CORRELATION BETWEEN THE SUBSTITUTION RATE AND RATE VARIATION AMONG SITES IN PROTEIN EVOLUTION

It is well known that the rate of amino acid substitution caries among different proteins and among different sites of a protein. It is, how ever, unclear whether the extent of rate variation among sites of a pr otein and the mean substitution rate of the protein are correlated. We used two approaches to analyze orthologous protein sequences of 51 nu clear genes of vertebrates and 13 mitochondrial genes of mammals. In t he first approach, no assumptions of the distribution of the rate vari ation among sites were made, and in the second approach, the gamma dis tribution was assumed. Through both approaches, we found a negative co rrelation between the extent of amen g-site rate variation and the ave rage substitution rate of a protein. That is, slowly evolving proteins tend to have a high level of rate variation among sites, and vice ver sa. We found this observation consistent with a simple model of the ne utral theory where most sites are either invariable or neutral. We con clude that the correlation is a general feature of protein evolution a nd discuss its implications in statistical tests of positive Darwinian selection and molecular time estimation of deep divergences.