Network centrality is the most important variable determining gene sequence evolution rates in Schizosaccharomyces pombe

Abstract

Which variables determine the rate of gene and protein sequence evolution is a central question in evolutionary genomics. In the model organism fission yeast (Schizosaccharomyces pombe), the determinants of the rate of sequence evolution have yet to be determined. Previous studies in other organisms have typically found gene expression levels to be most significant, with numerous other variables identified as having a smaller impact. Here, partial least squares regression (PLS) and partial correlation analysis are used to model sequence evolution rates in the fission yeast genome by a range of variables. Variable importance in projection (VIP) scores as well as partial correlation coefficients are calculated for each variable, and used as estimates of the influence of each independent variable on sequence evolution rate. Unlike many previous studies in other organisms, centrality in the PPI network is shown to be the most important variable, and gene expression found to be less influential. Considerable heterogeneity is found in the influence of different gene ontology terms as well as amino acid composition. However, the majority of variance in constraint in fission yeast remains unexplained by this study, indicating that variables not yet considered and stochastics probably have considerable impact on the rate of molecular evolution.