[Bio] / FigTutorial / tut_abs.html Repository:
ViewVC logotype

Diff of /FigTutorial/tut_abs.html

Parent Directory Parent Directory | Revision Log Revision Log | View Patch Patch

revision 1.3, Sat Oct 13 15:06:20 2007 UTC revision 1.4, Thu Nov 15 21:06:50 2007 UTC
# Line 4  Line 4 
4  <h1>An Abstract View</h1>  <h1>An Abstract View</h1>
5  <h2>by Ross Overbeek</h2>  <h2>by Ross Overbeek</h2>
6  </div>  </div>
7    <h2>Introduction</h2>
8    This strange document began as a tutorial for computer scientists and mathematicians.  It was supposed
9    to somehow introduce them to the computational issues in genome analysis.
10    It was requested by an instructor in a computer class.  Overbeek in attempting to respond to this request
11    formulated an abstraction that he began to believe had significance beyond the tutorial.
12    <p>
13    This document is a set of working notes relating to the abstract.  It is not organized properly as
14    an abstraction, a tutorial, or an essay on the role of bioinformatics in support of biological research.  It is,
15    however, organized properly as a working document that relates to all of these goals.
16    <p>
17    It begins with a development of the abstraction.  This will be suitable for mathematicians or computer scientists.
18    The abstraction is developed in four steps: the basic abstraction, the enhanced abstraction needed to support
19    basic bioinformatics support for biologists, and finally the third step which includes suport for the notion
20    of regulation.  The intent throughout this discussion will be to seek a minimal set of concepts needed to
21    effectively capture the essence of the required data.  Unlike almost all efforts to lay a foundation
22    for tutorials, software or research in biology, this effort focuses on leaving out as much as possible.
23    While we do believe that there is an almost unlimited complexity that can be introduced, and almost all of
24    it is needed for some specific goals, the vast majority of tools and discussions require (we believe) relatively few
25    concepts.  As they say, "the proof is in the pudding."
26    
27    <p>
28    The second section will feature a bit more tutorial comments.  It may well repeat much of what is in Part 1.
29    This part is offered as a way of easing a computer scientist of mathematician into the issues that need to be
30    considered, if they wish to try to do useful research relating to the genomics revolution.  Eventually, this part
31    will be dramatically expanded by giving condensed summaries of the machines of the cell broken into two broad
32    sets: the metabolic network and the cellular machinery not directly included in the metabolic network.  Loosely,
33    this separates what would be learned in a microbial biochemistry class (when they exist) from what would
34    be learned in a course on molecular biology.
35    <p>
36    The third part is an essay is an attempt to characterize our view on
37    <ul>
38    <li> what the main goals should be in current efforts to advance biological knowledge via genome research,
39    <li> what role bioinformatics researchers have played in the past, and
40    <li> what role they could productively play during the coming few years.
41    </ul>
42    As such, it is undoubtedly an arrogant formulation by a group of individuals with minimal background in
43    biology.
44    <p>
45    The fourth section will focus on the imlications of the abstractions in software development.
46    This is a bit of a radical proposal that makes sense to us (and is in an area that we can
47    legitimately claim expertise).
48    
49  <h2>What Is a Cell?</h2>  <h1>Part 1: The Abstractions</h1>
50    <h2>The cell: a Minimal Perspective</h2>
51    
52  A <b>cell</b> is a bag (i.e., a volume enclosed by a membrane) that contains three types of things: compounds, cellular machines, and a genome.  A <b>cell</b> is a bag (i.e., a volume enclosed by a membrane) that contains three types of things: compounds, cellular machines, and a genome.
53  <p>  <p>
# Line 59  Line 101 
101  </table>  </table>
102  <br><br>  <br><br>
103  <hr>  <hr>
104    The process of building a protein as a string of amino acids from the gene containing codons is
105    called <b>expressing</b> the gene.
