.6.1 . The mRNA mosaic is constructed out of a linear structure of codon aggregates. The complementarity rules, for linking the tiles of the mR NA mosaic together, simply line up the letter tiles next to each other to form a long sequence of codon aggregates. This sequence has two ends, the 3' end and the 5' end, [6] , [50] , and [38] . The observation stream saved the inputs about the tRNA mosaic. There exist at least 20 different mosaics of the tRNA. Each tRNA mosaic carries a specific amino aggregate. The 3D structure of the tRNA mosaic was not fully defined. The preliminary inputs showed that the tRNA mosaic is built out of a sequence of base tiles (A, U, G, C). Another lab input declared a new letter tile , I used to construct the tRNA anticodon. Moreover, lab inputs reported that the amino aggregate is always li nked to the 3' end of the tRNA mosaic. The tRNA anticodon aggregate should be linked, during translation, to the corresponding codon in the mRNA . The observation stream saved the incoming inputs from the thought experiments that deciphered mos t of the gene tic code, see Chapter VI . The stream also received the different sets that were identified in these thought experiments, i.e. , the initiation, termination, amino, and code sets. It also collected the reading rules needed to correctly read the genetic code in the mRNA mosaic.

PAGE 119

! )() ! The AD examined the matching rules for attaching the codon to the anticodon in the mRNA and tRNA mosaics , respectively. This attachment would construct a "temporary mosaic" that includes the mRNA and the tRNA mosaics as well as the ami no aggregates. Historically, the thought experiments , presented in Section 7.4.0 , were performed by Crick, [6] , and [38] . 7.4.2 Discovering Matching Rules At the time of this thought experiment, the observation stream's goal was to find the correct complementarity matching rules between the mRNA and the tRNA. Note that the expression stream monitored and tagged the actions of every visual stream of this discovery in parallel. The output of the expression stream was a symbolic mosaic that includes all of the possible pairings between the mRNA and the tRNA. The observation stream followed the assumption, collected from other thought experiments, that different types of tRNA mosa ics have common features. The stream received new inputs about a mosaic called the ribosome. This mosaic is the machine that rea ds the genetic code in the mRNA, calls the corresponding tRNA , and constructs the peptide mosaic . Further, previous lab experime nts observed the tight binding between the first two tiles of the mRNA codon and their corresponding tiles, which comprise the tRNA anticodon, shown in Figure 37 . These experiments also verified that the codon's first two tiles we re clearly distinguished. Based on that, the observation stream concluded that the first two tiles in the codon and anticodon aggregates would be attached together following the same RNA base pairs complementarity rules, i.e., G C and A U. The stream also received additional inputs observing the A T and I C pairing. The observation stream saved these pairings to the construction set. It only excluded the A T pairing since T is not part of the base tiles of either the tRNA nor mRNA mosaics, see Figure 37 .

PAGE 120

! )(" ! Figure 37 . The third tile's matching rule is unknown , [6] . With respect to the pairing between the first two tiles in the mRNA and tRNA, the expression stream added the observati on stream results into a symbolic mosaic. This mosaic is a table with four rows and five columns. The rows represent the four possible tiles of the mRNA, (A, U, C, and G). The columns represent all the possible tiles of the tRNA, (A, U, C, G, and I). The intersection between the rows and the columns will be either zero or one. When there is no existing complementarity matching between the mRNA and tRNA tile, the intersection of the corresponding row and column will be a zero. The value of one will indicate the existence of a complementarity matching between the tRNA and the mRNA tiles. The expression stream added all of the known RNA complementarity matching pairs to the table. Table 7 shows the symbolic mosaic created by the expr ession stream, named the RNA complementarity matching mosaic, for the complementarity matching between the third two tiles in the mRNA and the tRNA. !

PAGE 121

! )(' ! Table 7 . The RNA c omplementarity m atching m osaic b etween the f irst t wo t iles in the mRNA and tRNA . tRNA ! U A G C I mRNA " A 1 0 0 0 0 U 0 1 0 0 0 C 0 0 1 0 0 G 0 0 0 1 0 The expression stream also initiated a similar symbolic mosaic to save the complementarity matching for the last pairs in the mRNA and the tRNA and named it the wobble complementarity matching mosaic. Initially, it included the existence of a complementarity matching between the C and I tiles, as defined by the observation stream, Table 8 . Table 8 . The w obble c omplementarity m atching m osaic b etween the t hird t ile in the mRNA and tRNA . tRNA ! U A G C I mRNA " A 0 0 0 0 0 U 0 0 0 0 0 C 0 0 0 0 1 G 0 0 0 0 0 The observation stream focused itself on studying the matching rules that direct the attachments between the last, the third, tiles of the codon and anticodon aggregates ( Figure 37 ). The stream considered the degeneracy of the genetic code, i.e., the fact that several codon aggregates can call the same amino aggregate. From the lab results, the observation stream, in its analysis stage, reasoned several d ifferent facts about the third base tile of the codon. The streams found dozens of inputs showing that the tile C is equivalent to tile U and the stream did no t find any data that contradicted with these findings. The stream also observed some cases where tile I is equivalent to tile A and other cases where tile I is not equivalent to tile A.

PAGE 122

! )(= ! With the above inputs, the observation stream reasoned that the matching between the two aggregates must explain the fact that one tRNA mosaic could be called by more than one codon sequence. Many lab inputs explained that the difference in the codon sequence that recognizes the same tRNA is always in the third tile. For example, phenylalanine tRNA could be called by the codon sequence UUU or UUC. The observation stream reasoned that since two different codons could recognize the same tRNA, then the pairing between the codon and anticodon in the third tile must be non standardized, i.e., does not follow the DNA base pair matching rule as the first and second tiles of the codon and anticodon aggregates. The stream concluded that this non standardized pairing means that the pairing could wobble; the wobble pairing could support more options when compared to the standard pairing , but still close to it. In order to investig ate the wobble pairing, the observation stream triggered a thought experiment to study all of the pairing options for all possible combinations of base tiles according to the tiles' chemical structure by applying the non standardized position of the glycos idic (sugar) bonds. The stream must investigate the pairing options for A, U, G, C, and I tiles as the third tile of the anticodon aggregate. To accomplish this, the observation stream initiated the construction stream to construct the pairing between the two base tiles, A and G. However, the construction stream could not link one hydrogen bond of the G's NH group, not even via water. The observation stream reasoned that this failure was because the physical distance measurements, i.e. , the physical rules, between the G and A, if linked together, would be violated. The physical rules stated the possible physical measurements of each pair known from x ray crystallography structures. The physical rules are part of the environmental rules, Section 2.5.0 , which represent the physical constraints that define the distance that must be obeyed when linking two pairs together. A v isual representation of the construction stream's attempts to build

PAGE 123

! )(> ! the G A aggregate can be seen in Figure 38 . This failure led the observation stream to exclude the G A pairing possibility. Figure 38 . The c onstruction stream failed to build the G A aggregate , [6] . Then, the o bservation stream scheduled a thought experiment for the construction stream to establish the linkage between the U and C tiles. The observation stream sent, along with this thought experiment, all known information about structure of base tiles, the prefe rred bonds, and the structure of canonical pairing. In addition, the inputs showed t hat the first two positions contain the sugar phosphate of the RNA backbone were also provided to the construction stream. The construction stream started to construct the linkage, i.e. , find the correct local matching rule, between the U and C tiles, having full understanding of the acc urate geometries and lengths of the bonds. In order to build the U C aggregate, the construction stream brought the keto groups and the g lycosidic bonds, in both tiles, close together. In the U tile, the carbon and oxygen keto group form a double bond, i.e., C=O. The C=O group was linked through a hydrogen bond with the C tile. Similarly, the keto group of the C tile formed a hydrogen bond with the U tile, see Figure 39 , [44] .

PAGE 124

! )(# ! Figure 39 . The c onstruction stream succeeded in building the U C aggregate , [6] . Constructing the U C aggrega te led the observation stream to add the U C pairing to the local complementarity rules. Immediately, the expression stream tagged the new finding of a U C complementarity. The construction stream also built the U U aggregate by bringing the two keto and t he two glycosidic groups closer together. The observation stream saved the U U pairing to the local complementarity rules. Simultaneously , the expression stream also tagged the U U complementarity's existence. Once the observation stream finished its inte rnal thought experiment, the expression stream updated the wobble symbolic mosaic to include the new findings, as shown in Table 9 . Table 9 . The u pdated w obble c omplementarity m atching m osaic . tRNA ! U A G C I mRNA " A 0 0 0 0 0 U 1 0 0 1 0 C 1 0 0 0 1 G 0 0 0 0 0 Next, the observation stream triggered the construction stream to build the G U aggregate. The construction stream built the G U aggregate by moving the bonds approximately 2.5A¡ from their normal position, Figure 40 . The construction stream also observed the similarities of the structure of G and I, thus the stream built the aggregate I U following the same process. Following these two constructions, the observation stream included the G U and I U pairing in

PAGE 125

! )($ ! the local complementarity rules. Again, the expression stream tagged all the new findings and saved them in the wobble complementarity matching mosaic. Figure 40 . The G U a ggre gate , [6] . Lastly, the observation stream initiated the construction stream to link between the I and A tiles. The construction stream created the I A aggregate by bonding the A and I tiles together applying the same the DNA an gles and positions of the respective glycosidic , see Figure 41 . Table 10 shows the updated wobble complementarity matching rules built by the expression stream after monitoring the internal thought experi ments created by the observation and construction streams. Figure 41 . The I A a ggregate , [6] . Table 10 . The u pdated w obble c omplementarity m atching m osaic . tRNA ! U A G C I mRNA " A 0 0 0 0 0 U 1 0 1 1 1 C 1 0 0 0 1 G 0 0 0 0 0

PAGE 126

! )(% ! After finding all correct pairings between different combinations of the base tiles (A, U, G, C and I), the observation stream declared a final set of local complementarity rules. This set is a sp ecial type of scope set that is used by the streams to test the correctness of the pairings within this set. The scope sets are defined by the observation stream to discard incorrect options , while looking for the solution of a problem. After saving the lo cal complementarity rules set, the observation stream passed it to the validation stream to investigate the validity of the physical rules for every pairing within this set. The validation stream measured the positions of bonds for each pairing and compare d them with the allowed physical measurements known from the x rays. For example, looking to the pair U C , where the U is the anticodon tile and C is the codon tile, the stream visualized the actual position of the U C pair by applying the physical rules t hat control the position of those two tiles, as shown in Figure 42 . Figure 42 . The p hysical p osition of the b onds for the p air U C . The validation stream found that the distance between the U and C tiles in the U C aggregate is smaller than the distance shown in the x rays. The validation stream concluded that this aggregate violated the physical rules and excluded the U C pairing from the set of the local matching rules. The validation stream continued co mparing the distances of every pairing with the actual distance known from the x rays, including the normal base pair matching rules. The

PAGE 127

! )(& ! validation stream excluded the U U and C U pairing from the local complementarity rules set , because they also violate d the physical rules, see Figure 43 . On the other hand, the validation stream verified seven di fferent positions for the tiles pairing s . Figure 43 . The physical position of the bonds violates the physical rule: too close bonds . Consequently, the expression stream eliminated the pairing between the identified tiles from the wobble complementarity matching mosaic. The final set of local complementarity rules included all of the pairs that did not violate th e physical rules, see Figure 44 . Historically, they were called the wobble rules. The observation stream reasoned that these rules would be applied to build a new aggregate by linking the third codon's tile to its corresponding an ticodon's tile. A n a pplication of the wobble complementarity rules would allow the full attachment of the mRNA and tRNA mosaics. Figure 45 shows the complementarity matching rules between the codon and anticodon as constructed by the construction stream , while Table 11 shows the final wobble complementarity matching symbolic mosaic built by the expression stream. The RNA complementarity matching and the wobble complementarity matching mosaics are essentia l during translation to link the anticodon to its corresponding codon. During translation,

PAGE 128

! ))( ! visual streams must check the wobble complementarity matching mosaic before allowing any linkage between the tRNA and the mRNA mosaics. Figure 44 . The physical position of the bonds obeys the physical rule . Figure 45 . The c omplementarity r ules that c ontrol the b inding of mRNA and tRNA , [6] . Table 11 . The f inal w obble c omplementarity m atching m osaic . tRNA ! U A G C I mRNA " A 0 0 0 0 1 U 0 0 1 0 1 C 0 0 0 0 1 G 1 0 0 0 0

PAGE 129

! ))) ! 7.5.0 The Watson Model After finding the process rules to build the amino tRNA and the wobble complementarity rules to attach the anticodon to the codon aggregates, the next step was t o find the rules that control the translation process. In this section, the Watson thought experiments [46] were rediscovered . These are the thought experiments that generated the first theoretical framework to explain the elo ngation cycle, i.e., the second stage of protein synthesis after the initiation stage and before the termination stage. The AD began by triggering the observation stream to prepare the first thought experiment with the goal of revealing the Watson model. The observation stream started its preliminary stage to collect the inputs related to the problem coming from the impression stream. The first input was from the experiments confirming the existence of the mRNA, which carries the genetic message. As expla ined in Section 6.6.1 , mRNA is a linear information mosaic that is built out of four different tiles (A, U, G, and C). The observation stream retrieved the definition of the tiles, aggregates, and matching rules needed to build the mRNA mosaic and saved them into the construction set that would be used when constructing a visual model of this mosaic , [46] . The observation stream next accepted the results from the thought experiments that identified the process rules needed to create the amino tRNA mosaic, see Section 7.3.0 , and added these process rules to the construction set. The observation stream triggered a thought experiment to construct the amino tRNA mosaic. The stream sent the thought experiment along with the construction set needed to build the amino tRNA to the construction stream to start building the amino tRNA mosaic.

PAGE 130

! ))" ! In order t o build the amino tRNA, the construction stream retrieved the information that t he amino aggregate must be attached to the tRNA at a specific region, the 3` end. The construction stream also had inputs that declared an anticodon area in the middle of the tRNA. The construction stream first found that for each amino aggregate there is a specific type of tRNA. However, the stream reasoned that these specifics are not necessary for this discovery , as the only important part of the tRNA is the anticodon area. Therefore, the construction stream erased all the particulars and constructed th e tRNA as a strand with three base tiles (either A, U, G, I, or C) in the middle of the strand representing the anticodon. The stream visualized the strand as taking the shape of a rectangle, see Figure 46 . Figure 46 . The tRNA m osaic b efore f olding as r epresented by the c onstruction s tream . The construction stream then folded the tRNA to mimic the x ray pictures showing the folding of the tRNA. With no clear inputs on how tRNA folds, the construction strea m folded the tRNA in a manner similar to DNA folding, as in Figure 47 . The construction stream saved the actions it applied to create the body of the tRNA in the construction set as part of the process rules needed to build any tR NA mosaic.

PAGE 131

! ))' ! Figure 47 . The tRNA m osaic a fter f olding as r epresented by the c onstruction s tream . At the same time, the expression stream created a corresponding shell to save the steps followed by the construction stream in building the body of the tRNA. The first actions tagged by the expression stream were the creation of the anticodon aggregate from the five different tile possibilities (A, U, G, I, or C). The next step was generating the body mosaic as a long rectangle. Then, the expression stream saved the actions of adding the anticodon to the body mosaic. As a result, the expression stream generated a new shell to save the operations followed by the construction stream to build the tRNA. The initial shell of "Construct tRNA body " is shown in Figure 48 . Figure 48 . "Construct tRNA body" s hell . After creating the initial body of the tRNA, the construction stream reasoned that the anticodon aggregate in tRNA mosaic would be attached to the codon aggregate in the mRNA mosaic following the wobble rules. Thus, the anticodon area must be constructed at the bottom of the tRNA to make the linkage between the two mosaics possible, as shown in Figure 47 . The Generate (anti codon, Complementarity rules, tRNA tiles set ) Construct tRNA body ! Attach (body, anti codon, Complementarity rules) Generate (body , Construct ion rules)

PAGE 132

! ))= ! expressi on stream simultaneously tagged the process of folding the tRNA and saved it in the "Construct tRNA body" shell, see Figure 49 . Figure 49 . The m odified "Construct tRNA body" s hell . In order to build the amino tRNA, the construction stream added the amino aggregate at the 3` of the tRNA mosaic, see Figure 50 (a). Attaching the amino aggregate to the tRNA must follow the process rules explained in Section 7.3.3 . The expression stream simultaneously retrieved the associated shell and modified it to include the new findings by the construction stream, as shown in Figure 51 . Figure 50 . Three types of tRNA mosaics built by the construction stream: (a) amino tRNA, (b) peptide tRNA, and (c) deacylated tRNA. Generate (anti codon, Complementarity rules, tRNA tiles set ) Construct tRNA body ! Attach (body, anti codon, Complementarity rules) Generate (body , Construction rules) Fold (DNA transfor mation rules) I2J ! IKJ ! I3J !

