Computational Biology Colloquium | Kelley Harris, Ph.D.

Kelley Harris, Ph.D.
Assistant Professor, University of Washington, Department of Genome Sciences
Lab Website

Using archaic introgression to probe enhancer function and evolution

Thursday, October 25, 2018
2 PM
RRI 101

Abstract: After Neanderthals interbred with humans, their DNA was selectively purged from many functional regions of the genome. Protein coding regions, for example, tend to have less Neanderthal DNA than regions that evolve more neutrally, and this imbalance implies that Neanderthal genes put hybrids at a fitness disadvantage compared to contemporaries without hybrid ancestry. As such, the landscape of Neanderthal introgression provides a snapshot of selection that occurred about 50,000 years ago as a result of the differences that arose over 300,000 years of Neanderthal/human divergence. To deduce how much of this selection may have been driven by gene regulation rather than gene sequence, we examine how high-confidence Neanderthal variants annotated by Sankararaman, et al. are distributed in enhancer regions annotated by the ENCODE project, calculating the binomial enrichment/depletion of Neanderthal variants in enhancers compared to a set of control variants matched for allele frequency and the local intensity of background selection. We find that enhancers appear to have tolerated Neanderthal introgression slightly better than genes or promoters have, but are still significantly depleted of introgression more than expected based on their relatively modest level of sequence conservation. We find that highly pleiotropic enhancers active in many tissue types are the most intolerant to introgression, whereas tissue-specific enhancers are relatively tolerant to Neanderthal DNA, perhaps because changes to the sequences of tissue-specific enhancers are easier to buffer and therefore less damaging to regulatory function. Every cell type profiled by ENCODE appears somewhat intolerant to the presence of Neanderthal DNA in its complement of enhancers, but the magnitude of this intolerance varies by tissue type. Introgression depletion is correlated across tissues with skew in the site frequency spectrum toward rare variants, but this correlation is imperfect, indicating that the introgression landscape provides additional information about genomic functionality that cannot be gleaned from conservation information alone. Fetal enhancers, in particular, seem to tolerate new mutations better than brain enhancers do, but of the two categories, fetal enhancers are less tolerant of introgression. Fetal enhancers that are classified as human accelerated regions (HARs) show the strongest introgression intolerance of all, suggesting that enhancers in this pathway have experienced more functionally significant changes since the human/Neanderthal split than other pathways that exhibit more enhancer sequence conservation overall.

Host:  Mark Chaisson