106    <br>
107    A <b>subsystem</b> (i.e., an abstract cellular machine) is a set of functional roles.
108    Each protein implements one or more functional roles.  The set of functional roles
109    implemented by the protein is called the <b>function of the protein</b>.  The function of a  multifunctional
110    protein that implements {functional-role-1,functional-role-2} is normally written as
111    <i>functional-role-1 / functional-role-2</i>.
112    <br><br>
113    A <b>populated subsystem</b> is a subsystem with an attached spreadsheet.  Each column
114    in the spreadsheet corresponds to a functional role in the subsystem, and each row corresponds to
115    a specific genome.  Each cell in the spreadsheet contains the genes from the corresponding genome
116    that implement the designated functional role (there may be 0 or more such genes).
117    <br><br>
118    We do not actually know what machines are present in a cell.  We are in the midst of a grand
119    effort to clarify which are there and what they do.  The formulation of subsystems as abstract machines
120    in which each row of the subsystem describes a specific cellular machine that is believed to be present,
121    represents a way to maintain a collection of estimates or assertions.
122    <p>
123    A <b>protein family</b> is defined to be a set of proteins that implement the same functional roles and
124    are similar over the entire lengths of the proteins.
125    <p>
126    We seek a situation in which each protein occurs in one or more subsystems and in a single protein family.
127    <p>
128    In any specific cell, sets of specific cellular machines are
129    switched on and off as units.  That is, they are <i>co-regulated</i>.  We will call such a set
130    of <i>co-regulated cellular machines</i> a <b>regulon</b> (note that a regulon is often a set containing
131    a single cellular machine).  A <b>state</b> of a cell will be defined
132    as the set of regulons that are operational at a point in time.  Thus, a state amounts to the set
133    of cellular machines that are operational at one instant.
134    <p>
135    Microarrays are, for a given genome, two lists of genes that "changed expression levels" between two states of a
136    cell.  Basicaly, the first list contains genes that were "active" during the first state, but not the second; and the
137    second list contains genes that were "active" in the second but not the first.  If a cellular
138    machine utilizes protein <i>X</i>, and <i>X</i> is in the first list, and if <i>X</i> is used in
139    only one cellular machine, then it would be reasonable to infer that you could say that the machine was
140    active in the first state, but not the second.
141    
142    <h2>The cell: the Enhanced Formlation Needed to Support Bioinformatics</h2>
143    
144    In the enhanced abstraction, we need to losen up some concepts.  In particular,
145    <ul>
146    <li> A <b>genome</b> is a set of strings in a 4-character alphabet.  Each of the strings
147    is called a <b>contig</b>.  Note that the concept as formulated covers both incomplete genomes and
148    genomes with multiple replicons.
149    
150    <li>The genes within a genome are of two distinct types:
151    <ol>
152    <li>those that describe how to construct a protein (i.e., prtein-encoding genes), and
153    <li>those that describe how to construct a string of RNA (i.e., how to construct a string in the
154    4-character RNA alphabet {A,C,G,U}).
155    </ol>
156    <br><br>
157    <li>The location of a gene is generalized to be a set of regions within the genome (that are
158    concatenated to form the instructions needed to construct either a protein or a string of RNA).
159    <li>A protein is a character in an alphabet that now includes the 20 character codes from
160    the basic abstraction plus a very limited set of extra codes.
161    We already have cases in which <i>selenocyctein</i>  and <i>pyrrolysine</i> appear as nonstandard
162    translations of codons, and there may eventually be more.
163    
164    <li>Each protein-encoding gene has both a DNA sequence (by defintion) and a translation.  However,
165    the translation is not required to exactly match what a codon-by-codon translation of the DNA sequence
166    would produce.  This allows us to handle the very rare instances in which selenocystein occurs as the translatin
167    of TGA or pyrrolysine occurs as a translation of TAG (and others, if necessary).
168    </ul>
169    
170    This loosened up formulation represents a very minimal set of changes.  They should be left out of the
171    basic tutorial for computer scientists and mathematicians.
