|
Author
|
Topic: Self-Disclosive Design (SDD) as an ID Research Program
|
William A. Dembski
Member
Member # 7
|
posted 14. March 2002 17:53
In a recent speaking tour of Canadian universities, I met with a colleague who works in bioinformatics. He described some research of his that holds exceptional promise for ID. I shall leave him unnamed because this research is at the beginning stages and also because ID researchers who are publicly identified as such face enormous pressures.
Briefly, what this researcher found was that even though DNA sequence information coding for a given protein within a given organism appears to give no information about the folding of that protein, by looking at an array of homologous DNA sequences for the same protein type across multiple organisms does give reliable 3-dimensional information about the folding of the protein (he has published some preliminary results here).
This is remarkable because on a Darwinian hypothesis, evolution cobbles together organisms and their various components, so homolous structures should not give any information about deep underlying design plans. Yet, that is what seems to be happening here -- the evolutionary phylogeny holds information about the underlying design. It is as though biology is not only rife with design but with self-disclosive design, in which the design assists in discovering the method of design. Alternatively, the design comes with a "design manual" thrown in for good measure.
Let me be quick to note that such self-disclosive design is not a prediction of ID that, if not borne out, would falsify the theory. Design can be real without being self-disclosive. Rather, it is a possibility that the designs in biology come with embedded information that assists in discovering the "how" of the design. If you will, the designs in biology could be such that they facilitate discovery into the nature of the design and where the facilitation is such that it was itself designed.
This possibility, it seems to me, if confirmed across a wide range of biological phenomena, would not only be enormously fruitful in biological investigation (biology would in that case be providing us with the key to its own secrets), but would provide overwhelming confirmation of ID over blind evolutionary processes. Blind evolutionary processes can conceivably reward designs that assist the organism, but there is no reason that such processes should reward scientists by doing their work for them. Self-disclosive design would be a higher order design that holds no advantage to the organism but only to the scientist intent on understanding the organism.
The possibility of self-disclosive design is just that at this point -- a possibility. But it seems borne out in some cases already (i.e., in the work of my bioinformatics friend). More significantly at this point in the debate, it constitutes a conjecture that flows out of ID but cannot be squared with any non-telic approach to evolution.
IP: Logged
|
|
charlie d.
Member
Member # 159
|
posted 14. March 2002 18:24
Maybe I am misunderstanding, but how is this different than the strategy widely used in structural biology of giving evolutionary conserved aminoacid residues more "weight" in determining protein 3-D structures? (The logic behind it being, of course, that the more evolutionarily conserved a residue, the more likely that the residue is important for secondary and tertiary protein structure)
IP: Logged
|
|
Drosera
Member
Member # 139
|
posted 15. March 2002 04:06
I think we're talking about homology modelling, which is indeed relatively new although not radically so. Web searches turn up lots of hits, as does Pubmed.
Comparative or Homology Modelling
Predicting reliable regions in protein sequence alignments
quote:
Bioinformatics 2002 Feb;18(2):306-14
Predicting reliable regions in protein sequence alignments.
Cline M, Hughey R, Karplus K.
Center for Biomolecular Science and Engineering Department of Computer Engineering, Jack Baskin School of Engineering, University of California, Santa Cruz, CA 95064, USA.
Motivation: Protein sequence alignments have a myriad of applications in bioinformatics, including secondary and tertiary structure prediction, homology modeling, and phylogeny. Unfortunately, all alignment methods make mistakes, and mistakes in alignments often yield mistakes in their application. Thus, a method to identify and remove suspect alignment positions could benefit many areas in protein sequence analysis. Results: We tested four predictors of alignment position reliability, including near-optimal alignment information, column score, and secondary structural information. We validated each predictor against a large library of alignments, removing positions predicted as unreliable. Near-optimal alignment information was the best predictor, removing 70% of the substantially-misaligned positions and 58% of the over-aligned positions, while retaining 86% of those aligned accurately.
...or in the less-than-new category:
quote:
Protein Eng 1992 Jun;5(4):305-11
Towards an automatic method of predicting protein structure by homology: an evaluation of suboptimal sequence alignments.
