Chalmers Conferences, 9th European Conference on Mathematical and Theoretical Biology

Modeling the impact of selection in protein coding genes
Carina Farah Mugal, Ingemar Kaj

Last modified: 2014-04-01


By combining methods for diffusion approximation of the Wright-Fisher reproduction mechanism with Poisson random measure tools, we construct a stochastic model for the evolution of protein coding sequences under positive or negative natural selection. The model incorporates the genetic code in the sense that non-synonymous mutations, those mutations which represent replacements of amino acids during the protein generating process, evolve under selection whereas synonymous mutations evolve neutrally. In this framework we investigate the ratio of non-synonymous to synonymous divergence between gene sequences, by decomposing the contributions due to fixations, lineage-specific polymorphisms and ancestral polymorphisms. With these findings we are able to assess properties of the dN/dS ratio and of the neutrality index. These are measures of wide-spread use in evolutionary genetics studies designed to help accept or reject the presence of selective forces.