PAGE 133

! ))> ! Figure 51 . "Construct amino tRNA" s hell . Next, the observation stream declared another type of tRNA that carries the peptide mosa ic during protein synthesis and named it peptide tRNA. The observation stream triggered the construction stream to build the structure of the peptide tRNA. The construction stream built the peptide tRNA mosaic using the same reasoning in building the amin o tRNA mosaic. The only difference is the attachment of a peptide mosaic at the 3` of the tRNA instead of the amino aggregate. Since the peptide mosaic is a chain of amino aggregates and the matching rules of attaching the amino aggregates were already kn own, the construction stream generated a peptide mosaic and then attached it to the amino aggregate to build the peptide tRNA, see Figure 50 (b). Simultaneously, the expression stream called the algorithmic mosaic to construct th e peptide mosaic, which was defined while building the peptide mosaic. Then, the expression stream generated a "Construct peptide tRNA" shell following the construction stream actions, as in Figure 52 . Figure 52 . "Construct peptide tRNA" s hell . The construction stream found that during the process of elongation, tRNA with no amino or peptide attached to it must appear. Thus, the stream declared another type of tRNA with nothing Call Construct Active a mino Construct amino tRNA ! Detach (Ad.P, Active amino) Call Construct tRNA body Attach (Active amino, tRNA, Complementarity rules) Call Construct peptide mosaic Construct peptide tRNA ! Call Construct tRNA body Attach (peptide , tRNA, Complemen tarity rules)

PAGE 134

! ))# ! attached to it and named it d eacylated tRNA, see Figure 50 (c). The expression stream also reasoned that deacylated tRNA can be reconstructed following the "Construct tRNA body" shell in Figure 48 . The construction stream then built the mRNA mosaic as identified in the construction set. In order to erase the particulars, the construction stream did not include any of the specifics of the base tiles and visualized them as simple shapes with four different colors, each color representin g one of the actual base tiles in the tRNA, see Figure 53 . The expression stream retrieved the "Construct information mRNA" shell, from Section 6.6.2 , and saved it as part of the elongation process. Figure 53 . Constructing the mRNA m osaic . Some inputs showed that the growing peptide mosaic is attached to the amino aggregate at its R amino terminal, i.e., the amino aggregate COOH group of the last amino aggregate in the peptide mosaic would link to the complementary R amino group in the amino aggregate. The observation stream concluded that, after the attaching the new amino aggregate to the growing peptide mosaic, the peptide mosaic would still be attached to its peptide tRNA. The observation stream saved this new finding as part of the process rules for elongation. The observation stream also received snapshots from previous thought experiments that represented the information, known at that time, of the elongation process, dep icted in Figure 54 .

PAGE 135

! ))$ ! Figure 54 . The translation process as defined by previous thought experiments , [46] . The observation stream realized that besides the mRNA, amino tRNA, pept ide tRNA, and deacylated tRNA, the studies showed another mosaic, the ribosome that could be used as a site of protein synthesis. The observation stream discovered which the genetic code that guides the construction of any protein does not exist in the rib osome , but is carried by the mRNA that would be attached later to the ribosome. Next, the observation stream initiated a series of thought experiments to identify the construction set that would be used to build the ribosome mosaic and attach mRNA and tRNA to it. 7.5.1 Defining the Structure of the Ribosome In order to define the structure of the ribosome, the observation stream began to analyze its chemical structure. Inputs from several lab experiments showed that ribosomes are constructed out of two smaller mo saics. The stream named these mosaics the 50 s and the 30 s mosaics. Each of these mosaics is built out of smaller protein and RNA mosaics. The RNA mosaics that build the ribosome are special single strand RNAs called ribosomal RNA (rRNA). The observation stream reasoned that the specifics of the 30 s and 50 s mosaics are not required for this discovery. Only a general shape of the two ribosomal mosaics is needed. Thus, the

PAGE 136

! ))% ! observation stream erased the particulars and concluded that the ribosome consists of two smaller mosaics named the 50 s and the 30 s mosaics. The 30 s is smaller than the 50 s mosaic. Next , the observation stream saved the definition of the two ribosomal mosaics into the construction set to be applied , while building the ribosome mosaic , [46] . 7.5.2 Identify the Matching Rules between the mRNA and the Ribosome After defining the general structure of the ribosome mosaic, the observation stream reasoned that to read the genetic message in the mRNA, the ribosome mosa ic must be first linked to the mRNA mosaic. The stream began to investigate the rules that control this linking. It found that the mRNA should be linked to either the 50 mosaics or to the 30 mosaic of the ribosome. In order to identify the correct location where the mRNA is linked to the ribosome, the observation stream received the results of lab experiments that monitored two cell free systems. The first system included poly U mRNA and 30 s mosaics and the other one included poly U mRNA and the 50 s mosai cs. The results of the first experiment found the poly U was attached to the 30 s mosaic , while the other experiment found no connection between the poly U and the 50 s mosaics. Based on these lab results, the observation stream concluded that during elong ation, mRNA mosaics attached to the 30 s mosaic of the ribosome. The stream saved this conclusion into the set of process rules of elongation to be used later by the construction stream. In order to find the complementarity matching rules between the 30 s mosaic and the mRNA, the observation stream analyzed the results from the Takanami and Zubay lab experiment, [57] . In the Takanami and Zubay experiment, a mixture of 14 C labeled poly U and ribosomes were treated with a small am ount of ribonucleases , which are used to detach base tiles of the mRNA, i.e., the mRNA nucleotides are disassembled after treating the mRNA with ribonucleases. The lab results found that not all the base tiles of the mRNA were freed, 25 30 base tiles were still

PAGE 137

! ))& ! attached to each other. Receiving these results, the observation stream concluded that the unaffected base tiles must be protected by the ribosome. Thus, the stream declared that, while the mRNA attaches to the ribosome mosaic, the mRNA would link t o the 30 s mosaic. However, not all of the base tiles of the mRNA mosaic would be inside the 30 s mosaic at once, only 25 consecutive mRNA base tiles would be inside the 30 s mosaic, while the rest of mRNA mosaic would be outside the ribosome. The observat ion stream declared new process rules stating that the mRNA mosaic enters the ribosome from the 30 s mosaic. Only 25 base tiles stay inside the 30 s ribosomal mosaic during the process of elongation. This rule was saved into the set of process rules to be used by the construction stream when constructing the elongation process , [46] . 7.5.3 Finding the Plugging Sites for tRNA into the Ribosome Mosaic After attaching the mRNA mosaic to the ribosome, the next step of the elongation proce ss is to read the genetic code and call for the correct tRNA mosaic. The observation stream had no information about the way the tRNA is attached to the ribosome. The stream began the preliminary stage by gathering the results from lab experiments that tes ted the rules of binding tRNA with ribosomes. Results from lab experiments found that amino tRNA binds to the ribosome by hydrogen bonds, and possibly Mg ++ bridges, to a special cavity in the ribosome. The observation stream added a new construction rule, to build the ribosome, stating that inside the ribosome there must be a special location for the amino tRNA binding. The observation stream named it the A site. On the other hand, the observation stream did not find any precise inputs on the exact numbe r of amino tRNA that bind to each ribosome. However, some initial inputs indicated that only one amino tRNA molecule is attached firmly to the A site. Additionally, the inputs showed

PAGE 138

! )"( ! the absence of any covalent bond in the linkage of the growing peptide mo saic to its site in the ribosome. Instead, the peptide mosaic is bound to the terminal tRNA mosaic that fits in a cavity with in the 50 s mosaic. Moreover, many studies found that the peptide mosaic always attaches to the 50 s mosaic of the ribosome into a site different from the amino tRNA binding site , [46] . Thus, the observation stream declared at least two plug in sites for the tRNA to bind to the ribosome. The first is the A site , which holds the incoming amino tRNA before the formation of the polypeptide chain. The second is the binding site (P site), which is the place of the polypeptide chain formation. The observation stream updated the construction rules to build the ribosome. The construction stream would apply these r ules when building the ribosome mosaic. 7.5.4 Constructing the Mechanism of Elongation The observation stream finished the identification of the construction set (to include the 30 s mosaic, 50 s mosaic, mRNA, amino tRNA, peptide tRNA) as well as the matching ru les that include the relationship between the different mosaics according to all of the above inputs and conclusions. The observation stream triggered the construction stream to start finding out how all the mosaics interact to build the polypeptide mosaic . Once the construction stream received the construction set and the matching rules from the observation stream, it started its task to run the process of elongation. Along with the discovery streams, the expression stream would build an algorithmic mosai c to represent the elongation process. The construction stream realized that the ribosome mosaic is the machine responsible for processing the whole mechanism. It is the mosaic that is responsible for reading the mRNA mosaic and constructing the peptide mo saic. In order to build the ribosome mosaic, multiple separated mosaics must be joined in a specific order, i.e., the process rules of elongation.

PAGE 139

! )") ! However, the construction stream needed to apply and define the transformation rules for building the ribosom e, mRNA, and tRNA mosaics. The construction stream first generated a general shape of the two ribosomal mosaics, the 30 s and 50 s, see Figure 55 . The stream then added two rectangular holes to the 50 s mosaic to represent the A s ite and the P site. Following that, the construction stream moved the two ribosomal mosaics towards each other and linked them to form the ribosome mosaic, as shown in Figure 56 . At the same time, the expression stream started bui lding the corresponding shell by tagging the actions of the construction stream while building the ribosome, see Figure 57 . Figure 55 . A general s hape of the 50 s and 30 s m osaics . Figure 56 . Constructing the r ibosome m osaic .

PAGE 140

! )"" ! Figure 57 . The shell corresponded to the construction of the ribosome mosaic. Next , the construction stream attached the mRNA mosaic to the 30 s mosaic. As defined by the observation s team, the construction stream visualized 25 30 base tiles of the mRNA inside the 30 s mosaic. It also allowed the mRNA to enter to the 30 s mosaic from a general position since the rules that control this entrance are not yet known, see Figure 58 . Figure 58 . The mRNA enters the ribosome mosaic. In parallel, the expression stream marked the mRNA attachment to the ribosome as the first in the elongation shell. The expression stream create d the instruction ag gregate "Attach (mRNA, ribosome, Matching rules)." The Attach instruction aggregate takes the two operands (mosaics) and links them together following the matching rules declared earlier by the observation stream. Generate (50 s, construction set) Construct the ribosome Add (50 s, A site, construction set) Add (50 s, P site, construction set) Generate (30 s, construction set) Attach (30 s, 50 s, construction set)

PAGE 141

! )"' ! The expression stream saved these rules in to a shell that includes "written description" , while monitoring the rules declaration by the observation stream. The construction stream found that the next logical step in elongation must be a tRNA in the P site. Since the process rules declared that th e P site is the place of the polypeptide chain formation, the construction stream allowed the peptide tRNA to occupy the P site. The expression stream marked the entrance of the peptide tRNA to the P site as the next instruction aggregate in the elongation shell. A snapshot of the visual stream of the starting state of elongation is shown in Figure 59 . Figure 59 . The starting state in the mechanism of elongation identified by the construction stream . Next, the construction stream identified the first step of the elongation cycle. This is where the ribosome read the codon aggregate in front of the A site following the reading rules. Then, the construction stream called the wobble complementarity matching mos aic, see Table 11 , to find the corresponding anticodon. Based on this, the amino tRNA with the correct anticodon entered the ribosome. The amino tRNA moved to the ribosome following the tRNA movement rules. It entered the A site a nd bound to the mRNA following the wobble complementarity

PAGE 142

! )"= ! matching rules, see Figure 60 . All of these actions were also tagged by the expression stream and saved into the elongation shell as instruction aggregates. Figure 60 . The amino tRNA enters into the ribosomal A site . Once the amino tRNA is in the A site, see Figure 61 , the peptide formation would start. Peptide formation is the building up of the peptide mosaic by adding th e incoming amino aggregates to the growing peptide mosaic following the transformation rules identified by the observation stream. The matching rules for linking the amino aggregates to the peptide mosaic were already coming to the construction stream via the impression stream from the output of the AD , [10] . The construction stream also received the process of building the peptide mosaic as inputs and linked the amino aggregate and the peptide mosaic flowing that process. The r esult after peptide formation is a peptide tRNA in the A site is shown in Figure 62 . At the same time, the expression stream tagged the construction stream's actions and saved them in the elongation shell.

PAGE 143

! )"> ! Figure 61 . The amino tRNA occupies the A site . Figure 62 . The p eptide f ormation . The construction stream then reasoned that the tRNA in the P site must be now a deacylated tRNA, i.e., had nothing attached to it. The construction s tream moved deacylated tRNA out of the ribosome to keep the P site free for the next peptide tRNA, see Figure 63 . The construction stream observed that the polypeptide mosaic is always seen in the P site. Therefore, the stream con cluded that the binding of the protein and the ejection of the free amino tRNA must be done simultaneously. Once the deacylated tRNA started to move outside the ribosome, the peptide tRNA in the A site would move to the P site, as seen in Figure 64 . The movement of the peptide tRNA from the A site to the P site triggered a corresponding movement of the mRNA. This movement is governed by the reading rules that were discovered in the previous experiments, see Section 6.5.0 , Figure 65 . The construction stream ended one cycle of elongation and found that

PAGE 144

! )"# ! the next step is to repeat the cycle by reading the codon aggregate in front of the A site and calling for the correct amino aggre gate. In order to end the elongation cycle, the stream reasoned that the ribosome must read a terminal codon. Having the current codon as a terminal codon would end the elongation step. If the current codon was not a terminal codon, then the process would continue with the next elongation cycle. Figure 63 . The tRNA leaves the ribosome . Figure 64 . The peptide tRNA translocate s into the P site .

PAGE 145

! )"$ ! Figure 65 . The next amino tRNA enters the ribosome . Since the construction stream constructed the elongation as a continuing process, the expression stream kept tagging all of the actions and immediately adding them into the elongation shell. Figure 66 shows the final sh ell describing the Watson model for elongation built by the construction stream. Figure 66 . The "Elongation" shell for the Watson model . Attach (mRNA, ribosome, Matchi ng rules) Read (codon, Reading rules) Call (amino tRNA, anti codon) Attach (codon, anti codon, wobble matching rules) Attach (peptide, amino, matching rules) Move (amino tRNA, current location, A site, tRNA movement rules) Move (mRNA, current locatio n, one codon forward, mRNA movement rules) Move (peptide tRNA, A site, P site, tRNA movement rules) Move (deacylated tRNA, P site, outside, tRNA movement rules) Call Elongation anticodon : = Find (codon, wobble complementarity matching rules ) Move (a mino tRNA, current location, P site, tRNA movement rules) Flag : = Compare (current codon, termination set) If Flag then Call terminate Elongation

PAGE 146

! )"% ! 7.6.0 The Watson extension Model In this section, the AD was applied to rediscover the Watson extens ion model , [63] . At the beginning of the Watson extension thought experiments, the AD launched the observation stream to retrieve the outputs from the Watson thought experiments, see Section 7.5.0 . The observation stream, in the preliminary stage, received two important visual inputs via the impression stream from the Watson thought experiments. The first visual input was the introduction of at least two binding sites (A site and P s ite), which serve as a physical location for the tRNA to plug into the ribosome during elongation. The second visual input was the identification of the translocation rules, part of the movement rules, that control the movements of a newly synthesized pept ide tRNA from the A to P site and the movements of mRNA by one codon aggregate at a time. The observation stream also retrieved the algorithmic mosaic that describes how to build different types of tRNA mosaic s based on the mosaic or aggregate that they a re plugged into, as shown Figure 48 , Figure 51 , and Figure 52 . It also retrieved the transformation rules to build the mRNA mosaic and the ribosom e. After collecting these initial inputs about the process of elongation, the observation stream triggered the construction stream to build a visual model of the different types of the tRNA mosaics , as well as the mRNA mosaic , [63] . The construction stream generated "general shape" tiles from the base tiles set and aligned them to build a long string of tiles. This alignment was controlled by the alignments rules of the mRNA tiles. The stream marked every three t iles as one codon aggregate, see Figure 67 . Simultaneously, the expression stream saved the "Construct information mRNA" shell, Section 6.6.2 , Figure 18 , to be the used when r ebuilding the mRNA mosaic.