Saqi MA, Bates PA, Sternberg MJ.
Biomolecular Modelling Laboratory, Imperial Cancer Research Fund, London, UK.
A major problem in predicting protein structure by homology modelling is that the sequence alignment from which the model is built may not be the best one in terms of the correct equivalencing of residues assessed by structural or functional criteria. A useful strategy is to generate and examine a number of suboptimal alignments as better alignments can often be found away from the optimal. A procedure to filter rapidly suboptimal alignments based on measurement of core volumes and packing pair potentials is investigated. The approach is benchmarked on three pairs of sequences which are non-trivial to align correctly, namely two immunoglobulin domains, plastocyanin with azurin and two distant globin sequences. It is shown to be useful to reduce a large ensemble of possible alignments down to a few which correspond more closely to the correct (structure based) alignment.
Dr. Dembski writes,
quote:
In a recent speaking tour of Canadian universities, I met with a colleague who works in bioinformatics. He described some research of his that holds exceptional promise for ID. I shall leave him unnamed because this research is at the beginning stages and also because ID researchers who are publicly identified as such face enormous pressures.
Briefly, what this researcher found was that even though DNA sequence information coding for a given protein within a given organism appears to give no information about the folding of that protein, by looking at an array of homologous DNA sequences for the same protein type across multiple organisms does give reliable 3-dimensional information about the folding of the protein (he has published some preliminary results here).
I think the issue is simply that directly calculating a correct 3-D protein structure from a sequence of amino acids is simply very, very difficult computationally. We are talking about the relative influences of thousands of variously charged atoms on each other.
quote:
This is remarkable because on a Darwinian hypothesis, evolution cobbles together organisms and their various components, so homolous structures should not give any information about deep underlying design plans.
???? I don't see how this follows at all. Homology modelling, as far as I can tell, is simply based on the commonly-made observation that protein structure is more conserved than amino acid sequence (which is more conserved than DNA sequence) over long time periods.
Therefore, if you suspect homology based on weak amino acid sequence similarity, then you might try and predict the structure of the protein by comparison with the homologous protein where the structure has been solved. Presumably this would be cheaper and faster than the time-consuming and expensive task of solving structures directly via crystallography etc. (which can, however, be used as a test of the method).
quote:
Yet, that is what seems to be happening here -- the evolutionary phylogeny holds information about the underlying design. It is as though biology is not only rife with design but with self-disclosive design, in which the design assists in discovering the method of design. Alternatively, the design comes with a "design manual" thrown in for good measure.
Let me be quick to note that such self-disclosive design is not a prediction of ID that, if not borne out, would falsify the theory. Design can be real without being self-disclosive. Rather, it is a possibility that the designs in biology come with embedded information that assists in discovering the "how" of the design. If you will, the designs in biology could be such that they facilitate discovery into the nature of the design and where the facilitation is such that it was itself designed.
This possibility, it seems to me, if confirmed across a wide range of biological phenomena, would not only be enormously fruitful in biological investigation (biology would in that case be providing us with the key to its own secrets), but would provide overwhelming confirmation of ID over blind evolutionary processes. Blind evolutionary processes can conceivably reward designs that assist the organism, but there is no reason that such processes should reward scientists by doing their work for them. Self-disclosive design would be a higher order design that holds no advantage to the organism but only to the scientist intent on understanding the organism.
See above. It seems to me that the homology modelling method follows quite logically from a few simple observations about common descent and divergence of proteins.
quote:
The possibility of self-disclosive design is just that at this point -- a possibility. But it seems borne out in some cases already (i.e., in the work of my bioinformatics friend). More significantly at this point in the debate, it constitutes a conjecture that flows out of ID but cannot be squared with any non-telic approach to evolution.
Bioinformatics is indeed an interesting topic, considering the absolutely ubiquitous concept of homology within the field. It seems to me that bioinformatics is one of the most fantastic examples of an entire new hot field being born directly out of modern evolutionary theory plus DNA sequencing plus computers.
Drosera
IP: Logged
|
|
James A. Barham
Member
Member # 50
|
posted 15. March 2002 09:04
Homology (that is, common evolutionary descent) is not the only explanation for the non-randon DNA-protein correlations discussed by Dembski and Drosera.