172    
173    <h2>The cell: Adding the Concepts Needed to Discuss Transcriptional Regulation</h2>
174    
175    In the final version of the abstraction, we add the minimal set of notions needed to support
176    analysis of transcriptional regulation.  An <b>operon</b> is a set of contiguous genes that are all
177    on the same strand and are all co-regulated.  We consider a gene that is not co-regulated with any adjacent genes
178    to be an operon composed of just itself.  A <b>binding site</b> is a small region of DNA (normally
179    occurring a short space ahead of an operon) that acts as a switch turning the operon "on" or "off". When
180    a specific protein or expressed RNA called a <b>transcriptional regulator</b> binds the site, it flips the switch.  One or more
181    specific transcriptional regulators can bind a specific site (i.e., sets of
182     sites are associated with each specific transcriptional regulator).  The effect of a regulator binding at a site
183    always has the same effect (either activating or deactivating the operon), but which effect depends on
184    the site-regulator pair.
185    
186    <h1>Part 1: Tutorial Notes</h1>
187    
188    <h2>Notes for The Basic Abstraction</h2>
189    
190  We will be speaking about organisms that are a single cell.  At some point life began on earth.  We will be speaking about organisms that are a single cell.  At some point life began on earth.
191  The single-celled organisms that we know of replicate producing copies of themselves that have  The single-celled organisms that we know of replicate producing copies of themselves that have
192  genomes which usually have very, very similar content to that of the parent cell.  <b>Evolution</b> is the  genomes which usually have very, very similar content to that of the parent cell.  <b>Evolution</b> is the
# Line 75  Line 203 
203  although the  although the
204  numbers are rapidly growing for all three domains.  numbers are rapidly growing for all three domains.
205  <p>  <p>
206  This minimal notion of a cell is enough to explain some of the basic  The minimal notion of a cell is enough to explain some of the basic
207  problems in bioinformatics:  problems in bioinformatics:
208    
209  <h3>Identify the genes within a genome</h3>  <h3>Identify the genes within a genome</h3>
# Line 277  Line 405 
405  and trees in order to compare extant versions of proteins and gain insight into their historical origins  and trees in order to compare extant versions of proteins and gain insight into their historical origins
406  is considered basic to the task at hand.  is considered basic to the task at hand.
407    
408  <h2>Some Random Facts that You Should Absorb</h2>  <h3>Some Random Facts that You Should Absorb</h2>
409    
410  Most genomes of bacteria contain between 400,000 and 12,000,000 characters.  Most genomes of bacteria contain between 400,000 and 12,000,000 characters.
411  Normally, the genes in a genome  Normally, the genes in a genome
# Line 334  Line 462 
462  the specific role of the gene in sustaining the functionality of the machine.  the specific role of the gene in sustaining the functionality of the machine.
463  <br><br>  <br><br>
464    
465  <h2>Imposing a Structure on Characterizing the Inventory</h2>  <h23Imposing a Structure on Characterizing the Inventory</h2>
466    
467  One central goal of bioinformatics is to support an accurate characterization of the cellular  One central goal of bioinformatics is to support an accurate characterization of the cellular
468  machinery for each cell.  It is of major importance to biologsts that we be able to support  machinery for each cell.  It is of major importance to biologsts that we be able to support
# Line 393  Line 521 
521  handled by fairly general procedures.  handled by fairly general procedures.
522  </ul>  </ul>
523    
524  <h2>States of the Cell</h2>  <h3>States of the Cell</h2>
525    
526  The notion of <i>subsystem</i> was introduced as an <i>abstract machine</i> -- that is, as an  The notion of <i>subsystem</i> was introduced as an <i>abstract machine</i> -- that is, as an
527  attempt to create a framework for understanding variations within specific celular machines via  attempt to create a framework for understanding variations within specific celular machines via
# Line 417  Line 545 
545  and how transitions between are managed all are now parts of the picture being filed in.  and how transitions between are managed all are now parts of the picture being filed in.
