Presenter: Frank V. Mannino
Advisor(s): Spencer V Muse
Author(s): Frank V Mannino and Spencer V Muse
Graduate Program: Bioinformatics

Title: Extensive site-to-site rate variability of synonymous substitution rates

Abstract: We investigate the presence of site-to-site variability of synonymous substitution rates using a recently developed model of codon evolution. In an analysis of all 13 protein coding genes from the completed mitochondrial genomes of a total of 111 taxa sampled from ten animal clades, we find extensive variability of synonymous rates in the majority of genes (77% of the analyzed data sets showed statistically significant levels). Furthermore, the magnitude of synonymous rate variation as measured by the coefficient of variation is comparable to that of nonsynonymous rates, including three data sets where there is more synonymous rate heterogeneity than nonsynonymous heterogeneity. These findings raise interesting questions about underlying mechanisms of selection at silent sites, and they have important implications for analyses--- especially studies of adaptive evolution--- that are based on the premise that silent rates are equal across all sites in a gene. In addition, we test for the significance of correlation between synonymous and nonsynonymous rates using a new bivariate distribution of rate variation. Our results demonstrate significant rate correlation in many of the data sets.