Another explanation, or partial explanation, is the existence of intrinsic physical affinities between nucleotide sequences and amino acids. There is in fact mounting empirical evidence for this. See:
R.D. Knight et al., "Selection, History and Chemistry: The Three Faces of the Genetic Code," Trends in Biochemical Science, 1999, 24: 241--247.
R. Morchio et al., "Proteins, Nucleic Acids and Genetic Codes," Rivista di Biologia/Biology Forum, 2001, 94: 37--58.
Such intrinsic affinities between codons and amino acids are of course what one would expect from the self-organization point of view.
IP: Logged
|
|
William A. Dembski
Member
Member # 7
|
posted 15. March 2002 10:47
It is remarkable how just about any evidence can be conveniently co-opted into one's theoretical framework. Apparently Darwinism can perfectly account for self-disclosive design and self-organization can as well. I expected this reaction. So does that mean that theory choice (ID, Darwinism, self-organization) in the case of biological evolution is purely a matter of preference? Is there always empirical equivalence? Or could there be a type of self-disclosive design that would provide compelling evidence for one position over another?
IP: Logged
|
|
James A. Barham
Member
Member # 50
|
posted 15. March 2002 11:47
Bill:
You have a valid point about the interpretation of evidence in light of one's preferred theoretical framework, but I would make the following observations:
First, no one observation is ever going to be conclusive. It is rather the overall balance of observations that counts.
Second, the real problem is that we are all grasping at straws. No one really understands what is going on in the cell. We are trying to piece it together as we go. In short, our disagreements are a symptom of the immaturity of biology as a theoretical science. As theoretical biology matures, this will undoubtedly all get sorted out.
Now, the above is admittedly a cop-out. What can I say in reply to your comment in the light of our current knowledge? I guess I would like to venture the following:
(1) It seems to me that the standard mechanistic-reductionist interpretation of natural selection most naturally predicts either no correlations between DNA and proteins at all, or only ones based on strict homology. Admittedly, selection theorists try to use homoplasies---i.e., convergences---as evidence in favior of their viewpoint, too, but I think that just shows the vacuousness of selection theory. As Dennett says, it is a "universal acid" that explains everything under the sun---naturally, everything that exists is "selected" in a broad enough sense of that term---and so explains nothing at all.
(2) ID, it seems to me, most naturally predicts convergences of the sort we encounter in human engineering, based on design requirements under similar constraints---such as the streamlining we find in ariplanes (and birds) and submarines (and fish). That is certainly confirmed in nature at all levels. But why would ID per se predict physical affinities? As I undertand it, an important claim of ID is that organisms are like manmade machines in which the functionality is arbitrary with respect to the physical "instantiation." This is the basic meaning of the "boundary conditions" notion, isn't it? If we drop this requirement, then ID seems to collapse into self-organization theory.
(3) Self-organization theory, I think, is the only one of the three frameworks that predicts---indeed, requires---physical affinities of some sort. To the extent that physical affinities are empirically confirmed, then that is evidence in favor of self-organization theory in contrast to either natural selection or ID.
James
IP: Logged
|
|
John Bracht
Member
Member # 5
|
posted 15. March 2002 15:39
I agree with Drosera and charlie d. on this one, in the sense that I think that what William Dembski calls "self-disclosive design" is a result of structural constraints on proteins and therefore is compatable with either evolution or design.
Drosera and charlie d. both point out that homology would predict that the important residues would be conserved and hence statistical analysis should help us determine which residues are important for 3D structure. Once we know which residues are important, we have some very helpful clues as to what the 3D structure might be.
On the other hand, these structural constraints will apply to any designer as well. If a designer is trying to achieve a certain 3D structure and function, they will have to include the essential residues at key positions in the sequence. Other nonessential residues (with regard to 3D structure) might vary from organism to organism and provide different (nonstructural) organism-specific functions.
The bottom line is that these statistical studies seem to be getting at important information regarding the folding and 3D structure of proteins. These structural constraints, it would seem, apply to natural selection and intelligent design equally.