PAGE 147

! )"& ! Figure 67 . The mRNA m osaic . Next, the construction stream built three types of tRNA mosaics. According to the construction set, the tRNA is constructed out of five tiles (A, U, G, C, and I). The construc tion stream reasoned that the specifics of tRNA mosaics are not needed. The only aggregate that needed to be specified is the anticodon. The construction stream erased the particulars and generated a "general" shape of the tRNA body representing the deacyl ated tRNA. In order to build the amino tRNA, the construction stream attached an amino aggregate to the 3` end of the tRNA following the process rules in Section 7.3.3 . Similarly, the stream built the peptide tRNA by attaching the peptide mosaic to the 3` end of the deacylated tRNA. The different types of the tRNA mosaic as built by the construction stream can be seen in Figure 68 . Figure 68 . Three D ifferent T ypes of tRNA M osa ics While the construction stream was building the three different types of tRNA mosaics, the expression stream was observing and tagging the actions of the construction stream. The expression stream reasoned that the only difference in building the tRNA , when compared to the Watson model , was to not fold the tRNA mosaic. Thus, the stream modified the "Construct tRNA body" shell based on the new findings, see Figure 69 . The shells for attaching the peptide

PAGE 148

! )'( ! and the amino mosaic are the same as the ones generated by the Watson thought experiments, shown in Figure 49 , Figure 51 , and Figure 52 . Figure 69 . The m odified "Construct tRNA body" s hell . Next, in the analysis stage, the observation stream initiated a thought experiment to validate the correctness of the Watson model based on any new inputs. The observation stream also triggered the validation stream to find any contradictions to the rules identified in the Watson model. The validation stream also received additional information: lab reports of observations of more than two tRNA bindings to the ribosome of a rat liver at the same time. This led the stream to question the number of sites for binding tRNA mosaics. For further investigation, the validation stream studied the Woese [31] and Pestka [62] two site models. These two models were identified to have two physically d istinct sites located in the 50 s mosaic. Both of th e se sites have the same function; they can hold a tRNA molecule. Contrarily, the 30 s mosaic has only one binding site that aligns precisely with one site of the 50 s mosaic , creating a ribosome with two sites, the A and P site. The movement rules of these models were declared to allow the tRNA to enter one site and occupy that site leaving the ribosome. In other words, the tRNA does not fo llow a sequential movement when it enters the A site, translocates to the P site, and then exits the ribosome. These two models were invalidated by the validation stream , after it received confirmed inputs that identified two physically distinct sites in the 30 s mosaic that were not a result of the attachment between 30 s and the mRNA mosaics. The binding sites of the 30 s Generate (anti codon, Complementarity rules, tRNA tiles set ) Construct tRNA body ! Attach (body, anti codo n, Complementarity rules) Generate (body , Construction rules)

PAGE 149

! )') ! mosaic, after the plugging between the 30 s and 50 s mosaics, have the same functional properties as the A and P site of the ribosome mosaic. Thus, the validation stream declined the rule of having o nly two binding sites in the ribosome and concluded that the ribosome must have more than two binding sites. Then, the validation stream called the observation stream to investigate the correct number of tRNA binding sites , [63] . The observation stream accepted more inputs from the Wettstein and Noll lab experiment, [40] , showing that both the amino tRNA and the peptide tRNA were tightly bound to the ribosomes, [43] . They stay attached to the ribosome after washing the ribosome even at low concentrations of Mg ++ . In contrast, deacylated tRNA could be easily washed from the ribosome or exchanged with an uncharged tRNA. Uncharged tRNA is a tRNA before merging with the amino aggregate. More evidence confirmed the existence of an exit site, or E site, in the bacterial ribosomes. Having the discovery of the E site, the observation stream received the results of the Nierhaus thought experiments , which declared a t hree site model of elongation, [48] . The observation stream triggered the validation stream to investigate the corr ectness of the Nierhaus results, [63] . 7.6.1 Validating the Nierhaus m odel The validation stream ran a thought experiment to study the results coming from the Nierhaus thought experiment. The stream received the Nierhaus conclusion that the ribosome mosaic contained a third physically distinct binding site, the E site, as we ll as the A site and the P site, [48] . Having other lab reports observing the existence of the E site, the validation stream confirmed that the ribosome has three tRNA binding sites in existence, the A , P and E site , [63] .

PAGE 150

! )'" ! Further more , the validation stream received lab results confirming that deacylated tRNA is the only type of tRNA mosaic that binds to the E site. The validation stream accepted these rules , since they did not con tradict with any known lab data. The validation stream passed the new finding to the observation stream, which declared a new process rule: only deacylated tRNA can bind to the E site during the process of elongation. This process rule states that the deac ylated tRNA is constructed in the P site after it is unplugged from its peptide mosaic. The deacylated tRNA would not leave the ribosome from the P site , instead it would translocate, i.e. move, to the E site; then from that site it would leave the ribosom e. The observation stream saved the new process rule into the construction set and triggered the validation stream to continue evaluating the Nierhaus thought experiment. In Nierhaus' thought experiment, it was concluded that deacylated tRNA binds to the ribosome in the E site with codon anticodon hydrogen bonds. This conclusion was reached based on the observation of three tRNA mosaics binding to the ribosome mosaic with an mRNA mosaic attached to it. Only one tRNA mosaic was seen to bind to the P site in the ribosome in the absence of the mRNA mosaic. However, the validation stream had the outputs of the lab experiment showing two tRNA attached to the P site and the E site of a ribosome in the absence of the mRNA. The validation stream questioned the codo n anticodon bond at the E site. The stream received the inputs stating that, in the presence of mRNA, a ribosome with amino tRNA and peptide tRNA mosaics ( attached to the A site and the P site , respectively ) accepts a binding to deacylated tRNA only. Anoth er lab experiment detected the binding of two deacylated tRNAs into the ribosome in the absence of mRNA. With or without mRNA, the affinity constant of deacylated tRNA for the E site is th e same. The same affinity value means that the attachment degree of the deacylated tRNA to the ribosome is not affected by removing the mRNA from the

PAGE 151

! )'' ! ribosome. Finally, different lab experiments reported the absence of the codon anticodon attachment at the E site. All of these inputs guided the validation stream to invalid ate Nierhaus' conclusion that deacylated tRNA binds to the ribosome with the codon anticodon bonds. The stream further concluded that the deacylated tRNA binds to the ribosome with no codon anticodon and passed this conclusion to the observation stream. Th e observation stream saved into the construction set a new process rule stating that the deacylated tRNA occupies the E site without any attachment to the mRNA mosaic. The observation stream then triggered a new thought experiment to investigate the match ing rules associated with the E site. 7.6.2 Declaring the Matching Rules for Occupying the E Site The observation stream began a new thought experiment to study the nature of the E site, find its correct position inside the ribosome, and define the matching rule s that control the placement of the tRNA inside it. The stream received lab inputs confirming that the 50 s mosaic has only one binding site. This site was assumed to be part of the P site. However, the observation stream received lab results that spotted the binding of two deacylated tRNAs to the 50 s mosaic at the same time. While analyzing the Nierhaus model, the observation stream identified the process rules of the E site to be the location on a ribosome that only allows deacylated tRNAs to occupy it a nd that the deacylated tRNAs are independent from the mRNA, i.e., no hydrogen bonds are formed between mRNA and tRNA in the E site. The observation stream next re identified the location of the ribosome's three sites. Additional r esults from one lab exper iment reported that the E site is located in the 50 s part of the ribosome. Another lab experiment observed the existence of the P site and the A site in the 30 s part of the ribosome. This observation was based on the isolation of the 30 s mosaic

PAGE 152

! )'= ! with ami no tRNA and peptide tRNA bindings. Hence, the observation stream declared a new transformation rule stating that the 30 s mosaic contains two sites, the A and P site. Moreover, looking to the results that isolated 30 s mosaics with mRNA bindings, the obs ervation stream confirmed the binding place of mRNA to be on the 30 s mosaic. The stream also had new inputs, which indicted that the mRNA enters and leaves from the same site. Therefore, the observation stream saved into the construction set a new process rule indicating that mRNA runs inside the ribosome making a U turn shape inside the 30 s ribosome mosaic. Two transformation rules to build the ribosome were saved into the construction set. The first is that the ribosome is constructed out of the 30 s an d the 50 s mosaics. Secondly, the ribosome contains three binding sites that tRNA can occupy during translation. The A and P site are located inside the 30 s mosaic and the E site is located inside the 50 s mosaic. The process rule for the A site is simpl e. It only allows the entrance of the amino tRNA. The process rule of the P site allows all types of tRNA to occupy it. The E site is exclusive to deacylated tRNA. Hydrogen bonds are built, following the wobble rule, between the mRNA and tRNA only in the A and P site s . While in the E site, deacylated tRNAs are independent from the mRNA. The observation stream triggered the construction stream to build the ribosome applying the new matching rules. In order to build the ribosome, the construction stream gen erated the two ribosomal mosaics, the 30 s and the 50 s. The construction stream added three binding sites based on the process rules identified by the observation stream, the E site in the 50 s mosaic and the P site and the A site between the 50 s and the 30 s mosaics. At the same time, the expression stream tagged the actions of the construction stream. The expression stream translated every tag into an instruction

PAGE 153

! )'> ! aggregate and saved it in the "Construct the ribosome" shell corresponding to the process o f building the ribosome. Then, in the visual scene, the construction stream merged the two mosaics following the local matching rules, resulting in a bigger mosaic, the ribosome. The expression stream tagged every action executed by the construction strea m to build the ribosome and updated the shell as needed. Once the construction stream finished building the ribosome, the expression stream marked the "Construct ribosome" shell as the first shell to run it at any time the process of translation needed to be rerun. The final visual model of the ribosome built by the construction stream and the final corresponding shell can be seen in Figure 70 and Figure 71 respectfully. Figure 70 . The ribosome mosaic built by the construction stream and tagged by the expression stream . Figure 71 . The shell corresponded to the construction of the ribosome mosaic . Generate (50 s, construction set) Add (50 s, E site, construction set) Generate (30 s, construction set) Add (30 s, A site, construction set) Attach (30 s, 50 s, construction set) Add ( 30 s, P site, construction set) Construct ribosome

PAGE 154

! )'# ! 7.6.3 Reconstructing the Translation Process After collecting all of the necessary rules, the observation stream initiates the construction stream to start modifying the Watson model and visualizing the process of translation. In order to start the translation process, the construction stream called the visual models for the different mosaics involved in the process, which include the ribosome, the tRNA, and the mRNA mosaics. Next, the construction stream allowed the ribosome to attach to the mRNA according to the matching and process rules generated by the observation s tream. The construction stream visualized the mRNA movement inside the ribosome as a U turn shape, see Figure 72 . Immediately , the expression stream generated a "Link ribosome to mRNA" shell that includes all the instruction aggre gates explaining the current actions of the construction stream, see Figure 73 . Figure 72 . The mRNA mosaic takes a U turn shape inside the ribosome . Figure 73 . The "Link the ribo some to the mRNA" shell generated by the e xpression s tream . In order to create a moving scenario of the initiation process, the construction stream moved the ribosome along the mRNA , while reading the genetic code according to the reading rules identified in thought experiments in Chapter VI . Every time the ribosome had read one code Attach (mRNA, ribosome, construction set) Link ribosome to mRNA Move (mRNA, current location, exit location, movement rules)

PAGE 155

! )'$ ! aggregate in the mRNA, the initiating set had to be checked. Based on the lab inputs, translation starts only when the start codon aggregate is read by the ribosome. In previo us thought experiments, an initiation set, which included all of the start codon aggregates, was built. The only start codon aggregate within this set is "AUG." The construction stream received the initiation set as an input. If the codon aggregate did not belong to the initiating set, the construction stream would loop again to read another aggregate. The expression stream observed this loop and tagged the resulting actions. The expression stream again watched the actions of the construction stream runnin g the initiation process and saved the actions as instruction aggregates in the "Initiation" shell. Once the ribosome read the start codon aggregate, the ribosome would call for the amino tRNA carrying the correct anticodon according to the wobble compleme ntarity matching mosaic, see Table 11 . Since the initiation set only includes one start code aggregate, the corresponding fMet tRNA would always be the first amino tRNA to enter the ribosome. fMet tRNA is the amino tRNA mosaic wit h the fMet amino aggregate attached to it. The fMet tRNA would always occupy the P site to allow the next codon to occupy the A site of the ribosome. Once the fMet tRNA is in the P site, it would bind with the corresponding codon using its anticodon and fo llowing the wobble complementarity matching rules from Section 7.4.2 . The result of this binding is the creation of the codon anticodon aggregate between the fMet tRNA and the mRNA mosaics. While the construction stream ran the initiation process, the expression stream tagged the current shell as "Initiate." Figure 74 shows multiple snap shots of the initiation process as visualized by the construction stream . T he corresponding shell generated by the ex pression stream is given in Figure 75 .

PAGE 156

! )'% ! Figure 74 . Different snapshots of the initiation process by the construction stream : (1) The current codon is not an initiation codon the construction stream will re ad the next codon, (2) The current codon is an initiation codon, (2) call the initiator tRNA the fMet tRNA, and (4) place the fMet tRNA in the P site and create the codon anticodon aggregate. Figure 75 . The "Initiate" shell corres ponded to the process in Figure 74 . After the fMet tRNA is in place, the construction stream starts the elongation process by reading the mRNA codon aggregate that faces the A site of the ribosome following the associated reading rules. According to this reading, the construction stream first checked that the codon is not a termination codon. If it is not, the construction stream would call the ! " # " $ " % " Read (current codon, reading rules) Flag : = Compare (current c odon, initiation set) Else Call Start Translation If Flag then Call Initiation Start Translation Initiation Move (fMet tRNA, current location, P site, tRNA movement rules) Attach (codon, anti codon, Wobble matching rules)

PAGE 157

! )'& ! corresponding amino tRNA following the wobble complementarity mosaic, Table 11 . The incoming amino tRNA would enter the A site and would link, using its anticodon, to the codon aggregate in the mRNA. The linkage between the codon and anticodon aggregates must follow the wobble complementarity matching rules, Section 7.4.2 . Again, the expression stream monitored the whole process and tagged every movement run by the construction stream with a corresponding instruction aggregate. Once the incoming amino aggregate is in the A site, the const ruction stream reasoned that the next step must be transpeptidation , where the peptide tRNA donates its polypeptide mosaic to the amino tRNA. The translocation process triggered the movement of different mosaics following the process and movements rules, i .e. , translocation. The processes rules simultaneously allowed the peptide tRNA in the A site to move to the P site, the mRNA to move by one codon following the reading rules, and the deacylated tRNAs to unplug from its corresponding codon in the mRNA befo re it moves to the E site. At this point, the construction stream's state is the following: the A site is empty, the P site includes the peptide mosaic, and the E site includes the deacylated tRNA. The construction stream reasoned that the next step is to repeat the elongation process by reading the mRNA codon that faces the A site. The expression stream saved all of the instruction aggregates, that it defined by monitoring the construction stream, and finalized the current shell by naming it "Elongation." The expression stream then started to monitor the construction stream while repeating the process of elongation. The goal of the monitoring is to update the "Elongation" shell based on any modification of the visual model by the construction stream. Figure 76 shows snapshots of a single elongation cycle as built by the construction stream and Figure 77 shows the first version of the "Elongation" shell.

