Genomic Patterns of Gene Evolution.

Abstract

The bounty of genomic sequence and supporting genomic datasets such as expression data enables study of the genomic patterns of gene evolution. Such studies can address long standing questions in the field of molecular evolution, such as the molecular basis of phenotypic change, the relative contributions of selection and drift, and the origin of new genes. Here, I first examine a genomic dataset of nervous system genes and find, contrary to previous reports, that brain genes as a group have not experienced accelerated evolution in the human lineage. From this finding I infer that widespread changes in protein coding sequence are not responsible for the unique features of the human brain. I also compare the prevalence of positively selected genes in human and chimpanzee on a genome-wide scale. After careful control for the differences in genome sequence quality between human and chimpanzee, I find that, consistent with the predictions of neutral theory given the smaller effective population size in humans compared to chimpanzees, humans have fewer positively selected genes and have experienced less purifying selection as well. Finally, I use simulation based on genomic evolutionary rate patterns in Drosophila to examine phylostratigraphy, a method to infer gene age. I find that the method substantially overestimates the age of young genes and underestimates the age of old genes, especially those that evolve rapidly. Furthermore, spurious correlations can result simply from bias in the measurement of gene age. Taken together, these studies suggest that genomic data can yield previously unattainable insights about the history and process of evolution, but that genomic results must be critically evaluated to ensure they reflect true biology, rather than artifacts of data or method.