Yury Barbitoff*, Polina Malysheva, Polina Bogaichuk, Nadezhda Pavlova and Alexander Predeus
Institute of Bioinformatics Research and Education, Belgrade, Serbia
barbitoff [at] bioinf.institute
Abstract
Natural selection is arguably the most important driving force of microevolution which has a significant impact on the patterns of genetic variation across the genome. In loci with important cellular and organismal functions, negative selection acts to eliminate deleterious derived alleles, leading to evolutionary conservation of sequences. At the same time, positive selection enhances the transmission and fixation of adaptive alleles, leaving a characteristic footprint called a “selective sweep”. While many methods have been developed to detect signals of both positive and negative selection, little is known about the interplay of these two forces. In this work, we tried to comprehensively identify loci bearing combinations of negative and positive selection signals across the human genome. To this end, we integrated a wide array of metrics, including; (i) alignment-based measures of evolutionary conservation across vertebrate genomes (phastCons, phyloP, GERP++); (ii) selective constraint estimates obtained from standing genetic variation in human populations (pLI, LOEUF, z-scores for observed-to-expected missense variant ratios); and (iii) measures of recent positive selection covering different evolutionary timescales (iHS, DRC150, SDS). In total, we identified a set of 163 genes bearing a combined selection signal (CSS) according to at least one negative and positive selection metric. Intriguingly, we found that these genes have a broader range of molecular functions despite having a relatively restricted expression profile. Overrepresentation analysis showed a strong enrichment of genes involved in nervous system function, including synaptic transmission and neurogenesis. Of note, CSS genes were also overrepresented among highly pleiotropic genes affecting multiple independent traits (across a collection of over 1,000 complex and Mendelian traits in humans and mice). Taken together, our results emphasize the complex network of connections between natural selection, gene function, and phenotype, and indicate that mutations in genes controlling high-level organismal functions might have stronger fitness effects.
Keywords: natural selection, pleiotropy, complex trait
Acknowledgement: We thank JetBrains Ltd. for providing computational resources for the project.