PAGE 158

! )=( ! Figure 76 . One cycle of the elongation process as visualized by the construction stream . Figure 77 . The "Elongation" shell corres ponded to the process in Figure 76 . ! ! # ! $ ! % ! & ! ' ! Read (curre nt codon, reading rules) Flag : = Compare (current codon, termination set) Call (amino tRNA, wobble complementarity matching rules) Move (amino tRNA, current location, A site, tRNA movement rules) Attach (codon, anti codon, wobble complementarity matchi ng rules) Attach (peptide, amino, protein matching rules) Move (mRNA, current location, one codon forward, mRNA movement rules) Move (deacylated tRNA, P site, E site, tRNA movement rules) Move (peptide tRNA, A site, P site, tRNA movement rules) Call E longation If Flag then Call Termination Elongation

PAGE 159

! )=) ! When the construction stream repeated the elongation process and re ached the point where the deacylated tRNA had to be translocated to the E site, the stream realized that the E site was occupied by another deacylated tRNA from the previous elongation cycle. The construction stream concluded that this old deacylated tRNA must be removed from the E site before the new deacylated tRNA could enter it. With no incoming inputs, the construction stream assumed two different methods to remove the previous deacylated tRNA. The first method is to allow the deacylated tRNA to be dis solved once it enters the E site. The other assumption is to push the old deacylated tRNA out of the ribosome immediately before the new deacylated tRNA enters the ribosome. Since, as previously mentioned, the expression stream is monitoring the process, i t tagged the new actions and added new instruction aggregates to the shell. The expression stream identified that the "Elongation" shells need to include the two possibilities of removing the deacylated tRNA from the ribosome. The two resulting "Elongatio n" shells are shown in Figure 78 and Figure 79 . Figure 78 . The first scenario for a new version of the " Elongation " s hell . Read (current codon, reading rules) Flag : = Compare (current codon, termination set) Call (amino tRNA, wobble complementarity matching rules) Move (amino tRNA, current location, A site, tRNA movemen t rules) Attach (codon, anti codon, wobble complementarity matching rules) Attach (peptide, amino, protein matching rules) Move (mRNA, current location, one codon forward, mRNA movement rules) Move (deacylated tRNA, P site, E site, tRNA movement rules) Move (peptide tRNA, A site, P site, tRNA movement rules) Call Elongation If Flag then Call Termination Elongation Dissolve (deacylated tRNA, E site)

PAGE 160

! )=" ! Figure 79 . The second scenario for a new version of the "Elongation" s hell . The construction stream repeated the elongation process until it read a termination codon; at this point the construction stream ended the elongation phase and entered the termination phase, see Figure 80 . The construction stream terminated the process of translation by detaching the two ribosomal mosaics from each other, disassembling the mRNA, freeing all the tRNA mosaics, and releasing the new peptide mosaic from the peptide mosaic, Figure 81 . In parallel with the construction stream terminating the process, the expression stream tagged all of the steps completed by the construction stream and saved them in the "Termination" shell, depicted in Figure 82 . Figure 80 . Reading a termination codon . Read (current codon, reading rules) Flag : = Compare (current codon, termination set) Call (amino t RNA, wobble complementarity matching rules) Move (amino tRNA, current location, A site, tRNA movement rules) Attach (codon, anti codon, wobble complementarity matching rules) Attach (peptide, amino, protein matching rules) Move (mRNA, current location, one codon forward, mRNA movement rules) Move (deacylated tRNA, P site, E site, tRNA movement rules) Move (peptide tRNA, A site, P site, tRNA movement rules) Call Elongation If Flag then Call Termination Elongation Move (deacylated tRNA, E site, outs ide, tRNA movement rules)

PAGE 161

! )=' ! Figure 81 . Terminating protein synthesis . Figure 82 . The "Termination" s hell . Once the construction stream ended the c onstruction of the translation process, the expression stream grouped all of the shells generated during the process. The expression stream then created an algorithmic mosaic by grouping these shells. The new algorithmic mosaic describes exactly how to re run all of the Watson extension thought experiments. 7.7.0 The Alternative Model: Rebuilding the Elongation Model In this section, the work of Nierhaus and Rheinberger in [48] and [49] were visited. At t he beginning of this thought experiment, the AD initiated an observation stream to collect the necessary data relating to protein synthesis. The observation stream, in its preliminary stage, collected all inputs related to the Watson model. In this model, the ribosome moves one codon further towards the 3' end of the mRNA mosaic each time an amino acid is added to the peptide Termination Detach (ribosome, deacylated tRNA ) Detach (30 s, 50 s ) Detach ( 50 s, peptide) Detach ( 30 s, mRNA) Detach (ribosome, peptide tRNA )

PAGE 162

! ) == ! mosaic. The observation stream also accepted the fact that the ribosome contains two different sites that tRNA mosaics may occupy, t he A site for amino tRNA mosaics and the P site for peptide tRNA mosaics. Next , the observation stream saved in the construction set the process rules defined during the Watson thought experiments and from the later experiments examining the three stages o f protein translation: initiation, elongation, and termination. The stream also accepted the new inputs about the three stages of translation. At the beginning of the initiation phase, the different mosaics that participate in protein translation must be a ssembled. The two ribosomal mosaics, 50 s and 30 s, first link to each other, then the ribosome attaches to the mRNA mosaic. The ribosome reads the genetic code on the mRNA codon by codon until it reaches the start codon, a special type of codon aggregate belonging to the initiation set. The initiation set is a subset of the codon set that includes all possible codon aggregates. Once an initiation codon aggregate is found, the start, or initiated, amino tRNA is called to occupy the P site. The start amino tRNA is the fMet tRNA mosaic. After the fMet tRNA occupies the P site, the initiation stage ends and the elongation phase starts , [48] , and [49] . The observation stream received inputs from lab ex periments which found that, at the beginning of every elongation cycle, two factor aggregates, EF Tu and GDP, must connect to each other to construct the EF Tu GDP aggregate. The EF Tu GDP aggregate transforms into EF Tu GTP after a chemical interaction. At the same time, the tRNA mosaic is energized and transformed into the amino tRNA mosaic following the process rules from Section 7.3.2 . The amino tRNA mosaic then interacts with the EF Tu GTP aggregate to form the EF Tu GTP a mino tRNA aggregate that will enter the ribosome and bind to the A site with the help of the EF Tu and GTP factors.

PAGE 163

! )=> ! The amino tRNA entering the ribosome must decode for the correct codon aggregate following the codon anticodon wobble complementarity match ing rules in Section 7.4.2 . The lab results reported that, after the binding, the EF Tu GTP aggregate detaches from EF Tu GTP amino tRNA. Next , the amino tRNA in the A site will be ready to accept the growing peptide mosaic fro m the peptide mosaic in the P site. The peptidyltransferase, which is the integral part of the ribosome's 50 s mosaic, will release the peptide mosaic from the peptide tRNA in the P site and transfer it to bind to the amino aggregate attached to the amino tRNA in the A site. After the peptide transfer, the tRNA located in the P site becomes deacylated tRNA and the tRNA located in the A site becomes a peptide tRNA. Once the peptide mosaic transfer is completed, the two tRNA mosaics translocate. The deacylat ed tRNA in the P site leaves the ribosome and the peptide mosaic in the A site moves to the P site. The elongation cycle will continue until a termination codon aggregate is read in the mRNA mosaic ending the elongation phase. Termination codons are nonsen se codon 5 aggregates that are saved in the termination set, a subset of the codon set that includes all possible codon aggregates. When the elongation stage ends, the ribosome enters the termination phase. The new peptide mosaic is separated from its tRNA and released from the ribosome. Later, the resulting peptide mosaic is folded to form an active protein mosaic. Simultaneously, the 30 s and the 50 s mosaics are disconnected from each other and the mRNA detaches from the current ribosome. The observation stream then analyzed the incoming inputs. The stream accepted the rules controlling the initiation and termination phase identified by previous thought experiments. Lab results defined the different factors that are needed for each one of the protein phas es. These factors bind to the ribosome to induce the different reactions needed to complete the protein !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! !!!!!!!!!!!!!!!!!!!! ! 5 Nonsense codons are the codons that do not signal to any amino tRNA, [44] .

PAGE 164

! )=# ! synthesis. Three groups of factors were identified: initiation, termination, and elongation factors. These factors are essential to attaching or detach ing different mosaics. The observation stream reasoned that these factors must be part of local transformation rules for linking or unlinking the translation mosaics. The observation stream also accepted the identification of a ribosome mosaic from the Wat son thought experiment, see Section 7.5.0 . The stream declared three types of tRNA based on the inputs from the Watson model: amino tRNA, peptide tRNA, and deacylated tRNA. Following the Watson model, only one type of tRNA can bind to the ribosome after translocation (i.e. , when the peptide tRNA occupies the P site). The observation stream triggered a new thought experiment to define the matching rules that control the process of translation. 7.7.1 Preparing the Construction Set for Elongation: the Observation Stream Based on inputs coming from different lab and thought experiments, including the thought experiments in Section 7.6.0 , the observation stream identified the ribosome mosaic to consist of two d ifferent mosaics with three binding sites. According to incoming inputs, the stream declared the 30 s mosaic to have two binding sites, an A site an d a P site, and the 50 s mosaic to have three binding sites, an A site, a P site, and an E site. In order to construct the ribosome, the P site and A site of the two ribosomal mosaics must be aligned. Other lab inputs found that tRNA mosaics occupied the three binding sites in this order: P site, E site, A site. The observation stream defined a new process rule stating that the amino tRNA could bind to the A site and the P site simultaneously, while peptide tRNA could only bind to one site at a time, either the P site or the A site. Next, the observation stream declared the exclusion rule, as part of the proces s rules, which prevents two peptide tRNA mosaics from linking to the ribosome at the same time. The

PAGE 165

! )=$ ! exclusion rule provides a channel for the constructed peptide mosaic. Only one peptide mosaic could enter this channel for the current ribosome. Then, the o bservation stream re identified the rules that would be applied during the elongation process. The observation stream accepted the elongation factors stating that at the beginning of every elongation cycle, the EF Tu and the GDP factor aggregates must be c onnected to construct the EF Tu GDP aggregate that would be transformed later into the EF Tu GTP. The observation stream also accepted the transformation rules for the amino tRNA from Section 7.3.3 . The observation stream saved these rules into the construction set as part of the elongation process. The observation stream then identified the process rules that would control the movement of the different types of tRNA mosaics. Since the peptide tRNA always donates its peptide to the new amino tRNA during elongation, the observation stream reasoned that the deacylated tRNA would not have an entry site to the ribosome. Instead, the deacylated tRNA mosaic would be generated after the peptide tRNA detached from its peptide mosaic. Fur ther, the observation stream declared that deacylated tRNA could bind to the three binding sites of the ribosome in this order: first, occupy the P site; second, move to the E site; finally move to the A site. Next, the observation stream defined the proc ess rules for the amino tRNA movement as follows: the entrance site for amino tRNA would be the A site, however , it would not bind to the P site immediately. The amino tRNA would bin d to the P site only after it moved from the A site. On the other hand, am ino tRNA could bind to the A site immediately without occupying the P site first.

PAGE 166

! )=% ! Finally, the observation stream identified the process rules for the peptide tRNA movement as the following: peptide tRNA could bind directly to the P site, and it could bin d directly to the A site if the P site were occupied with any tRNA type. At the end of this thought experiment, the observation stream saved the process rules of the different tRNA mosaic movements to the elongation construction set. The stream then initi ated a thought experiment to reconstruct the elongation cycle based on these new findings and sent it to the construction stream. 7.7.2 Reconstructing Elongation The construction stream visualized a new model for the elongation process taking into account the c onstruction set including all of the rules declared by the observation stream. First, the construction stream built the new structure of the ribosome. The stream generated a general shape of the two ribosomal mosaics , the 30 s includes two binding sites wh ile the 50 s includes three. The stream then attached the two ribosomal mosaics. In order to link the 30 s mosaic to the 50 s mosaic, the construction stream found that the two mosaics must first be aligned. The construction stream identified the alignmen t rules as follows: the A sites of the two mosaics must be aligned. The alignment rules are part of transformation rules used to transform one mosaic in to another. Similarly, the P sites of the two mosaics must be aligned. After the alignment of the two mo saics, the construction stream joined them following the local matching rules. The result was a ribosome with three binding sites : the A site, the P site and the E site. The A site and the P site each have two parts, one in the 30 s and the other in the 50 s. The E site is located next to the P site towards the 5` end of the mRNA mosaic, as shown in Figure 83 .

PAGE 167

! )=& ! Figure 83 . The new structure of the ribosome mosaic . In parallel, the expression stream observed the actions of the construction stream and tagged the steps of building the ribosome. The result of this tagging is a shell that contained all the instruction aggregates needed to construct the ribosome at any time. Once the construction stream finished b uilding the ribosome, the expression stream saved the current shell as "Construct ribosome" to be used later in building the algorithmic mosaic of elongation, see Figure 84 . Figure 84 . The "Construct ribo some" shell generated by the expression stream . Next, the construction stream retrieved the declaration for tRNA and mRNA mosaics from the construction set. The construction stream reasoned that since two tRNA mosaics can occupy the A site and the P site at the same, the codon anticodon interaction must be maintained Construct ribosome Generate (50 s, construction set) Add (50 s, E site, constr uction set) Add (30 s, A site, construction set) Generate (30 s, construction set) Attach (30 s, 50 s, construction set) Add (30 s, P site, construction set) Add (50 s, A site, construction set) Add (50 s, P site, construction set) Align (30 s, 50 s , alignment rules)

PAGE 168

! )>( ! simultaneously. Therefore, the stream concluded that the tRNA structure was L shaped. If the tRNA took the L shape, then it would be possible to have two adjacent tRNA mosaics linked at the sa me time to the codon aggregate in the mRNA. The stream received data from Rich, [2] , that the tRNA mosaic diameter must be 20 A¡ and the codon aggregate length must be 10 A¡ ; based on this new definition, the construction strea m re identified the body of the tRNA and saved the new finding in the construction set to build the tRNA. The visual model of the three types of tRNA mosaics can be seen in Figure 85 . The expression stream observed the steps of bu ilding the different types of tRNA and saved every one of them in a separate shell. These shells did not differ from the shells generated by the Watson extension model, in Section 7.6.0 . The only difference was the "Construct t RNA body" shell with the new L shaped body of the tRNA mosaic, as shown in Figure 69 , [48] , and [49] . Figure 85 . Three different types of t RNA mosa ics as visualized by the construction stream . The construction stream also found that, to make the peptide mosaic translocation possible, the two tRNA mosaics in the A site and the P site must have their 3' ends close together in the center of the peptidyl transferase position in the ribosome. Further, the construction stream received an input, showing that the mRNA mosaic enters and exits the ribosome from almost the same position. This led the construction stream to conclude that the mRNA mosaic must turn inside the ribosome at least once. The construction stream saved the definition for building the mRNA mosaic, see Section 6.6.1 . The stream started to build a visual model of the mRNA inside the ribosome. Figure 86 shows snapshots of linking the mRNA to the ribosome. These steps did

PAGE 169

! )>) ! not differ from the way the mRNA enters the ribosome in the Watson extension model. Therefore, the shell generated by the expression stream of this visual scene was the same as the one in Figure 73 . Figure 86 . The mRNA enters the ribosome . Subsequently, the construction stream mimicked the process rules controlling the movements of different mosaics during elongation. The f irst rule declared that two tRNA mosaics have to be attached to the ribosome before and after translocation. Also, when the deacylated tRNA moves from the P site to the E site, it must stay attached to the ribosome. Further, based on the Watson model, the construction stream found that the tRNA's release is induced by the EF G factor and the peptide tRNA following a one dimensional diffusion from the A site to the P site. The construction stream rejected this rule and re identified a new movement rule that allows the deacylated tRNA in the E site to be released from the ribosome only when a new amino tRNA enters the A site. Then the stream mimicked the initiation stage of translation. Since no new inputs were received that contradict the initiation process defined in Watson extension model, see Section 7.6.0 , the result of constructing the visual model of initiation and the corresponding shell were similar to the Watson extension model, excluding the ribosome visual model, in Figure 74 and Figure 75 . Elongation started by reading the codon in front of the A site. The ribosome then called the corresponding amino tRNA to enter empty A site. Before the amino tRNA could enter the

PAGE 170

! )>" ! r ibosome, it must be attached to the elongation factors. The construction stream built a visual model that represents the process of preparing the elongation factors and attaching them to the amino tRNA that would enter next to the A site, as seen Figure 67 . Figure 87 . The starting cycle of elongation . After that, the EF Tu GTP amino tRNA entered the A site based on the reading of the adjacent code aggregate in the mRNA, see Figure 88 . The construction set in the codon anticodon rules was retrieved from the Crick thought experiment in Section 7.4.2 . Following the process rules, the amino tRNA would be ready to accept the peptide mosaic in the P site aft er it unlinked from the EF Tu GTP aggregate. Once EF Tu GTP aggregate is released, translocation occurs. During translocation, the construction stream would link the peptide mosaic in the P site to the amino aggregate in the A site, as in Figure 89 .

