Better Models
by BradfordIn their PLOS One paper titled Stylus: A System for Evolutionary Experimentation Based on a Protein/Proteome Model with Non-Arbitrary Functional Constraints, Douglas D. Axe, Brendan W. Dixon and Philip Lu of the Biologic Institute, introduced new software known as Stylus. Stylus is intended to more realistically model proteins and evolutionary constraints with more accurate linkage of protein structure to function. The nature of some current models, compromises accurate assessments of protein transition possibilities and the mapping of sequences to changes in function.
Human languages have been used for the purpose of analogy to biological polymers like polypeptides. The analogous nature of letters to amino acids and whole sentences to function, suggests linguistic models. Stylus however is based, not on alphabets common to European languages, but rather on characters of Chinese origin. This is intended to generate more relevant biological models.
The intent of Stylus is to represent proteins which have a 20 letter alphabet. Unlike DNA, the manner in which proteins fold is critical to their function because protein structure (a three dimensional protein property) is essential to understanding protein function whereas sequencing and nucleotide identity explain DNA coding properties. The utility of Chinese pictograms lies in their analogy to structure. Quoting from the paper:
One such insight is that protein-like models (in contrast to RNA models) tend to show sparse connectivity between regions of sequence space that encode different structures [1]. In other words, stepwise paths through sequence space that accomplish a structural transformation without passing through unstructured intermediates appear to be rare. This clearly fits expectations for real proteins, where reorganization of core structure would seem to require complete loss of structure (and therefore function) along the way [4]. It also fits experimental observations, which show that the expected deterioration is common not only for transitions between different folds [5] but also, more surprisingly, for transitions between different sequences encoding the same fold [6].
The authors are pointing out that gradual changes, that would produce new protein innovations, are made problematic by properties innate to proteins. Even small alterations of proteins can radically alter protein structure. Effects of slight sequence changes on protein folding also are difficult to predict. Differences between protein families can entail few transition possibilities. A natural inference is an expectation of distinct structural origins for different functional groups of proteins.
July 23rd, 2008 at 1:36 pm
Bradford, let me make sure I understand the analogy: Chinese characters actually represent words or concepts, not sounds? And this would better correspond to proteins? Now can't one add onto Chinese characters by adding a mark or something, and expand their meaning? Say, for example, from "ball" to "green ball"?
Comment by Bilbo — July 23, 2008 @ 1:36 pm
July 23rd, 2008 at 3:42 pm
No, not really. At least not in any sort of arbitraty way. Many of their characters are sort of composites of other characters though, like you might find 小 (small) and 亻(an alternate form of 人, or "people") as part of another larger character, say 你 (you). So the three radicals that form "you" can sort of be thought of as meaning "small equal people." That doesn't mean you can just toss a 小 into another character to make it a "small whatever."
Comment by Todd Berkebile — July 23, 2008 @ 3:42 pm
July 24th, 2008 at 10:14 am
Hi Bilbo,
My knowledge of Chinese symbols is very limited. Apparently sometimes symbols suggest that which they represent. Here's an interesting site with information about both pictographs and ideographs. Sometimes the relationship between symbols and meaning can entail arbitrary assignment of symbols to meanings. I believe I have an inkling as to what the software designers had in mind when they chose Chinese characters to model protein folding dynamics but since I do not have it from the source I'll instead focus on protein properties.
A major focus of the linked paper seems to be evolutionary transitional states where new functions are enabled by changes in the amino acids of proteins. Modularity is featured in that within major protein families certain commonalities are shared by family members. For example, alpha helices, are structural elements found in proteins. They can be particular to a protein group. A basic structural feature may be common to proteins found in group x while a different feature may mark proteins of group y. Differences between proteins in group x may involve small variations in protein structure. Slight refinements of a basic shared template.
Interestingly enough sometimes a single amino acid substitution, if at a critical juncture, can have a large impact on the folding of that protein thereby affecting protein shape. Structure is critical to protein function. Changes in protein structure impact function giving rise to significant phenotypic effects from small changes in genotype. The idea of newer software is more accurate predictions of how changes in the primary structure of proteins impacts the tertiary structure of those same proteins.
Comment by Bradford — July 24, 2008 @ 10:14 am
July 28th, 2008 at 12:31 pm
I have read the paper. It looks very interesting. The main point of course is to build a heuristic (possibly partial) correspondence between Chinese characters (which have a hierarchical structure) with proteins. If there is even a subset that can be corresponded, that would be interesting. But more than that, since Chinese characters, at least for some, are pictographs, which have visual correspondence with the real things. For example, the word 'bird' in Chinese looks like a bird with head, peak, eyes and body etc. If there is indeed a correspondence, it would be very surprising- that a language construct, the visual imagery and proteins share some common structural characteristic. Nobody would expect it, I suppose. What I am saying is that they have now proposed a tool to find whether there is such a common structure in nature. I do not see Darwinism has anything to say about this.
Comment by yc2008 — July 28, 2008 @ 12:31 pm