The difficulty in modelling protein secondary and tertiary structure was well-stated by Drosera:
quote:
I think the issue is simply that directly calculating a correct 3-D protein structure from a sequence of amino acids is simply very, very difficult computationally. We are talking about the relative influences of thousands of variously charged atoms on each other.
I think this is the reason why statistical data from comparing multiple sequences is needed to help determine the 3D structure: protein sequences innundate us with a wealth of information and we need to know what information to ignore from a structural perspective. In a sense, we need some sort of higher-order information telling us what is important and what isn't. Statistical analysis provides that higher-order information and allows us to throw out the nonessential residues and focus on those residues that hold clues to 3D structure.
Is it fair to call this "self-disclosive design"? As noted above, the pattern seems to arise more from structural constraints than from a tendency of nature to help us discover her secrets. In other words, I don't think there is any contingency here. Thus, I conclude that these patterns are an inherent part of the structural requirements of protein folding, and while they do help us to discover nature's secrets, they cannot help us distinguish design from unguided evolution.
John Bracht
[ 15 March 2002, 16:43: Message edited by: John Bracht ]
IP: Logged
|
|
Jules
Member
Member # 181
|
posted 15. March 2002 17:52
I tread into unknown territory here, as I do anytime I discuss biology. I'm not sure I understood Dr.Dembski's opening post, but it sounded as if his colleague's research was trying to predict the shape of proteins by the location of certain amino acids in their sequence. Have I missed the boat already?
If not, then there may be an argument for design here. From the little I remember of discussions of protein homology back at ARN (oh how I miss my home), the standard minimum for homology was placed at 20% sequence similarity. This presented a problem in a number of instances where similar protein structures with similar functions were found in organisms, but the sequence similarity was less than 20%. Non-IDists argued that this could still be considered homology. However, if there is a way to predict protein structure based on key amino acids in the chain, then a designer could merely select a protein chain that had those key amino acids, regardless of homology.
No doubt I've misunderstood everything. If by some miracle I haven't, then there does indeed seem to be evidence of design here.
IP: Logged
|
|
Jules
Member
Member # 181
|
posted 17. March 2002 09:41
Interpreting silence is always a tricky business. In the present case, some might think that no one has responded because I have revealed so large and fundamental a misunderstanding of the topic, that short of writing a chapter to a book there is no hope of enlightening me. I prefer to think of it as evidence that what I have said is unassailable.
IP: Logged
|
|
Mike Gene
Member
Member # 149
|
posted 17. March 2002 19:21
Dembski: Let me be quick to note that such self-disclosive design is not a prediction of ID that, if not borne out, would falsify the theory. Design can be real without being self-disclosive. Rather, it is a possibility that the designs in biology come with embedded information that assists in discovering the "how" of the design. If you will, the designs in biology could be such that they facilitate discovery into the nature of the design and where the facilitation is such that it was itself designed.
It's possible to approach this perspective from the opposite angle. That is, how are proteins rationally designed by intelligent bioengineers? ID theorists attempting to reverse engineer life have much to use from scientists trying to engineer proteins and other molecules, as the approaches appear complementary:
quote:
Design of a protein requires that both a structure and a sequence are specified. The basic forces that determine the noncovalent interactions within the polypeptide chain, with the surrounding solvent, and with ligands are relatively well understood: van der Waals and electrostatic interactions, hydrogen bonds, the hydrophobic effect, and the favorable packing interactions associated with the condensed state of protein interiors. However, the number of conformations a particular polypeptide can potentially adopt as well as the number of different sequences that can be built into even a small protein is vast. Furthermore, many of these sequences and their conformations are distinguished only by relatively small energy differences. The combination of the immense combinatorial complexity and subtle energetic differences turns the seemingly simple basic interactions into a dauntingly complex landscape of virtually infinite possibilities. The ability of an algorithm to explore this vast landscape and seek out preferred solutions that have to be distinguished from closely related inferior possibilities is therefore a crucial component of any rational design approach. All design methods use the same general approach to reduce the immense complexity of the search problem. The structure of a protein backbone is chosen a priori, kept fixed, and redecorated with different amino acid sequences that are predicted to be structurally compatible with that fold. This "inverse folding" approach therefore removes the backbone conformational degrees of freedom from the design problem.