PAGE 171

! )>' ! Figure 88 . The amino tRNA is ready to accept the peptide mosaic . Figure 89 . Peptidyltransferase . After translocation, the A site contained the new peptide tRNA and the P site contained t he deacylated tRNA. Next, these two tRNA mosaics moved to the next site, the peptide tRNA to the P site and the deacylated tRNA to the E site, see Figure 90 . The last step of this elongation cycle occurred when the new amino tRNA entered the A site and the deacylated tRNA exited the E site, shown in Figure 91 and Figure 92 . The elongation stage ended when the ribosome read a termination codon in the mRNA and the new peptide mosaic left the ribosome.

PAGE 172

! )>= ! At every step in constructing the elongation process, the expression stream tagged and saved the actions, applied by the construction stream, in the "Elongation" shell. The expression stream generated the complete algorithmic mosaic ass ociated with the alternative model, see Figure 93 . Figure 90 . The tRNA m osaics t ranslocation . Figure 91 . The new amino tRNA is ready to enter the elongation cycle and the deacylate d tRNA is ready to move out of the cycle.

PAGE 173

! )>> ! Figure 92 . The beginning of a new elongation cycle . Figure 93 . The algorithmic mosaic for the alternative elongation model built by the expression stream . Build (EF Ts GTP, construction set) Read (current codon, reading rules) Flag : = Compare (current codon, termination set) Call (amino tRNA, wobble complementarity matching rules ) Build (EF TU GDP, construction set) Attach (EF Ts GT P, EF TU GDP, Matching rules) Attach (amino tRNA, EF TU GTP, complementarity matching rules) Move (EF TU GTP amino tRNA, current location, A site, tRNA movement rules) Attach (codon, anti codon, Wobble matching rules) Attach (peptide, amino, protein ma tching rules) Trigger (EF TU GTP) Move (deacylated tRNA, P site, E site, tRNA movement rules) Move (peptide tRNA, A site, P site, tRNA movement rules) Move (deacylated tRNA, E site, outside, tRNA movement rules) If Flag then Call Termination Elongati on Move (mRNA, current location, one co don forward, mRNA movement rules) Call Elongation Bring closer (peptide, amino)

PAGE 174

! )># ! 7.7.3 Verifying the New Three site Model: The Validation Stream The validation stream set out to evaluate the correctness of the new three site model proposed by the construction stream. The validation stream first checked the correctness of the L shape of the mRNA mosai c inside the ribosome. The alternative three site model was built based on the input showing two adjacent codon anticodon linkages between the mRNA and two tRNA mosaics at the same time. The validation stream first looked for data indicating that the tRNA mosaic pulls the mRNA mosaic through the ribosome. Also, lab evidence showed that various tRNA mosaics could be translocated inside the ribosome without interaction with the mRNA mosaic. The validation stream reasoned that if the tRNA mosaic is responsibl e for pulling the mRNA mosaic through the ribosome, then having two tRNA mosaics interacting with the mRNA mosaic is important for the mRNA mosaic's movement. The mRNA tRNA linkage must be preserved throughout the translocation step. The validation stream found that it is not realistic to lose half of the interaction with the mRNA mosaic by releasing the bindings of one of the tRNA mosaics before the translocation as was identified by the Watson model. Also, the validation stream considered keeping the mRNA interaction with the two tRNA mosaics after the translocation , since that would set the mRNA mosaic. Setting the mRNA is important to facilitate the specified exposure between the codon and anticodon aggregates, i.e., between the mRNA and tRNA mosaics, at the A site. The correct linkage between codon and anticodon aggregates is essential for correct decoding of the genetic code. Further, the validation stream received information that the codon anticodon linkage at the P site increases the decoding accurac y. Therefore, the stream reasoned that having two codon anticodon connections at the same time would be very important for moving the mRNA mosaic and increasing the decoding accuracy.

PAGE 175

! )>$ ! Further, the validation stream reasoned that the mRNA's turn inside the ribosome proved the three site model, because this turn allows the adjacent codon anticodon connections pre and post translocation required in this model. The validation stream justified that the mRNA turn must occur in the region of the mRNA that is used in decoding the genetic code. Therefore, the stream concluded the correctness of the mRNA's L shape inside the ribosome. Next , the validation stream investigated the possibility of having four binding sites in the ribosome , where amino tRNA could bind if a peptide tRNA bound to the A site. The stream received inputs from the Rheinberger and Nierhaus lab results that excluded any possibility of a fourth binding site for the tRNA mosaics during elongation, [49] . The validation st ream did not find any new inputs contradicting the Rheinberger and Nierhaus conclusions. Thus, the validation stream verified the three sites existence in the ribosome. Finally, the validation stream declared the correctness of the new model built by the c onstruction stream in Section 7.7.2 and named it the alternative model. 7.8.0 The Hybrid Model of Elongation In this section, a series of thought experiments were created based on the work of Moazed, and Noller in [32] , which resulted in a new elongation model, the hybrid model. For these thought experiments, the AD was initiated to find the correct process of protein translation. The AD started by running the observation stream to the retrieve current ly k nown inputs about protein synthesis. First, the observation stream retrieved the construction set and the visual models needed to build the different types of tRNA mosaic, the ribosome, and the mRNA from previous thought experiments including the Watson extension results, Section 7.6.0 . The stream then received the inputs that divided the translation algorithm into three smaller processes, which run sequentially

PAGE 176

! )>% ! to construct the peptide mosaic. The first process requires the o bservation stream to focus all streams to identify the rules that control the codon anticodon recognition process, Section 7.4.0 . The second requirement is to find the peptide bond formation process. The third constraint is to look for the process rules that guide the movement of the tRNA and mRNA, the ribosome shifting, and the peptide creation. The observation stream found that the problem has three different branches to examine, the first two of which were known, thus the obs ervation stream saved into the construction set the rules that control the codon anticodon recognition and peptide bond formation , [32] . The observation stream then focused itself to find the matching rules, which control the t ranslocation or movements of tRNA mosaics during the elongation process. The observation stream accepted all possible visual inputs about translocation. The first obvious input was the Watson two site model, Section 7.5.0 . The observation stream also saved the inputs about the three binding sites of the ribosome, the A , P , and E site. The stream found inputs from a lab experiment reporting that amino tRNA mosaics, after binding to the ribosome, do not participate in the pepti de construction until the release of the EF Tu aggregate. The observation stream began to investigate the possibility of having a fourth binding site in the ribosome. It also began to determine the process rules that control the entrance of the tRNA mosaic s into the ribosome's binding sites. In order to study the movement of the tRNA inside the ribosome, the observation stream received information stating that tRNA interacts with the RNA of the ribosome, rRNA. Several in vitro systems tested the tRNA bindi ng to the ribosome and reported that the rRNA bases became protected when the tRNA is bound to the ribosome. During monitoring of the tRNA binding to the P site or A site of the 30 s mosaic within the ribosome, the lab results confirmed

PAGE 177

! )>& ! that certain bases of rRNA became protected. Likewise, when tRNA binds to the P , A , or E sites of the 50 s ribosome, another set of rRNA bases becomes protected. The observation stream suspected that those bases are directly involved when tRNA binds to the ribosome and rea soned that if the rRNA state is examined, in the lab, before and after formation of the current peptide mosaic, then the chemical changes in the rRNA base will determine the movement of the tRNA inside the ribosome. 7.8.1 Declaring the First Process Rule: The O bservation Stream In order to observe the movement of the tRNA mosaics inside the ribosome, the observation stream accepted the inputs of a lab experiment examining the binding state of the P site before and after the formation of the peptide mosaic. The lab results showed that before the peptide formation, the bases of the P site most likely bind to tRNA mosaic and were protected by this mosaic. The protected bases were found in the 30 s and 50 s of the ribosome. However, it was reported that the rRNA bas es on the 50 s lose their protection after peptide formation. The observation stream reasoned that losing rRNA protection in the P site was not a result of moving the peptide mosaic to the A site. The resulting deacylated tRNA in the P site could still pro tect the rRNA bases. Further, lab inputs detected footprints of the rRNA bases in the P site of the 30 s continuing after peptide b ond formation. The rRNA protected bases in the 30 s P site persisted after the formation of peptide bond. The lab experiment also found an E site footprint after the transfer of the peptide mosaic. Having all these inputs, the observation stream concluded that after the peptide formation, deacylated tRNA moved partially from the P site to the E site. The observation stream reac hed this conclusion based on the absence of the P site's rRNA footprints in the 50 s mosaic and the presence of the P site's rRNA footprints in the 30 s mosaic.

PAGE 178

! )#( ! The observation stream declared a new process rule controlling the movement of deacylated tRNA after donation of the peptide mosaic. The observation stream found that tRNA mosaics move in states between the binding sites of the ribosome. The previous lab results showed that deacylated tRNA can occupy the E site and the P site at the same time, thus the observation stream declared that placement of the tRNA inside the ribosome must be described with regard to both binding sites of the 30 s and the 50 s mosaics. For example, in all previous elongation models, the placement of the tRNA inside the ribos ome might be declared as the A site, the P site, or the E site. However, in the hybrid model, if any tRNA mosaic occupies the A site of both ribosomal mosaics, then this tRNA must be declared to be in the A/A state. Similarly, the tRNA in both P sites will be declared to be in the P/P state and the tRNA in the E site will be in the E/E state. Figure 94 shows the three classical states of the tRNA mosaics inside the ribosome. Figure 94 . The three classical states the tRNA mosaics can move to during elongation. Based on lab results, the observation stream identified a new state of the deacylated tRNA, the P/E state. In the P/E state, the deacylated tRNA binds to the E site of the 50 s mosaic and to the P sit e of the 30 s mosaic, see Figure 95 . The stream concluded that deacylated tRNA would enter the P/E state after the peptide transformation from the P site to the A site. The observation stream saved this new state as part of the pr ocess rules controlling the movement of deacylated

PAGE 179

! )#) ! tRNA mosaics inside the ribosome. Next, t he observation stream initiated the next thought experiment to find all possible states tRNA mosaics can move to during the elongation process . Figure 95 . The P/E state where deacylated tRNA occupied the P site of the 30 s mosaic and the E site of the 50 s mosaic at the same time. 7.8.2 The Observation Stream: Declaring the Second Process Rule Following the previous thought experiment , the observat ion stream began to find the process rules controlling the translocation of the tRNA mosaics from one state to another. The observation stream looked for results from a lab experiment with a much more complicated technique to examine all possible states of the tRNA mosaics during translocation. The impression stream transformed the written results of this lab experiment into visual streams and forwarded them to the observation stream. The lab results confirmed that before peptide mosaic transformation, the rRNA bases of the P site in both the 30 s and the 50 s ribosomal mosaics were protected. These results led the observation stream to reason that the peptide tRNA must be occupying the P site of both 30 s and the 50 s mosaics before the actual peptide trans formation, i.e., the peptide tRNA would be in the P/P state. The stream saved into the construction set a new process rule associated with the

PAGE 180

! )#" ! P/P state: Before translocation, the peptide tRNA would be attached simultaneously to the P site of both the 30 s and the 50 s. In addition, the lab experiment found protected rRNA bases in the E site during the transfer of the peptide mosaic. Also, footprints of the rRNA bases were detected in the P site of both ribosome mosaics. In contrast, the same lab experimen t reported the protection of rRNA bases in the A site in the 30 s and the loss of the rRNA bases of the 50 s mosaic. In other words, during or after the transformation of the peptide mosaic, the rRNA footprints of the A site in the 30 s, the P site in the 30 s, the P site in 50 s, and E site in 50 s were all seen , while those of the A site in the 50 s were not found. These results led the observation stream to confirm that after the peptide mosaic transfers, the tRNA mosaic in the P/P state moves to the P/ E state. The observation stream concluded that, after the transfer of the peptide mosaic, the peptide tRNA in the A site moves to the A/P state, where the peptide tRNA resides in the 30 s's A site and the 50 s's P site simultaneously, as seen in Figure 96 . The observation saved this new state with its rules into the construction set as part of the process rules of elongation. Figure 96 . The A/E state where peptide tRNA occupied the A site of the 30 s mosaic and the P site of the 50 s mosaic simultaneously.

PAGE 181

! )#' ! The lab results showed that after the translocation of the peptide mosaics, the footprints of the 30 s's A site were abolished , while the 30 s P site and the 50 s P site and E site bases were still protect ed. The observation stream concluded the last step of peptide translocation is for the peptide tRNA to move from the A/P state completely to the P/P state, i.e. , it moves out of the A site in the 30 s and resides completely in the P sites in the 50 s and 3 0 s. The observation stream reasoned that the peptide tRNA movement to the P/P state requires the deacylated tRNA to empty the P site and move from the P/E state completely to the E/E state without leaving the ribosome. The stream saved, in the constructio n set, the process rules associated with the P/P and P/E states. The construction stream would apply the defined states during construction of the visual model of the translation process. 7.8.3 The Observation Stream: Declaring More Process Rules In order to con tinue the investigation of all possible states for tRNA movements and their associating process rules, the observation stream received information from a lab experiment confirming that the existence of deacylated tRNA inside the ribosome prevents the pepti de tRNA from binding completely to P site, i.e., moving to the P/P state. The stream reasoned that these lab results indicated that the peptide tRNA mosaics bind initially to the A site before they translocate to the P site , [32] . Further, the observation stream monitored a lab experiment that initially inserted the deacylated tRNA into the P site, and then bound the peptide tRNA to the ribosome. The lab experiment then examined the ribosome rRNA bases to see where this binding occurred. The results confirmed the previous lab findings. The observation stream concluded that the initial binding of peptide tRNAs during elongation must be to the A site.

PAGE 182

! )#= ! Then, the observation stream received input from two other lab experiments , [32] . The first one examined the locations of deacylated tRNA and peptide tRNA before the translocation and the other tested their locations after wards . The first lab experiment started with a ribosome with a peptide tRNA linked t o its A site. During the experiment, a deacylated tRNA was allowed to link to this ribosome to find all possible locations that this deacylated tRNA could occupy before the translocation. The results of the first experiment showed that some of the rRNA bas es in the E site of the 50 s were protected. At the same time, footprints of the P site's 30 s were detected. Also, a fragile protection of the rRNA in the P site's 50 s was found. The observation stream confirmed that before the translocation, the deacyla ted tRNA would bind initially in the P/E state. Moreover, after binding the peptide tRNA would not bind to the ribosome, with its P site blocked since no tRNA could occupy the P site. The results found that some ribosomal rRNA bases located in the A site w ere protected. Also, footprints of the rRNA bases in the P site of the 30 s and the E site of the 50 s were found. On the other hand, the A site of the 50 s included unprotected rRNA. The observation stream concluded that, before the translocation, while t he deacylated tRNA occupies the P/E state, the peptide tRNA would stay in the A/P state , where part of it enters the A site of the 30 s and the other part remains in the P site of the 50 s. Once the translocations of the tRNA mosaics were performed, with t he help of the EF G and GTP lab factors, the lab results did not detect any footprints of the A site in ribosomal mosaics. The footprints of the E site and the P site were detected. The observation stream reasoned that, after the translocation, the peptide tRNA moved from the A/P state to the P/P state. For the peptide mosaic to completely occupy the P site, the observation stream also reasoned that the deacylated tRNA must leave the P site and move from the P/E state into the E/E state.

PAGE 183

! )#> ! 7.8.4 Building the Model: The Construction Stream The observation stream identified five states that the tRNA can move through during elongation. The three standard states, which are the A/A, P/P and E/E states, as well as the A/P state and the P/E state. After saving all of the s tates and the process rules controlling every state into the construction set, the observation stream triggered the construction stream to visualize the elongation process applying the new ly defined rules. The construction stream focused itself on buildin g a new model mimicking the movement of tRNA mosaics during elongation. The initiation and termination processes were the same as those in the Watson extension model. The construction stream also used the rules of the Watson extension model, to include the reading of the codons, calling the correct tRNA, linking between the tRNA and the mRNA, building the peptide mosaic, constructing the mRNA and the tRNA mosaics, attaching the mRNA to the ribosome, and moving the mRNA and the ribosome mosaics. In order to visualize the entire process, the construction stream started the creation of the hybrid model with the previous knowledge of how to construct the ribosome's two mosaics, the different types of tRNA mosaics, and the mRNA mosaic. Initially, the construction stream built the ribosome out of two mosaics: the 30 s mosaic with two different sites (the A site and the P site) and the 50 s mosaic with an A site, P site and E site inside it, see Figure 97 . The construction stream built a ge neral shape of the mRNA mosaic and the different types of the tRNA mosaics in a manner similar to the construction stream of the Watson extension model. At every step, the expression stream was observing and tagging the process to generate the correspondin g shells of every thought experiment. The result was a shell similar to the one in the alternative model.

