Selection against archaic DNA in human regulatory regions
Although traces of archaic hominin DNA persist in all human populations outside Africa, these traces have been systematically depleted from the most functionally important regions of the human genome. This depletion suggests that many Neanderthal and Denisovan alleles had harmful effects on the fitness of hybrid individuals, but the nature of these harmful effects is poorly understood. Here, we show that Neanderthal and Denisovan alleles likely dysregulated gene expression in specific human tissues, causing systematic depletion of archaic introgression from enhancer regions annotated by the ENCODE RoadMap project. Highly pleiotropic enhancers that show activity in many tissues are more depleted of introgression than tissue-specific enhancers, and Neanderthal depletion is highly correlated with Denisovan depletion across sets of enhancers active in particular tissues. Fetal brain and fetal muscle are the tissues most depleted of archaic SNPs in their regulatory regions, and by analyzing the site frequency spectra of enhancers compared to control regions, we deduce that brain and muscle enhancers are likely depleted of introgression for different reasons. Brain enhancers appear to accumulate deleterious mutations unusually quickly, which may have caused inbred archaic populations to accumulate disproportionately large amounts of genetic load in these regions. In contrast, fetal muscle enhancers show no evidence of high deleterious mutation rates, and we hypothesize that their depletion of archaic DNA resulted from divergent selection for higher muscle mass in Neanderthals and Denisovans compared to humans.
