Genome-wide signatures of selection and convergent evolution in corvids
Final Report Abstract
Evolutionary change is driven by a combination of neutral and selective processes. Selection is the prevailing evolutionary force allowing organisms to adapt and seize novel opportunities (positive selection), but likewise to maintain systemic functionality (purifying selection). As selection efficacy strongly depends on the effective population size (Ne), its interplay with genetic drift needs to be considered. Specifically, genetic drift is expected to predominate and selection efficacy to be reduced in small Ne populations, leading to the accumulation of slightly deleterious sites. Although this increases the chances to accumulate novel genetic variants that turn out beneficial under certain environmental conditions, the fixation of slightly deleterious mutations is expected to reduce fitness and increase the risk of extinction. However, owing to a historical lack of genome-wide data, our knowledge about the interplay of selection and drift acting on the genomes of taxa across the animal kingdom is limited. Comparative genomic approaches are powerful to address this question. For this project, I had access to population genomic whole genome re-sequencing data for several species from the avian genus Corvus (crows, rooks, and jackdaws). While the species under investigation share similar life histories, they differ markedly in geographic distribution range sizes, allowing me to study the effect of Ne on the relative strengths of selection and genetic drift. I focused on estimates of the distribution of fitness effects (DFE) of new deleterious mutations that summarizes the proportions of slightly deleterious, deleterious and strongly deleterious mutations in the genome. The ratio of nucleotide diversity at nonsynonymous and synonymous sites, πN/πS, is a related measure for selection efficacy. Both can be estimated from population genetic data, summarized as the site frequency spectra (SFS) of a class of sites under selection (e.g. zero-degenerate sites) and a neutrally evolving site class (e.g. four-fold degenerate sites). Obstacles during the project were limited to struggles with bioinformatics software and pipelines that caused some delay but were solved by consulting colleagues at the host department at Uppsala university and the developers of the respective software causing problems. Further, biases from sampling on DFE estimates are found in analyses of an extensive dataset of 118 C. (corone) spp. individuals, but with minor consequences for interpretations of comparisons among species. In comparisons of genome-wide estimates of the DFE and πN/πS between Corvus species inhabiting islands and widespread mainland species, I find larger proportions of slightly deleterious sites and higher πN/πS values in island species. This is consistent with a stronger effect of genetic drift and reduced selection efficacy in small Ne populations as expected from evolutionary theory. This project highlights the value of whole genome polymorphism data from several closely related species in a comparative framework to address central questions arising from population genetic theory.