PAGE 184

! )## ! Figure 97 . The ribosome mosaic built by the construction stream . In order to start the elongation process, the construction stream built a start state , where the ribosome had already linked to the starting fMet tRNA in the P/P state and the mRNA in a U turn shape, see Figure 98 . Next , the construction stream began to mimic the elongation process by rea ding the current codon, according to the reading rules, see Section 6.5.0 , and calling the correct amino tRNA, based on the wobble complementarity mosaic, see Table 11 . Figure 98 . The i nitial s cene to s tart e longation . The construction stream looked to incoming lab results, via the impression stream, showing that when amino tRNA enters elongation it links to the EF Tu GTP aggregate. The construction stream built the EF Tu GTP a mino tRNA mosaic by binding the amino tRNA mosaic into the EF Tu GTP aggregate, which is constructed out of three tiles (EF, Tu, and GTP). The stream

PAGE 185

! )#$ ! allowed the amino tRNA portion of the EF Tu GTP amino tRNA mosaic to enter the A site of the 30 s ribosome and the EF Tu GTP portion to enter a special site (the stream named it the T site). The construction stream declared a new state, the A/T state and set the process rules such that this state is the first state for the amino tRNA during elongation, as show n in Figure 99 . When the amino tRNA is in the A/T state, the peptide tRNA would be shielded from the P site. At every step of applying the A/T state to elongation, the expression stream was tagging the process to b uild the elongation shell. The part of the elongation shell generated by the expression stream for this step is depicted in Figure 100 . Figure 99 . The EF Tu GTP amino tRNA is in the A/T state. Figure 100 . The algorithmic mosaic explaining the process in Figure 99 . At the end of the A/T state, the EF Tu and the GTP are disconnected from EF Tu GTP amino tRNA mosaic. Once the EF Tu GTP is released, the amino t RNA mosaic is allowed to Build (EF Ts GTP, construction set) Read (current codon, reading rules) Flag : = Compare (current codon, termination set) Call (ami no tRNA, wobble complementarity matching rules ) Build (EF TU GDP, construction set) Attach (EF Ts GTP, EF TU GDP, Matching rules) Move (EF TU GTP amino tRNA, current location, A/T state, tRNA movement rules) If Flag then Call Termination Attach (amin o tRNA, EF TU GTP, complementarity matching rules) E lon gation

PAGE 186

! )#% ! move completely to the A site of 50 s mosaic. At that time, the amino tRNA moves from the A/T state into the A/A state. When the amino tRNA is in the A/A state, it linked to the tRNA following the wobble matching rules, shown in Figure 101 . The expression stream continued tagging the movements of the different mosaics inside the visual thought experiment. Every new tag was added by the expression stream as an instruction aggregate to the elongation shell, s ee Figure 102 . Figure 101 . The amino tRNA moves from the A/T state to the A/A state.

PAGE 187

! )#& ! Figure 102 . The elongation shell after translating the operations from Figure 101 . The construction stream then reasoned that the peptide mosaic connected to the peptide tRNA must link to the amino aggregate in the amino tRNA. After the peptide mosaic movement, the peptide tRNA would be in the A/P state and the deacylated tRNA would be in the P/E state, see Figure 103 . Again, the expression stream added the operations explaining these steps to the elongation sell, as shown in Figure 104 . Figure 103 . The peptide mosaic links to the amino aggregates and the deacylated tRNA and the peptide tRNA moves to the states P/E and A/P respectively. Elongation Build (EF Ts GTP, construction set) Read (current codon, reading rules) Flag : = Compare (current codon, termination set) Call (amino tRNA, wobble complementarity matching rules ) Build (EF TU GDP, construction set) Attach (EF Ts GTP, EF TU GDP, Matching rules) Move (EF TU GTP amino tRNA, current location, A/T state, tRNA movement rules) If Flag then Call Termination Attach (amino tRNA, EF TU GTP, complementarity matching rul es) Detach (EF TU GTP, amino tRNA) Attach (codon, anti codon, Wobble matching rules) Move (amino tRNA, current location, A/A state, tRNA movement rules)

PAGE 188

! )$( ! Figure 104 . The expression stream continues building the elongation shell by watching the visual movie run by the construction stream. Finally, with the help of the EF G GTP factor, the deacylated tRNA moved completely to the E site and the peptide tRNA moved to the P sites in both the 50 s and 30 s mosaics. At this time, the deacylated tRNA i s in the E/E state and the peptide tRNA is in the P/P state. Simultaneously, the mRNA moved one codon toward the mRNA 3` end. By then, the elongation cycle has reached its end. I t would be repeated by reading the next codon to call the corresponded EF Tu G TP`amino tRNA, see Figure 105 . At the end of the first cycle of elongation, the expression stream finalized the first version of the elongation shell, as shown in Figure 106 . Build (EF Ts GTP, construction set) Read (current codon, reading rules) Flag : = Compare (current codon, termination set) Call (amino tRNA, wobble complementarity matching rules ) Build (EF TU GDP, construction set) Attach (EF Ts GTP, EF TU GDP, Matching rules) Detach (EF TU GTP, amino tRNA) Move (EF TU GTP amino tRNA, current location, A/T state, tRNA movement rules) Attach (codon, anti codon, Wobble matching rules) Attach (peptide, amino, protein matching rules) Move (deacylated tRNA, P/P state, E/P state, tRNA movement rules) Move (peptide tRNA, A/A state, P/A state, tRNA movement rules) Mov e (deacylated tRNA, E/E, outside, tRNA movement rules) If Flag then Call Termination Move (amino tRNA, current location, A/A state, tRNA movement rules) Attach (amino tRNA, EF TU GTP, complementarity matching rules) Elongation

PAGE 189

! )$) ! Figure 105 . The deacylated tRNA and the peptide tRNA move completely to the E site and the P site , respectively. Figure 106 . "Elongation" Shell When the construction stream repeated the elongation cycle and reached the step where the d eacylated tRNA wanted to move to the P/E state, it reasoned that the previous deacylated tRNA in the E/E state must leave the ribosome, see Figure 107 . The expression stream modified the Build (EF Ts GTP, const ruction set) Read (current codon, reading rules) Flag : = Compare (current codon, termination set) Call (amino tRNA, wobble complementarity matching rules ) Build (EF TU GDP, construction set) Attach (EF Ts GTP, EF TU GDP, Matching rules) Detach (EF TU GTP, amino tRNA) Move (EF TU GTP amino tRNA, current location, A/T state, tRNA movement rules) Attach (codon, anti codon, Wobble matching rules) Attach (peptide, amino, protein matching rules) Move (mRNA, current location, one codon forward, mRNA move ment rules) Move (deacylated tRNA, P/P state, E/P state, tRNA movement rules) Move (peptide tRNA, A/A state, P/A state, tRNA movement rules) Call Elongation If Flag then Call Termination Trigger (EF TU GTP) Move (amino tRNA, current location, A/A st ate, tRNA movement rules) Attach (amino tRNA, EF TU GTP, complementarity matching rules) Move (deacylated tRNA, E/P state, E/E state, tRNA movement rules) Move (peptide tRNA, P/A state, P/P state, tRNA movement rules) Elongation

PAGE 190

! )$" ! elongation shell by adding the new construc tion aggregate "Move ( deacylated tRNA , E/E state, outside, tRNA movement rules ) ." The final elongation shell can be seen in Figure 108 . Figure 107 . The previous deacylated tRNA must leave the ribosome befo re the new deacylated tRNA can enter the E site. Figure 108 . The f inal "Elongation" s hell . The construction stream repeated the elongation cycle several times until the terminal codon was read indicating the end of elongation phase and the entering into the termination stage. At the end of the discovery of the hybrid model, a new elongation model was generated. The Hybrid Elongation Build (EF Ts GTP, construction set) Read (current codon, reading rules) Flag : = Compare (current codon, termination set) Call (amino tRNA, wobble complementarity matching rules ) Build (EF TU GDP, construction set) Attach (EF Ts GTP, EF TU GDP, Matching rules) Detach (EF TU GTP, amino tRNA) Move (EF TU GTP amino tRNA, current location, A/T state, tRNA movement rules) Attach (codon, anti codon, Wobble matching rules) Attach (peptide, amino, protein matching rules) Move (peptide tRNA, P/A state, P/P state, tRN A movement rules) Move (deacylated tRNA, P/P state, E/P state, tRNA movement rules) Move (peptide tRNA, A/A state, P/A state, tRNA movement rules) Move (mRNA, current location, one codon forward, mRNA movement rules) If Flag then Call Termination Tr igger (EF TU GTP) Move (amino tRNA, current location, A/A state, tRNA movement rules) Attach (amino tRNA, EF TU GTP, complementarity matching rules) Move (deacylated tRNA, E/E state, outside , tRNA movement rules) Move (deacylated tRNA, E/P state, E/E state, tRNA movement rules) Call Elongation

PAGE 191

! )$' ! thought experiments considered only the movement of tRNA inside the ribosome. All other matching rules, definit ions of tiles and aggregates were adopted from the Watson extension model and similar models. Additionally, the visual scene and the corresponding shell of the termination process are also the same as those declared in the Watson extension model. 7.9.0 Results M ultiple thought experiments involved in the application of the AD to rediscovering the process rules for building the amino tRNA were identified, the matching rules that link tRNA and mRNA and the process rules for constructing the translation model were analyzed . It was demonstrated that the AD utilized mosaic reasoning to guide itself to the solution. In order to find translation process, the AD applied visual streams that use mosaic reasoning to build thought experiments and to focus themselves toward identifying the rules that control the translation process. In Section 7.4.0 , the AD applied mosaic reasoning to find the rules of binding between mRNA and tRNA. The AD first identified a scope set that includes all possible p airs based on the complementarity rules, and then excluded all the pairs that violated any physical rules. After that, the AD declared a scope set of all possible pairings between the last tiles of both the anticodon and the codon. In this discovery, an im mediate role for the validation stream was not found. Historically, the results of these thought experiments were not confirmed until the validation stream received lab inputs that verified the correctness of Crick's conclusions, [59] . Later on, new types of tiles in the tRNA anticodon were found and are located on the wobble tile of the anticodon. The wobble complementarity matching rules for the corresponding codon tiles were also apparent for the first time.

PAGE 192

! )$= ! Also , during tRN A and mRNA matching discovery, see Section 7.4.0 , the role of a sub set of the environmental rules, the physical rules, were found. The discovery visual stream used the physical rules to eliminate some of the pairs initially ide ntified by the observation stream . Defining the scope sets was very important in reducing the time needed to check all possible pairs. Visual streams can find the solution faster when the scope sets are identified. Instead of testing every possible pair, v isual streams only tested the physical rules for the pairs in the scope set. Further more , while the thought experiments were identifying the correct translation process, mosaic reasoning was demonstrated to be applicable to more than the discoveries of the structure of real world physical objects. This type of reasoning is also involved in discoveries related to the investigation of the functionality of any mosaic. Therefore, mosaic reasoning can be applied to finding the process that is used during the bui lding of any mosaic . By applying mosaic reasoning in the discovery of the process of translation, new components of mosaic reasoning w ere revealed. The process rules that indicate which mosaic moves first, were declared. These rules were also demonstrated during the discovery of the activating process of the tRNA mosaics, see Section 7.3.0 . The movement rules c ontrolling the motion of mosaics were found. Another type of transformation rules found in the translation discovery is the reading rules , where the ribosome moves specifically to read one codon in the mRNA. The two stages of the observation stream were applied in this discovery. The preliminary stage is the stage where the observation stream looks at a stream of visual inp uts and chooses the inputs that are related to the current problem. The visual inputs to any thought experiments are symbolic data that comes to the internal stream through the impression stream.

PAGE 193

! )$> ! Algorithmic mosaics, another type of information mosaics, we re introduced. This type of mosaic is built by the expression stream by monitoring the run of any discovery stream (observation, construction, or validation) and tagging the operations applied in any thought experiment. Every tagged operation is then saved into a corresponding shell as an instruction aggregate. The expression stream then generates a full algorithmic mosaic by grouping the different shells from different thought experiments. At the end of this chapter, the expression stream grouped the fina l versions of all of the generated shells, starting from the Watson communication thought experiments and ending with the Hybrid communication thought experiments. The result of this grouping is a complete translation algorithmic mosaic, as shown in Figure 109 . Attach (Ad, P, Complementarity rule) Attach (Ad.P, P, Complementarity rule) Construct ATP ! Attach (Ad.P~P, P, Complementarity rule) ()*+,*" IFLLMN!F46:/,6,512-;1O!-./,J! ! -,,+./" IFMN!FLLMN!F46:/,6,512-;1O! -./,J! ! Construct amino ! ()*+,*" IFMN!F46:/,6,512-;1O!-./,J! ! ()*+,*" I PM " N!F46:/,6,512-;1O!-./,J " ! -,,+./" I PM " N! FM ? FLLM N!F46:/,6,512-;1O!-./,J " ! -,,+./" I PM " ? FM ? FLLM N!QN!F46:/,6,512-;1O!-./,J " ! Construct Active amino ! -,,+./" IE9
PAGE 194

! )$# ! 2*3*)+,*" I251; 34945N!F46:/,6,512-;1O!-./,0N!1QPE!1;/,0!0,1 J ! Construct tRN A body ! -,,+./" IK49ON!251; 34945N!F46:/,6,512-;1O!-./,0J! ! ! 2*3*)+,*" IK49O N! F4501-.31;45!-./,0J ! (+00" F4501-.31!E31;7,! 26;54 ! ! Construct amino tRNA ! 1*,+./" IE9
PAGE 195

! )$$ ! Figure 109 . The final algorithmic mosaic for the translation process. 40536+,753 " 8970: ! IDA ? *0 ? G*RN!34501-.31;45!0,1J ! ;*+: ! I3.--,51!34945N!-,29;5B!-./,0J ! A/2B! U V! (5<=+)* ! I3.--,51!3 4945N!1,-6;521;45!0,1J ! (+00 ! I26;54 ? 1QPEN!W4KK/,!346:/,6,512-;1O!6213+;5B! -./,0 J ! 8970: ! IDA ? *H ? GXRN!34501-.31;45!0,1J ! -,,+./" IDA ? *0 ? G*RN!DA ? *H ? GXRN!@213+;5B!-./,0J ! 1*,+./" IDA ? *H ? G*RN!26;54 ? 1QPEJ ! >5?* ! IDA ? *H ? G*R ? 26;54! ? 1QPEN!3.--,51!/4321;45N!EY*!0121,N! 1QPE!647,6,51!-./,0J ! -,,+./ ! I34945N!251; 34945N!Z4KK/,!6213+;5B!-./,0J ! -,,+./" I:,:1;9,N!26;54N!:-41,;5!6213+;5B!-./,0J ! >5?* ! I:,:1;9, ? 1QPEN!RYE!0121,N!RYR!0121,N!1QPE!647,6,51!-./,0J ! >5?* ! I9,23O/21,9 ? 1QPEN!RYR!0121,N!DYR!0121,N!1QPE!647,6,51!-./,0J ! >5?* ! I:,:1;9, ? 1QPEN!EYE!0121,N!RYE!0121,N!1QPE!647,6,51!-./,0J ! >5?* ! I6QPEN!3.--,51!/4321;45N!45,!34945!84-W2-9N!6QPE!647,6,51!-./,0J ! @A ! A/2B ! ,/*3 ! (+00 ! *,-6;521;45 ! B)766*)" IDA ? *H ? G*RJ ! >5?* ! I26;54! ? 1QPEN!3.--,51!/4321;45N!EYE!0121,N!1QPE!647,6,51!-./,0J ! -,, +./" I26;54 ? 1QPEN!DA ? *H ? G*RN!346:/,6,512-;1O!6213+;5B!-./,0J ! >5?* ! I9,23O/21,9 ? 1QPEN!DYD!0121,N!4.10;9, N!1QPE!647,6,51!-./,0J ! >5?* ! I9,23O/21,9 ? 1QPEN!DYR!0121,N!DYD!0121,N!1QPE!647,6,51!-./,0J ! (+00 ! D/45B21;45 ! Termination Detach (ribosome, deacylated tRNA ) Detach (30 s, 50 s ) Detach ( 50 s, peptide) Detach ( 30 s, mRNA) Detach (ribosome, peptide tRNA )