The first rational design approaches used qualitative rules of protein structure applied by inspection . These experiments established that it is possible to create sequences de novo that adopt defined structures. Furthermore, they demonstrated that, by following a progressive design strategy [or "hierachic design"] in which increasing levels of complexity are iteratively introduced, new insights into the fundamentals of protein structure and function can be gained. One of the remarkable observations of these experiments was that it is surprisingly easy to obtain globally correct folds. However, the local details were found to be difficult to get correct. The interiors of these designed proteins show a high degree of disorder, which does not resemble the tightly packed, unique arrangement of natural systems. Global correctness in these designs apparently resulted from incorporation of the correct "binary pattern" of hydrophobic and hydrophilic residues, which sets up the geometric specification of the protein interior and exterior for the hydrophobic effect to act on. The difficulty in designing well-ordered cores can be viewed as a problem in specificity. The side chains in a disordered core adopt many alternative conformations of approximately equal energy, instead of assuming a single, specific arrangement. Hellinga -emphasis added
Thus, when Jules writes, However, if there is a way to predict protein structure based on key amino acids in the chain, then a designer could merely select a protein chain that had those key amino acids, regardless of homology , I have to agree. Structural similarity is a rather useless criterion for homology if one has reason to think ID may be involved nearby.
Drosera writes: Homology modelling, as far as I can tell, is simply based on the commonly-made observation that protein structure is more conserved than amino acid sequence (which is more conserved than DNA sequence) over long time periods.
Sure. But why does this commonly-made observation exist? We can start with this question and see where it leads us.
Barham: (2) ID, it seems to me, most naturally predicts convergences of the sort we encounter in human engineering, based on design requirements under similar constraints---such as the streamlining we find in ariplanes (and birds) and submarines (and fish). That is certainly confirmed in nature at all levels. But why would ID per se predict physical affinities? As I undertand it, an important claim of ID is that organisms are like manmade machines in which the functionality is arbitrary with respect to the physical "instantiation." This is the basic meaning of the "boundary conditions" notion, isn't it? If we drop this requirement, then ID seems to collapse into self-organization theory.
I don't know of any physical affinity between the sequence of codons and the 3-D shape of any protein. It is my impression that the papers James cited above deal with the origin of the genetic code, arguing that there is a chemical basis for codon assignment (at least for some codons).
IP: Logged
|
|
James A. Barham
Member
Member # 50
|
posted 17. March 2002 23:35
For Mike Gene:
Thanks for the clarification. I see that I was failing to make a crucial distinction between possible physical affinities between nucleotides and amino acids, on the one hand, and correlations between genes and three-dimensional, functional protein structures, on the other.
But can I ask you this? If there really are affinities at the level of the primary or linear structure, then that surely means that the notion that the gene-protein relationship is purely arbitrary is untenable, no? Would we not then have to posit something like a feedback from the 3-D structure onto the genome?
I suppose you might say that this is all that natural selection is saying. But surely a crucial part of the Darwinian scenario is unidirectionality of causal influence from the genome (which can supposedly be considered mechanistically) to the proteome (where the functionality lies). It is the necessity of this unidirectionality for the mechanistic-reductionist scheme that makes Crick's "central dogma" so important. But if there really is "feedback" from the proteome back onto the genome (as Caporale, von Sternberg, and others are arguing), then couldn't we say that it is the functional, 3-D level that is determining the linear level, at least as much as the othe way around?
In short, if there are any affinities at all---if the correlations are partly physically determined and not purely arbitrary, as functionalism and the machine analogy demand---then don't we have to say that teleology is an active and irreducible element in the picture, even if the affinities are only indirect?
IP: Logged
|
|
ordinary man
Member
Member # 188
|
posted 19. March 2002 11:01
quote:
Briefly, what this researcher found was that even though DNA sequence information coding for a given protein within a given organism appears to give no information about the folding of that protein, by looking at an array of for the same homologous DNA sequences protein type across multiple organisms does give reliable 3-dimensional information about the folding of the protein (he has published some preliminary results here).
