Detection of an ancient principle and an elegant solution to the protein classification problem

Abstract

This work is concerned with the development of a well-founded, theoretically justified, and least complicated metric for the classification of proteins with reference to enzymes. As the signature of an enzyme family, a catalytic domain is easily fingerprinted. Given that the classification problem has so far seemed intractable, a classification schema derived from the catalytic domain would be satisfying. Here I show that there exists a natural ab initio if nonobvious basis to theorize that the catalytic domain of an enzyme is uniquely informative about its regulation. This annotates its function. Based on this hypothesis, a method that correctly classifies potassium ion channels into their respective subfamilies is described. To put the principle on firmer ground, extra validation was sought and obtained through co-evolutionary analyses. The co-evolutionary analyses reveal a departure from the notion that potassium ion channel proteins are functionally modular. This finding is discussed in light of the prevailing notion of domain. These studies establish that significant co-evolution of the catalytic domain of a gene with its conjoint domain is a specialized, necessary process following fusion and swapping events in evolution. Instances of this discovery are likely to be found pervasive in protein science.