PAGE 196

! )$% ! CHAPTER VIII CONTRIBUTIONS AND FUTURE WORK 8.1.0 Overview After applying the AD to the discoveries of algorithms of admissible trajectories, genetic code, and translation, multiple new components that guided these discoveries were identified. This Chapter includes a detailed analysis of our contributions and the future work. 8.2.0 Contributions 8.2.1 New Matching Rules Matching rul es are part of the construction rules defined, by the visual streams, during the application of mosaic reasoning to solve a problem. These rules control the linkage of different tiles and aggregates building the final mosaic. They also determine the moveme nt of the mosaics within thought experiments. Matching rules include transformation, complementarity, interchangeability, and environmental rules. During this research, a new set of matching rules was identified. They describe special types of transformati ons for mosaics and aggregates. Transformation rules are the part of the matching rules that control the transformation of one mosaic into another. While applying the AD to the case studies in Chapters V, VI , and VII , additional types of transformation ru les were defined. These rules are the movement, reading, alignment, and process rules. The movement rules are a subset of the transformation rules. They control the motion of any mosaic within the visual stream. Due to the nature of the case studies in thi s research, the role of the movement rules was emphasized in almost every problem. In the admissible trajectory problem, the movement rules controlled the motion of the ghost , while scanning the final admissible tree to eliminate any redundancies, Section 5.5.2 . The movement rules were also

PAGE 197

! )$& ! utilized during the discovery of the translation process. These rules controlled the motion of the tRNA, mRNA, and ribosome mosaics , while constructing the protein mosaic. Also , these rules g uided the movement of the mRNA tiles when building the information mRNA mosaic. The movement rules are necessary in any discovery that requires movement of the tiles, aggregates, and the ghost. For example, in LG, the movement rules represent the reachabil ities of any piece on the abstract game board. The reading rules manage the reading of the information stored in information mosaics. These rules are utilized by the AD in any problem that requires information transfer between various entities. In this ty pe of problems, visual streams need to define how the information can be read in order to reach a correct solution. In this research, the reading rules were found in the g enetic code discovery, Chapter VI . During translation, the genetic information stored in the mRNA mosaic is read. Several runs of the AD were performed in order to define the correct reading rules for the mRNA mosaic. Once the correct reading rules were identified, the AD declared the discovery of the genetic code. The reading rules also w ere defined during the admissible trajectory case study in the Find Mid thought experiment, specifically, when the sum of forward and backward distances needs to be read for comparison with the length of the admissible trajectory. The alignment rules are needed for aligning two mosaics before attaching them together. While working on the admissible tree, Chapter V , the alignment rules were needed to put the From Mid tree in the correct position before attaching it to the To Mid tree. The alignment rules we re also identified when building the ribosome in the tra nslation case study in Chapter VII . Looking back to other discoveries visited by Stilman and Aldossary in [10] and [4] , the

PAGE 198

! )%( ! alignment rules we re also needed to align the tiles of the DNA and the protein before constructing the aggregates. The process rules are applied to order the different actions applied by internal visual streams on different mosaics within any thought experiment. For example , the process rules are used to construct the correct algorithm of translation. The process rules control the order in which mosaic is built, moved, or destroyed. In the initiation stage, the process rules force the construction stream to first build the r ibosome, construct the mRNA, attach them together, and then the ribosome can scan the mRNA to fin d the initiation code aggregate. Addressed in Chapter VII , t he process rules were emphasized in all thought experiments created to solve the protein translati on problem. These rules were first found in the discovery of the activation of the tRNA mosaics , Section 7.3.0 . They also were defined in the different translation models Sections 7.5.0 , 7.6.0 , 7.7.0 , and 7.8.0 . Next , the process rules controlled the correct mapping between the amino and code sets in the genetic code problem, Section 6.6.4 . Defining the correct process rules leads to creating the correct algorithm to solve an investigated problem. The physical rules , a subset of environmental rules, were identified as part of the matching rules. The environmental rules include the rules of the surrounding environment. For example, when finding a solution for a problem associated with falling objects, one must consider gravity, which is part of the environmental rules. The physical rules consider the known laws of physics for a problem. Th ese rules were first found while working on the discovery leading to the finding of the correct complementarity rules required to build the codon anticodon linkage, Section 7.4.0 . When looking for the correct co mplementarity rules between the mRNA and tRNA tiles,

PAGE 199

! )%) ! the visual streams must consider the physical distances between the base tiles as seen in the x rays. 8.2.2 Investigation of New Streams While working on this research, a new type of visual streams, the commun ication streams, were introduced, Section 2.4.0 . They include the expression and impression streams. The expression stream is a subset of the communication streams that transforms the actions of the internal streams into a symb olic language to be understood by the outer world. The impression stream transforms symbolic language into the internal visual streams. The output of the expression stream is a string of symbols and pictures that maps the results of the internal streams to the outer world. The expression stream usually employs secondary (conventional) languages. These could be written (visu al), vocal (sound), or computer generated strings. The communication thought experiments are the thought experiments created by the expr ession stream while translating the work of the PL , or by the impression stream when converting symbolic languages into the PL. The internal visual streams generate the internal thought experiments while working on a problem. In Chapters V, VI , and VII , mu ltiple communication thought experiments run by the expression stream were defined to translate the work of discovery streams. Usually, the results of the construction stream were translated. The outputs of this expression stream were an algorithmic mosaic or a symbolic mosaic, Section 8.2.3 with explaining the solution generated by the discovery streams. 8.2.3 New Types of Mosaics Before the start of this research, the only type of mosaic considered was a structural mosaic. This mea nt that the mosaic was involved in building physical or chemical objects. After working on several case studies, the variety of mosaics was expanded. Mosaics can be of different types

PAGE 200

! )%" ! based on their purpose. A structural mosaic is the one that represents t he physical structure of the object under investigation. An example of a structural mosaic is the DNA mosaic introduced by Stilman in [10] . In this research, three new types of data storage mosaics were defined: the informatio n, symbolic, and algorithmic mosaics. The information mosaic for the internal streams is responsible for saving data to be used later in any process. Internal streams build these mosaics if the problem requires the reading of data from a structural mosaic. Information mosaics are built out of "letter" tiles that are aligned next to each other to form a long "sentence." In particular, they can be one dimensional or two dimensional tables. These mosaics are created via a sequence of transformation phases. In order to read them, the reading rules must be employed. Examples of the information mosaics include the admissible tr ee mosaic, defined in Chapter V , and the information DNA and mRNA mosaics defined , in Chapters VI , and VII . The algorithmic mosaic is the d ata storage mosaic utilized by the communication streams. It contains the algorithm required to solve the investigated problem as constructed by the internal streams. The expression stream builds the mosaic to reflect the actions of an internal stream. The algorithmic mosaic is constructed from several instruction aggregates based on the actions of the objects within the internal thought experiments. Each instruction aggregate consists of two aggregates, operation and parameter aggregates. In order to form alize the structure of the information mosaic in any discovery, expression stream can choose one of the instruction aggregates that are defined in Table 2 . Creating the formal language of the algorithmic mosaic is another contribu tion of this research. The general form of the instruction aggregates is expected to be applicable to many discoveries. The

PAGE 201

! )%' ! generality of the algorithmic mosaic language will make it possible to add more instruction aggregates, if needed, while working on new case studies. Algorithmic mosaics were identified in all case studies in this research. In the admissible trajectory case study, they represent the algorithm used to build the final admissible tree. In the genetic code problem, they represent the algo rithm used to build the information mRNA mosaic. In activating the amino tRNA, they represent the algorithm used in constructing the active amino tRNA. Algorithmic mosaics were also built to represent the algorithms of initiation, elongation, and terminati on while working on the translation problem. The last type of mosaic defined in this research is the symbolic mosaic . The symbolic mosaic is also defined by the expression stream to save information produced by the internal stream. This information could i nclude the construction rules that the observation stream declared during the analysis stage. Several symbolic mosaics were introduced, such as the genetic mosaic that represents the mapping applied by the construction stream between the genetic code and a mino sets , Table 6 . Another symbolic mosaic is the mosaic that represents the translation from the DNA tiles to the mRNA tiles and vice versa, Table 5 , as well as the mosaic that includes the translation between the mRNA base and letter tiles, Table 3 . The mosaic that includes all possible code aggregates from the diamond code experiment is also an example of a symbolic mosaic, Table 4 . While working on the protein translation problem, we realized that the expression stream generated a symbolic mosaic, which included a set of complementarity matching rules, Table 11 . We also discovered that this mosaic was utilized several times by other visual streams during the investigation of the elongation process. Every time a visual stream had to attach a tRNA mosaic

PAGE 202

! )%= ! to a mRNA mosaic, the complementarity matching mosaic must be utilized to ensure that the wobble complementarity matching rul es are satisfied. 8.2.4 New Transformation Phases of Protein Synthesis During this research, two transformation phases were identified that are associated with the discovery of the protein synthesis process. They are transcription and translation, shown in Figure 11 0 . In the transcription phase, the genetic information, saved in the DNA information mosaic, is read according to the reading rules. Based on that reading, the DNA information mosaic should be transformed into multiple RNA in formation mosaics, each of which carries the genetic information of a single gene. Figure 11 0 . The s tages of p rotein s ynthesis . The new mRNA mosaic will enter the translation phase in which the genetic code it contains is read to create the protein mosaic. During translation, multiple mosaics are interacting with each other to produce the final protein mosaics. They are the mRNA, the deacylated tRNA, the amino tRNA, the peptide tRNA, and the ribosome mosaics.

PAGE 203

! )%> ! 8.2.5 New Scope Sets The sco pe sets are the subsets of construction sets used to focus the streams. These sets of tiles and aggregates limit the material utilized for mosaic construction. A number of scope sets were identified during our case studies. While working on the admissible trajectory case study, the observation stream defined the DOCK scope set to focus the construction stream , while building the admissible tree. Multiple scope sets were built while analyzing the genetic code problem. Identifying the correct amino and code sets was essential in discovering the genetic code. During the process of finding the correct matching rules between the mRNA and tRNA mosaics, the observation stream identified a scope set that included only the pairs that have the potential to be connec ted. Next , the observation stream triggered the construction stream to attach these pairs. Identifying the scope set, that included these pairs, focused the construction stream while defining the correct matching rules between the mRNA and tRNA mosaics. 8.2.6 Id entifying a New Stage of the Observation Stream The observation stream definition from previous work of Stilman et al., [27] [6] and [28] [30] , was updated to include two stages, the preliminary and the analysis stages. During the preliminary stage , the observation stream saves all visual inputs related to the problem. In contrast, the analysis stage requires the observation stream to take the saved inputs, analyze them, and create reasonable conclusions based on that analysis. Every time the AD starts its run, the observation stream must be initiated in its preliminary stage. The main goal of this stage is to choose the required inputs from an impression stream, which may includes millions of items. Selecting the correct inputs at the preliminary stage will increase the possibility of finding the correct solution to the problem.

PAGE 204

! )%# ! 8.2.7 Identifying New Thought Experiments During the study into revealin g the core of the PL, the algorithm of the AD was studied by monitoring its execution. Also, the "computer" running this algorithm was examined in a trial to find its instruction set, i.e., the human brain's mechanisms used to execute it. During this inve stigation, this algorithm was assumed to be based on visual streams (mental movies), which visually demonstrate a sequence of events interacting with each other. It is argued that those discovery streams could be recreated by analyzing publications of a sc ientist, via identifying his/her thought experiments involved in the discovery. In performing this research , it was assumed that all discoveries are made by running the same algorithm, mainly the AD. In order to understand the behavior of the AD (that led to each discovery), the sequence of thought experiments, following the available publications , was reconstructed. The reasoning that helped in recreating and morphing those streams was carefully followed. It revealed additional components of the AD activel y involved in the specific discovery. We revealed and executed numerous thought experiments that were generated by the visual streams for solving the admissible trajectory, genetic code, and protein translation problems. We confirmed that, for the above l ist of problems, the AD consistently applied its major tool, mosaic reasoning. Interestingly, this investigation revealed that mosaic reasoning in the admissible trajectory, genetic code, and translation problems generated many different types of mosaic s . Further more , the flow of data between different streams from different thought experiments is essential to discarding incorrect conclusions and confirming correct ones. A correct conclusion is then used by the AD to focus the different streams towards a di scovery. Also, it was found that the AD must look at the impression stream of inputs and accept only the inputs that are required. Focusing the inputs is a very important step that makes the discovery

PAGE 205

! )%$ ! possible, since analyzing to all inputs is impossible. In the future, the application of the AD will replay past discoveries until the complete correct structure of the AD is revealed. In Chapter V , we described an application of the AD to discovering the algorithm for generating admissible trajectories. Ever y thought experiment done by Stilman in [28] was revisited to reveal the behavior of visual streams (which run such thought experiments). From this analysis, it was concluded that mosaic reasoning was a major contributor in con structing the algorithm of admissible trajectories. In particular, the AD employed mosaic reasoning to construct the algorithm for building the mosaic tree of bundles of admissible trajectories. It was found that the AD had to define tiles, aggregates, and matching rules, as well as define the scope sets, to focus the construction stream. In addition, the approach of "erasing the particulars" was applied in the discovery of admissible trajectories. After revealing all of the thought experiments that led to the discovery of the admissible trajectories algorithm, an OpenGL animated program was developed, to simulate application of the AD to this discovery. In Chapter VI , the genetic code discovery was analyzed by visiting several thought experiments performed by several scientists. The visual streams were recreated, including the visual objects and sets having a major role in generating the algorithm for deciphering the genetic code. The conclusion is that the AD used mosaic reasoning to make this discovery. Mo saic reasoning was applied for the first time to analyze an existing mosaic rather than constructing a new one. This result of a successful application of mosaic reasoning, to study the structure of an existing mosaic, revealed its power and universal natu re of this p articular c omponent of the AD. Chapter VII includes many thought experiments run by the internal and communication streams to find the correct process of translation. The AD mainly applied mosaic reasoning to

PAGE 206

! )%% ! build different mosaics. Next , the AD moved those mosaics to study their interaction and investigate the process of building the protein mosaic. Figure 111 shows the role of different AD components utilized in finding the solution for every problem in this research . Figure 111 . The components of the AD utilized in each case study . 8.3.0 Towards Automation of the AD At the beginning of every trial, the AD reinitiates itself. It might happen several times until the AD makes a discovery. Studying t he execution of the AD requires focusing on one trial of the AD from initiation to termination. The AD halts when it reaches results, correct or not, and from the results of the current AD trial, the next trial will focus the streams in the most promising direction to find the answer. Our experience with execution of the AD shows that the AD runs in a specific number of steps, limited by a constant. This upper bound does not depend of the size of the input. However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

PAGE 207

! )%& ! before running these steps, the AD must a nalyze the input to find a pattern or behavior. This analysis is proportional to n , the size of the input. Overall, we evaluate the AD run time as O( n ) . The experience of human scientists shows that they do not look at the entire set of inputs. In making discoveries, scienti sts analyze the inputs, often sub consciously, looking for specific inputs based on their analysis. Sometimes, one specific input might affect the entire discovery. Indeed, the AD has categorized the inputs to be either restricted inputs prohibiting the streams from following certain directions, or permitted inputs that the streams can examine further. The permitted inputs are the ones that will be saved in the scope sets for the stream to focus the investigation accordingly. For example, while looking for the correct reading of the genetic code, the observation stream added to the restricted inputs the information that the code cannot be read base by base, or two bases at a time. Th e se restrictions forced the AD to discard or delete all s tream inputs related to such information. Further more , the stream identified that th e code could be read in triples, which guided the stream to "encapsulate" its inputs by looking only to the inputs that might explain the triple reading. The current level of knowledge might cause the stream to incorrectly include an input as a permitted one. This incorrect inclusion would affect the final conclusion of the current thought experiment, and the AD would terminate with incorrect outputs, which might be determi ned later , after any subsequent validation. In such case, the discovery would not happen until the AD categorizes the inputs correctly. If, and only if, the AD chooses all its inputs correctly , the final solution could be reached. In our pursuit of the go al of automating the AD, we should think about how the scientists thought while they were solving their problems. First, they eliminated all the impossible "solutions" based on any current information or knowledge. Then, they identified possible rules