546    
547    
548  <h2>Microarrays</h2>  <h3>Microarrays</h2>
549    
550  Microarrays are, for a given genome, two lists of genes that "changed expression levels" between two states of a  Microarrays are, for a given genome, two lists of genes that "changed expression levels" between two states of a
551  cell.  Basicaly, the first list contains genes that were "active" during the first state, but not the second; and the  cell.  Basicaly, the first list contains genes that were "active" during the first state, but not the second; and the
# Line 429  Line 557 
557  many microarrays, and if the specific cellular machines for the cell are known, then one could make  many microarrays, and if the specific cellular machines for the cell are known, then one could make
558  substantial progress in uncovering the exact composition of the regulons that make up the cell.  substantial progress in uncovering the exact composition of the regulons that make up the cell.
559    
560    <h2>Notes for the Enhanced Abstraction</h2>
561    
562    The process of <b>expressing a gene</b> amounts to using the gene to produce the functional component of
563    a machine (a protein for a protein-encoding gene, and an RNA for an RNA-encoding gene).
564    The process of expressing a protein-encoding gene takes a gene (a string of DNA formed by concatenating a sequence of
565    regions from contigs) and producing a protein is normally thought of as taking place in two steps.
566    <b>Transcription</b> is the process of a specific machine moving along the contig and making a copy of the
567    gene as RNA.  This string of RNA is then <b>translated</b> by a separate machine.  The machine that performs
568    the copying of the gene into a string of RNA is called an <b>RNA polymerase</b>.  The machine to translate
569    the RNA into a protein, the <b>ribosome</b>, is made up of both proteins and RNA components.
570    <p>
571    Machines can be made up of both protein and RNA components, although most machines are built from
572    just proteins. Some of the most fundamental questions in biology relate to how life started and the steps
573    required to gradually enrich the basic machinery to the point where this magnificent information storage and
574    maintenance system based on DNA, RNA and proteins could have arisen.  There is much that can be inferred by
575    reasoning back from what we now observe and reasoning forward from the relatively little we know of
576    what the early earth was like. One possible set of goals would be to first understand in detail the inventory
577    of components we now see in life forms, composing something analogous to a CAD/CAM system describing life forms.
578    Then, as a second step, to understand the sequence of transformations that led from some initial raw components
579    to initial life forms to those we have seen and characterized.
580    <p>
581    The need to allow occasional "nonstandard" characters in protein sequences and a loosening of the corespondence
582    between a gene and characters in the protein sequence it can be used to build results from the fact that
583    evolution has produced the existing genetic codes and they continue to evolve (either converging or diverging
584    depending on the outcome of basically random processes operating under selective pressure).
585    <br>
586    
587    <h2>Notes on the Abstraction Extended to Support Regulation</h2>
588    
589    There are two basically different regulatory mechanisms in the cell.  In one, you have a metabolic
590    network in which fluxes are tightly controlled by positive and negative feeback loops. This <b>metabolic
591    regulation</b> occurs very rapidly.  <b>Transcriptional regulation</b> occurs orders of magnitude more
592    slowly.  It is just this transcriptional regulation that we consider in this extension.
593    <p>
594    
595    As the cell changes state, regulons are activated or de-activated by
596    transcriptional regulators (either protein or RNA) binding to specific
597    sites in the DNA.  This model has the redeeming characteristic of
598    simplicity.  It is certainly the case that there are innumerable
599    important issues that it disregards (e.g., regulation based on DNA
600    packaging, due to small RNAs binding the RNAs produced by
601    transcription, etc.).  In forming any clear notion of transcriptional
602    regulation and how it is achieved, we will need to carefully separate
603    these different mechanisms, since they have fundamentally different
604    modes of control and operation.  We are arguing that the notion of a
605    protein or RNA being used to flip regulons on and off by binding to
606    control sites within the genome is a major form of regulation and
607    probably the right place to start any effort to formulate a useful
608    abstraction.
609    
610    <h1>The Role of Bioinformatics in Supporting the Genomic Revolution</h1>
611    
612    Within the growing genomics revolution, one can easily divide developments and
613    goals into those relating to advances in medicine and agricultue from those relating to
614    pure science.  Here we consider only issues relating to pushing advances in basic research.