Just a side note -- is Dembski using "homologous" in the sense that evolutionary biologists use the term -- as a sign of common descent? If so, then how can ID use homology as evidence for design when homology itself is taken as evidence that several DNA sequences were descended from a common ancestor? At the very least this argues against those creationists who deny that similar DNA sequences are evidence of common descent. In other words, getting many creationists to agree that there is such a thing as homology is a victory in itself. Now homology is being used as evidence against natural selection. How odd.
IP: Logged
|
|
Moderator
Administrator
Member # 1
|
posted 19. March 2002 11:46
ordinary man,
Your talk of "victory" indicates that you are here to fight a battle. Brainstorms is not the appropriate place to do this.
Indeed, in your quotation, Dembski discusses what he believes to be positive evidence for self-disclosive design, not evidence against natural selection.
Side notes which "stir the pot" are not welcome at ISCID. You have been granted a warning.
IP: Logged
|
|
charlie d.
Member
Member # 159
|
posted 20. March 2002 12:22
Well, trying to get back to the original issue, that is whether SDD can indeed be an ID research program, I have to say that after thinking about it, I am fairly convinced it cannot, at least as stated by Dembski.
First of all, I’d like to spell out what I think are the 2 alternative interpretations of the original post, which at this stage are still unresolved, pending some clarification from Dembski and/or his friend.
Interpretation #1: Dembski’s friend is using basic DNA database search tools to look for significant sequence similarity between structurally/functionally corresponding proteins (I don’t want to say homologous) from phylogenetically distant organisms in order to identify conserved residues/patterns (e.g. hydrophobicity, charge etc) that are likely to be more important for protein structure, and can therefore be used effectively in molecular modeling.
If this is the case, it would be old news. Also, far from Dembski’s later complaint that “any evidence can be conveniently co-opted into one's theoretical framework”, the entire rationale behind this strategy would be perfectly in keeping with, and in fact was long predicted by, a darwinian framework. So let’s ignore this boring possibility and focus on…
…interpretation #2. That is: Dembski’s friend, by comparing DNA sequences from genes of structurally/functionally corresponding proteins in phylogenetically distant organisms has found some sort of regularity/pattern/algorithm in the nucleotide sequences themselves that allows him to make (more) accurate predictions about the protein’s 3-D structure.
Except for the potential role that restrictions of codon affinities for certain aminoacids may have played in establishing the genetic code (as pointed out before in the thread), current molecular biology does not posit DNA as carrying any additional embedded information on secondary or tertiary protein structure beyond that determined by primary aminoacid sequence. As far as I can tell, no known mechanism would account for such structural encoding. Thus, I agree, this would make a GREAT subject for scientific inquiry.
A couple of interesting research plans could clearly be pursued without any preconceived model in mind (either darwinian or ID). One could i) use the information, whatever its origin and significance, to study protein structure, and ii) extend the database to multiple protein families and organisms to establish the generality of the “rules”. No problem with either of these, I suppose, in fact they would both be necessary.
Now, what would a darwinian do? I think the main focus of research would be to understand the origin of the observed regularities. For instance, one would try to elucidate how such structure-encoding patterns may have been established, by looking at their potential selective advantage for primordial organisms, much like studies on the adaptive logic behind the genetic code have pretty much undermined the original, neutralist “frozen accident” hypothesis. Also, the comparative analysis of structural changes during the evolution of homologous protein families, not only within their primary structure, but also their genes’ DNA sequences may be used to shed new light onto the evolution of such families. Well-known phylogenetic trees of related proteins could be re-analyzed to reassess the assumptions about adaptive evolution (see for instance Kirk’s “Ka/Ks ratio” thread for reference). Basically, one would ask the usual darwinian questions: what good is this feature to the organism? What are the constraints on this advantage? How could these advantages and constraints have played out during life’s history? Note that very clearly one could come up empty-handed with all three questions, but certainly there would be research going on.
