How Much Information is Provided by Human Epigenomic Data? An Evolutionary View

Here, we ask the question, \"How much information do available epigenomic data sets provide about human genomic function, individually or in combination?\" We consider nine epigenomic and annotation features across 115 cell types and measure genomic function by using signatures of natural selection as a proxy. We measure information as the reduction in entropy under a probabilistic evolutionary model that describes genetic variation across [~]50 diverse humans and several nonhuman primates. We find that several genomic features yield more information in combination than they do individually, with DNase-seq displaying particularly strong synergy. Most of the entropy in human genetic variation, by far, reflects mutation and neutral drift; the genome-wide reduction in entropy due to selection is equivalent to only a small fraction of the storage requirements of a single human genome. Based on this framework, we produce cell-type-specific maps of the probability that a mutation at each nucleotide will have fitness consequences (FitCons scores). These scores are predictive of known functional elements and disease-associated variants, they reveal relationships among cell types, and they suggest that [~]8% of nucleotide sites are constrained by natural selection.