615    Here is an overview of our perspective:
616    <ol>
617    <li> The different life forms that now exist were produced by an evolutionary process,
618    which leads to our view that comparative analysis is the key to understanding.  Biological
619    machines that exist in complex forms will often also still exist in simpler forms (usually
620    in simpler organisms).
621    <li> Unravelling exactly how a machine works is more easily done in simpler organisms.  They
622    are easier to work with, and it is easier to gather the data needed to support comparative analysis.
623    
624    <li> This leads to the view that we should try to understand single-celled organisms to lay
625    the foundation for analysis of multicelluar organisms.
626    
627    <li> The characterization of unicellular life will require access to orders of magnitude
628    more data than exist now (we have more-or-less complete genomes for about 1000 genomes, but
629    that represents a small fraction of a percent of extant single-celled life forms).
630    
631    <li> The immediate basic steps that are taking place are roughly:
632    <ol>
633    <li> Attempt to formulate a growing list of abstract machines that correspond
634    to the many specific machines that implement te same goal.  These abstract machines (subsystems)
635    represent the basic units that make up life forms.
636    
637    <li> Create protein and RNA families in which the members are all homologous (share a common ancestor),
638    remain similar over almost all of the sequence, and all implement a common function.
639    
640    <li> Build alignments for each protein family, along with phylogenetic trees that represent
641    an estimate of the history of how these specific sequences evolved.
642    
643    <li>Provide a computational framework to support continued maintenance and development of these
644    basic data types.
645    </ol>
646    
647    <li> A limited number of groups have progressed to the point where they can create models of
648    an organism that display predictive capabilities.  There are many forms of modeling.  In our view
649    it is important that we reach the state where we can routinely model states of the cell, transitions
650    between states, and metabolic characteristics of the cell.  We believe that it is now possible
651    to create fairly comprehensive representations of the metabolic networks of some bacteria.
652    In these cases, we have substantial amounts of physiological data, the number of abstract machines
653    in the cell is fairly limited, and it is possible to do compare the predictions against observed results.
654    
655    
656    </ol>
657    
658    
659    <br><br>
660    We do not actually know what machines are present in a cell.  We are in the midst of a grand
661    effort to clarify which are there and what they do.  Reaching a point where we have a near
662    complete overview of the basic inventory is arguably the highest priority at this point (we ignore
663    the medical revolution and numerous other wonderful advances, but...).
664    
665    The formulation of subsystems as abstract machines
666    in which each row of the subsystem describes a specific cellular machine that is believed to be present,
667    represents a way to maintain a collection of estimates or assertions.
668    <p>
669    A <b>protein family</b> is defined to be a set of proteins that implement the same functional roles and
670    are similar over the entire lengths of the proteins.
671    <p>
672    We seek a situation in which each protein occurs in one or more subsystems and in a single protein family.
673    The computational tasks imposed by such a goal are obvious:
674    <ul>
675    <li>We need to consruct databases that implement at least the following entities:
676    <ol>
677    <li>cells (i.e., each cell must have an ID and a set of attributes),
678    <li>genomes,
679    <li>genes,
680    <li>proteins,
681    <li>functional roles,
682    <li>subsystems, and
683    <li>protein families.
684    </ol>
685    <li> We need to add support for developing clues to function by integrating data
686    from sources like proximity within the genome, fusions, etc.
687    <li>We need to support a framework for the development of populated subsystems.
688    <li>We need to construct decision procedures for membership in protein families.  Some
689    of these procedures will be quite complex, although the majority of cases can be
690    handled by fairly general procedures.
691    </ul>
692    
693    
694    <h1> The Role of Abstraction in Setting the Stage for Software Development and Modeling</h1>

Legend:
Removed from v.1.3  
changed lines
  Added in v.1.4

MCS Webmaster
ViewVC Help
Powered by ViewVC 1.0.3