Patterns of ancient selection in modern humans around candidate sites

Fernando Racimo 1 ,2, Martin Kuhlwilm2, Montgomery Slatkin1
1University of California - Berkeley, Berkeley, CA, USA, 2Max Planck Institute for Evolutionary Anthropology, Leipzig, Germany


Though the recent sequencing of the high-coverage Denisovan and Neanderthal genomes has allowed us to find the genetic differences that set modern humans apart from archaic humans, the subset of such changes that rose to fixation due to selection is currently unknown. In this study, we look for patterns of positive selection on the modern human lineage at various classes of putatively functional changes using diversity scaled by divergence, as has been done previously on the human lineage since the split from chimpanzees. We also develop an approximate Bayesian computation (ABC) approach incorporating various statistics aimed at identifying ancient patterns consistent with selection around a candidate site. We fail to find an enrichment for signals of positive selection around nonsynymous changes relative to synonymous changes. It has been argued that the failure to detect this difference in changes on the human lineage may be due to varying levels of background selection which occlude the signal of positive selection. Indeed, when we control for the intensity of background selection (BS), we observe a significant difference between nonsynonymous changes in regions of low BS and matching regions of the genome, lending support to this hypothesis. We also identify a slight enrichment for positive selection at splice site changes.