On the other hand, in his original post Dembski immediately defines the problem in irreducible, unknowable terms, by stating, as his assumption, that “self-disclosive design would be a higher order design that holds no advantage to the organism but only to the scientist intent on understanding the organism”. Thus, like calling the complexity of biochemical cascades or molecular machines “irreducible” by definition, the intellectual boundaries of this budding ID-inspired research program are placed at its onset. Try as I may, I can’t think of any novel SDD-based experiments that can be carried out within this framework: any testable hypothesis would be "out of bounds". What questions can be asked? What answers would an ID scientist be looking for? Or am I just too engrained a darwinian?
Thus, I think the challenge for design scientists in SDD, like in IC/CSI, will boil down to the need to formulate models and allow for actual testable mechanisms. If Dembski or his friend would care to give more factual details, or maybe even just speculate on potential mechanisms/significance, we could discuss this further. If true, it surely sounds like an intriguing finding.
IP: Logged
|
|
YZ2
Member
Member # 91
|
posted 12. April 2002 15:53
I did some work previously that can lead to the possibility of such a research program:
1) Reference: Chiu, D.K.Y. and T. Kolodziejczak, Inferencing consensus structure from nucleic acid sequences, in Computer Applications in Biosciences (the former Bioinformatics journal), vol.7, no.3, pp.347-352, 1991.
This paper discusses an information-theoretic method based on statistical interdependency between molecular sites to infer secondary and tertiary bondings in transfer RNA and later in 5s rRNA as well. The method is not found to be successful to apply to more complex molecules. However, even tertiary bondings can be reflected by statistical interdependency is now fairly well accepted. The interesting aspect of this work is that statistical interdependency is not observed in conserved sites, but in certain non-conserved sites with varied type (varied among species).
2) Reference: Wong, A.K.C., D.K.Y. Chiu, W. Huang, A discrete-valued clustering algorithm with applications to biomolecular sequences, Information Science, 139 (2001), 97-112.
This paper described a more powerful clustering algorithm which found inherent groupings within cytochrome c sequences from different species. That is, species within groupings are much more similar than species between groupings. In another words, if the groupings are inherent, species may not be readily "evolving" from one group to the next. There are some errors in the algorithm evaluation when applies to the sequence data, but the main implication that inherent groupings exist in the cytochrome c data should be valid. The method is not based on heuristics in comparing sequences, but is based on selected statistical evidences in generating groupings. Again, interdependent (varied) sites are used, rather than conserved sites of the biomolecules as evidences.
A summary survey paper, discussing other aspects between sequences and high-level constructs such as biological functions (e.g. hereditory factors in cancer) can be found in
Chiu, D.K.Y. Discovering knowledge from statistical patterns in biomolecules, Recent Research Developments in Pattern Recognition, B. Chandrasekaran, M.D. Levine, C.H. Chan (eds.), Recent Research Development Series, Transworld Research Network Pub., 1 (2000): 283-300/.
An implication from the reports is that there are confirmation of an intuitive unity of biomolecules with respect to sequences, molecular structure, taxonomy and biological functions (and possible refutation if more evidences are accumulated). The surprising result is that the apparent discovered unity is NOT due to conserved molecular segments, but existing diverse varied sequence characteristics. Otherwise a simple searching algorithm will be sufficient to retrieve many of the properties of sequences. In other words, there are evidences of a HIGHER-ORDER formation of diverse biomolecules. To relate to the discussion of evolution or ID, the question is: Which comes first - the higher-order formation or the diverse individual sequences? My impression is that generating the higher-order formation from diverse individual sequences blindly is very unlikely. My next question is: Is the higher-order formation we observe today due to a common ancestry or a common design? The problem with the explanation of a common ancestry is that - What mechanism causes the common ancestry which already encodes the reflected unity to diverge into different sequences? As far as I understand, there is no obvious answer to this from natural selection which is based on a low-order formation of sequences. Even if some form of common ancestry is possible, it does not disregard the necessity of a common high-order design predating the common ancestry in such formation. Another point to make is that, the higher-order formation and the unity in biomolecules emerges empirically from observed data. The implications are generated from adopting the simplest explanation from observed data. As a research program, the approach
follows the traditions of Newton, Pascal and other earlier scientists who through observing the data, discovered many of the well known laws in Physics.
IP: Logged
|
|
|
Printer-friendly view of this topic