Testing for the Time Dependent Transition from a Rates Across Sites Model to a Covarion Model of Protein Evolution

Protein-coding gene sequences typically evolve constrained by the requirements for a protein to fold into its three dimensional structure. These constraints can dictate evolutionary rates at different sites, where residues in the hydrophobic core of a protein typically evolve more slowly than those on the surface. The exceptions to this are surface residues involved directly in functions of the protein such as binding, which are conserved. A gamma distribution of rates across sites is typically used to describe this process of protein evolution in what is called the RAS model. However, it has been proposed that this model is violated when a functional change in the protein occurs, and in this case, a covarion model of protein evolution would be more accurate. It has been proposed that protein structure may lead to violations of the gamma distribution over increasing evolutionary time. Because of this, I hypothesize that the rate of the transition is fold-dependent and that different protein folds will move from an RAS model to a covarion model at variable rates that are influenced by selective pressures. In preliminary analysis of an SH2 domain, there is a time dependent transition from an equal rates to a RAS to a covarion model of protein evolution.