PAGE 208

! )&( ! aft er reasoning about current inputs and testing the results. For example, in protein synthesis, the existence in any molecule, of a genetic code carrier different from the DNA was eliminated. Also, the reading of the genetic code from the sugar or phosphate tiles was excluded. Next, the path toward the correct solution began by permitting the genetic code reading from the base tile sequence of the DNA. Further, on the first trial of the AD, the permitted input was to read the code directly from the DNA. Some time later, it was determined that assigning the code reading directly from the DNA (without some intermediate substance) was incorrect. This invalidated the definition of the whole process of protein synthesis as it was constructed in the previous thought experiments. During the next trials, the AD determined a new type of permitted input. Those inputs showed that the genetic code should be transcribed first into the mRNA, and then the mRNA should be translated into a protein. This correct assignment led the AD to follow the correct path toward the solution. Specifically, the requirement that the DNA should be transcribed first, caused the need to find yet another requirement, which was the rules that guide the transcription. Therefore, the AD started to c ollect inputs that might lead to these rules. It then analyzed these inputs and reasoned using its visual streams and applying mosaic reasoning to mimic what it "saw" in the lab or x rays. The visual streams should first identify the tiles of the input mo saic(s) and then identify the rules that must be followed to create an aggregate out of those tiles. The tiles and aggregates, whether they were correct or not, must be part of the resulting algorithm , not part of the AD. For example, the tiles and the agg regates, along with the rules to construct them, are part of the translation algorithm , not part of the AD. However, how to identify them is part of the AD. Indeed, during multiple applications of the AD, it was found that the AD has common behavior,

PAGE 209

! )&) ! such as identifying the matching rules. That identification does not mean those matching rules are part of the AD, they are part of the current problem. Exactly what is part of the AD is the way that the AD identified them and determined whether the matching r ules are part of the new algorithm or not. 8.4.0 Future Work The scope of this research is expanding rapidly. New results directing research in different directions are frequently being published. T he future work is to continue working on the first stage of pro tein synthesis, Which is the transcription. Transcription will be examined via case studies and application of the AD to rediscover the algorithm that guided its discovery. Transcription is the construction of the mRNA mosaic by transforming the DNA mosai c. The constructed mRNA is used for the protein synthesis employing translation. By applying the AD to transcription, it is expected that the identification of thought experiments , that allow the refinement of all phases of the transformation involved in protein synthesis, will include all of the details of transcription and translation. By a further refinement of the AD, the details of this transformation, additional procedures, themes of visual streams and various data flows between visual streams will b e identified. After refining the details, the result of the analysis will lead to a manual application of the AD to discover the process of protein synthesis. Once the manual application is produced, the second stage, the implementation stage, i.e., the co mputer modeling stage, of the research will start. The translation portion of the implementation stage is complete. However, to test the correctness of the components of the AD, that guided the whole discovery of protein synthesis, the next step is to incl ude all of the visual streams that will be identified during the analysis of

PAGE 210

! )&" ! transcription. For this purpose, the thought experiments that led to the discovery of the protein synthesis will be implemented in software as GeneDiscover and applied to the disc overy of the genetic code. For this research, the major part of this discovery is the translation algorithm. This way, we will identify and implement all of the components of the AD initiated by visual streams that participate in the thought experiments ge nerating algorithms for transforming the DNA mosaic into the protein mosaic. The input of GeneDiscover will be the information mosaic of the DNA, the complete set of amino aggregate and the protein structure. GeneDiscover is expected to initiate all of the required visual streams. This way, the simulated AD will discover and construct the decoding algorithms, which will be utilized for generating the algorithm for transcription from the DNA mosaic to the mRNA mosaic. In order to approach our long term goal of making discoveries on demand, we will continue the application of the AD on multiple case studies from other fields of science such as physics and mathematics. This way, we will be able to generalize our findings for all of the discoveries. After gener alizing the AD components, the structure of the AD will be clear to the extent that we will be able to implement the full AD. The implementation of the AD means to develop a program that will receive stream of visual inputs and generate an algorithm that w ill solve any investigated problem. In t his way, we are going to build an artificial scientist who will be able to reason effortlessly at the level of the greatest human scientists.

PAGE 211

! )&' ! BIBLIOGRAPHY [1] A. Miller, Insights of Genius: Imagery and Creativity in Sci ence and Art , Copernicus, NY: Springer Verlag, 1996. [2] A. Rich, "How Transfer RNA May Move Inside the Ribosome," Ribosomes (eds. A. Nomura , A. Tissieres, , P. Lengyel) , Cold Spring Harbor Laborator , New York, pp.871 884, 1974. [3] B. Hayes, "The Invention of the Genetic Code," American Scientist , Vol. 86, No. 1, pp. 8 14, January February 1998, (Doi) 10.1511/1998.1.8 . [4] B. Stilman and M. Aldossary, "The Algorithm of Discovery: Programming the Mosaic of the Shorte st Trajectories," Proc. of the IEEE Int. Conf. on Intelligent Systems , p. 6, Sofia, Bulgaria, 2016. [5] B. Stilman and N. Alharbi, "Towards the Algorithm of Discovery, Rediscovering The Algorithm of Admissible Trajectories," The International Science and Pract ice Forum, "Science and Business" , pp. 9 15, Ukraine, June 2015. [6] B. Stilman and N. Alharbi, "Towards the Algorithm of Discovery: Identifying the rules that link tRNA and mRNA", PROCS11179, J., Procedia Computer Science , Aug 2017, pp. 479 484, (Doi) 10.1016 /j.procs.2017.06.050. [7] B. Stilman and N. Alharbi, "Towards the Algorithm of Discovery: Mosaic Reasoning for the Discovery of the Genetic Code", Proc. of the IEEE Int. Conf. on Intelligent Systems , p. 6, Sofia, Bulgaria, 2016. [8] B. Stilman, "Discovering the Di scovery of Linguistic Geometry," Int. J. of Machine Learning and Cybernetics, Springer , 2013, Vol. 4, No. 6, pp. 575 594, (Doi) 10.1007/s13042 012 0114 8, 20 p. [9] B. Stilman, "Discovering the Discovery of the Hierarchy of Formal Languages," Int. J. of Machin e Learning and Cybernetics , Springer , p. 25, 2012 , (Doi) 10.1007/s13042 012 0146 0, 2012. [10] B. Stilman, "Mosaic Reasoning for Discoveries," J. of Artificial Intelligence and Soft Computing Research , Vol. 3, No. 3, pp. 26, 2013. [11] B. Stilman, "Thought Experimen ts in Linguistic Geometry," presented at Proc. of the 3d Int. Conf. on Advanced Cognitive Technologies and Applications Ð COGNITIVE , Rome, Italy, Sep. 25 30, 2011. [12] B. Stilman , "Visual Reasoning for Discoveries," Int. J. of Machine Learning and Cybernetics, Springer , Vol. 6, No. 2, pp. 227 251, 2013, (Doi) 10.1007/s13042 013 0189 x.

PAGE 212

! )&= ! [13] B. Stilman, "A Formal Language for Hierarchical Systems Control," Int. J. Languages of Design , Vol. 1, No. 4, pp. 333 356, 1993. [14] B. Stilman, "A Linguistic Approach to Geometric Reasoning," Int. J. of Computers & Mathematics with Appl. , Vol. 26, No. 7, pp. 29 58, 1993. [15] B. Stilman, "A Linguistic Geometry for 3D Strategic Planning," Proc. of the 1995 Goddard Conf. on Space Appl. of Artificial Intel. and Emerging Information Technolo gies , pp. 279 295, NASA Goddard Space Flight Center, Greenbelt, MD, 1995. [16] B. Stilman, "Discoveries on Demand", Int. J. of Design & Nature and Ecodynamics , Vol. 11, No. 4, pp. 495 507, 2016. [17] B. Stilman , "Discovering the Discovery of the No Search Approach," Int. J. of Mach. Learning and Cybernetics , p.27, 2012, (Doi) 10.1007/s13042 012 0127 3. [18] B. Stilman, " Linguistic Geometry and Evolution of Intelligence," ISAST Trans. on Computers and Intelligent Systems , Vol. 3, No. 2, pp. 23 37, 2011 . [19] B. Stilman, "Ling uistic Geometry for Control Systems Design," Int. J. of Computers and Their Applications , Vol. 1, No. 2, pp. 89 110, 1994. [20] B. Stilman, "Managing Search Complexity in Linguistic Geometry," IEEE Trans. on Syst., Man, and Cybernetics , Vol. 27, No. 6, pp. 978 998, 1997. [21] B. Stilman, "Network Languages for Complex Systems," Int. J. of Computers & Mathematics with Appl. , Vol. 26, No. 8, pp. 51 80, 1993. [22] B. Stilman, "Network Languages for Concurrent Multi agent Systems," Intl. J. of Computers & Mathematics with App l. , Vol. 34, No. 1, pp. 103 136, 1993. [23] B. Stilman , "Proximity Reasoning for Discoveries," Int. J. of Machine Learning and Cybernetics , Springer , p. 31, 2014, (Doi) 10.1007/s13042 014 0249 x, 2014. [24] B. Stilman , "The Algorithm of Discovery: Making Discoveries on Demand," In S. Kunifuji et al. (Ed.), Advances in Intelligent Systems and Computing , Springer Int, pp. 1 16, 2016, (Doi)10.1007/978 3 319 27478 2_1. [25] B. Stilman, "The Primary Language of the Human Brain," PROCS11179, J., Procedia Computer Science , Aug 2 017, pp. 448 462, (Doi) 10.1016/j.procs.2017.06.047 . [26] B. Stilman, "Translations of Network Languages," Int. J. of Computers & Mathematics with Appl. , Vol. 27, No. 2, pp. 65 98, 1994. [27] B. Stilman, From Primary to Conventional Science, unpublished. [28] B. Stilman, Linguistic Geometry: From Search to Construction , Boston: Kluwer Academic Publishers , 2000.

PAGE 213

! )&> ! [29] B. Stilman, V. Yakhnis, O. Umanskiy, "Revisiting History with Linguistic Geometry," ISAST Trans. on Computers and Intelligent Systems , Vol. 2, No. 2, pp. 22 38, O ct. 2010. [30] B. Stilman, V. Yakhnis, O. Umanskiy, "The Primary Language of Ancient Battles," Int. J. of Machine Learning and Cybernetics , Vol. 2, No. 3, pp. 157 176, 2011. [31] C. Woese, "Molecular Mechanics of Translation: a Reciprocating Ratchet Mechanism," Natu re , Vol. 226, pp.817 820, 1970, (Doi) 10.1038/226817a0 . [32] D. Moazed, and H. F. Noller, "Intermediate states in the movement of transfer RNA," Nature , Vol. 342, pp. 142 148, 1989, (Doi) 10.1038/342142a0 . [33] F. Crick, "On protein synthesis," The Symposia of the Society for Experimental Biology , Vol. 12, No. 113, pp. 138 163, 1958. [34] F. Crick, " From DNA to protein on degenerate templates and the adapter hypothesis: a note for the RNA Tie Club," Wellcome Library for the History and Understanding of Medicine, Francis Harry Compton Crick Papers , 1955, Internet : http://profiles.nlm.nih.gov/ps/access/SCBBGF.pdf [35] F. Crick, " On the Genetic Code," Nobel Lectures, Physiology or Medicine 1942 1962, Elsevier Publis hing Company, Amsterdam, 1964 , Internet: http ://www.nobelprize.org/nobel_prizes/medicine/laureates/1962/crick lecture.html. [36] F. Crick, L. Barnett, S. Brenner, and R. Watts Tobin, "General nature of the genetic code for proteins," Nature , Vol. 192, No. 4809 , pp. 1227 1232, 1961. [37] F. Crick, What mad pursuit: a personal view of scientific discovery , New York: Basic Books, 1988. [38] F. H. Crick, "Codon Anticodon Pairing: the Wobble Hypothesis," J. of Molecular Biology , Vol. 19, No. 2, pp. 548 555, 1966. [39] F. Lipmann, "Polypeptide Chain Elongation in Protein Biosynthesis," Science , Vol. 164, No. 3883, pp. 1024 1031, 1969, (Doi) 10.1126/science.164.3883.1024. [40] F. O. Wettstein, and H. Noll, "Binding of transfer ribonucleic acid to ribosomes engaged in protein synthesis: n umber and properties of ribosomal binding sites," J Mol Biol , Vol. 11, pp. 35 53, 1965. [41] F. Sanger, " The chemistry of insulin, " Nobel Lectures, Chemistry 1942 1962 , Elsevier Publishing Company, Amsterdam, 1964 , Internet: http://www.nobelprize.org/nobel_prizes/chemistry/laureates/1958/sanger lecture.pdf . [42] G. Gamow, "Possible Relation between Deoxyribonucleic Acid and Protein Structure," Nature , Vol. 173, No.4398, pp. 318 318, 1954.

PAGE 214

! )&# ! [43] H. Noll, "Chain initiation and control of protein synthesis," Science , Vol. 151, pp. 1241 1245, 1966. [44] H.S. Stoker, General, Organic, and Biological Chemistry , Boston: Cengage Learning, 2015. [45] J. Berg, J. Tymoczko, and L. Stryer, Biochem istry , 5 th ed., W.H. Freeman, 2002 . [46] J. D. Watson, "The synthesis of proteins upon Ribosomes," Bull. Soc. Chim. Biol. , Vol. 46, pp. 1399 1425, 1964. [47] J. Neumann, The Computer and the Brain , 2nd ed., NY: Yale U. Press , 1958. [48] K. H. Nierhaus, "Structure, assembly, and function of ribosomes," Curr Top Microbiol Immunol , Vol. 97, pp. 81 155, 1982. [49] K. H. Nierhaus, and H. Rheinberger, "An alternative model for the elongation cycle of protein biosynthesis," Trends in Bioche mical Sciences , Vol. 9, No. 10, pp.428 432, 1984. [50] K. Moitra, A Journey Through Genetics [Part 1] , San Rafael: Morgan & Claypool, 2014. [51] L. Pray, "Discovery of DNA structure and function: Watson and Crick," Nature Education , Vol. 1, No.1, p. 100, 2008. [52] M. B uzzoni, "Thought Experiments," New Catholic Encyclopedia Supplement 2012 2013: Ethics and Philosophy , Vol. 1, pp. 1534 1536, 2013. [53] M. Hogland, Toward the Habit of Truth: A Life in Science , New York: W. W. Norton, 1990. [54] M. Lorch. " The most beautiful wrong ideas in science ." Internet: http://www.chemistry blog.com/2012/08/16/the most beautiful wrong ideas in science , Aug.16, 2012, Jun. 5, 2016 . [55] M. Nirenberg, "H istorical review: Deciphering the genetic code a personal account," TRENDS in Biochemical Sciences , Vol. 29, No. 1, January 2004. [56] M. Nirenberg, and J. Matthaei," The Dependence Of Cell Free Protein Synthesis In E. Coli Upon Naturally Occurring Or Synthet ic Polyribonucleotides," Proceedings of the National Academy of Sciences of the United States of America , Vol. 47, No. 10, pp. 1588 1602, 1961. [57] M.Takanami, and G. Zubay, "An estimate of the size of the ribosomal site for messenger RNA binding," Proceedings of the National Academy of Sciences , Vol. 51, No. 5, pp. 834 839, 1964. [58] N. Kresge, R. Simoni, and R. Hill, "The Discovery of tRNA by Paul C. Zamecnik," The Journal of Biological Chemistry , Vol. 280, No. 40, pp. e37 e39, 2005.

PAGE 215

! )&$ ! [59] P. B. Moore, T.A. Steitz, "T he roles of RNA in the synthesis of protein," Cold Spring Harbor Lab. Press , Vol. 3, No. 11:a003780, 2010, (Doi) 10.1101/cshperspect.a003780. [60] S. Deheaene, "A Few Steps Toward a Science of Mental Life," Mind , Brain and Education , Vol. 1, No. 1, pp. 28 47, 2007. [61] S. Kosslyn, W. Thompson, I. Kim, and N. Alpert, "Representations of mental images in primary visual cortex," Nature , Vol. 378, pp. 496 Ð 498, 1995. [62] S. Pestka, "Studies on the formation of transfer ribonucleic acid ribosome complexes. XIX. Effect of a ntibiotics on peptidyl puromycin synthesis on polyribosomes from escherichia coli," The Journal of Biological Chemistry , Vol. 247, pp. 4669 4678, 1972. [63] S. V. Kirillov, and Y. Semenkov, "Extension of Watson's model for the elongation cycle of protein biosyn thesis," Journal of Biomolecular Structure and Dynamics , Vol. 4, No. 2, 1986, (Doi) 10.1080/07391102.1986.10506345 . [64] V. Nanjundiah,"George Gamow and the Genetic Code," Resonance, Journal of S cience Education , Vol. 9, No. 7, pp. 44 50, July 2004.