Abundant Raw Material for Cis-Regulatory Evolution in Humans

TLDR

Gene expression changes driven by cis‑regulatory DNA evolution underlie many phenotypic shifts, yet the population distribution of such variation, including microsatellite‑mediated effects, remains poorly characterized. This study surveys experimentally validated functional cis‑regulatory polymorphisms for the first time. The data comprise over 140 polymorphisms affecting 107 genes in humans, the species with the most available regulatory data. Functional cis‑regulatory variation is pervasive, with 63 % of surveyed genes showing at least two‑fold expression differences, humans heterozygous at >16,000 such sites versus <13,000 amino‑acid sites, and widespread polymorphic transcription‑factor binding that creates epistasis, GxE interactions, pleiotropy, and overdominance.

Abstract

Changes in gene expression and regulation—due in particular to the evolution of cis-regulatory DNA sequences—may underlie many evolutionary changes in phenotypes, yet little is known about the distribution of such variation in populations. We present in this study the first survey of experimentally validated functional cis-regulatory polymorphism. These data are derived from more than 140 polymorphisms involved in the regulation of 107 genes in Homo sapiens, the eukaryote species with the most available data. We find that functional cis-regulatory variation is widespread in the human genome and that the consequent variation in gene expression is twofold or greater for 63% of the genes surveyed. Transcription factor-DNA interactions are highly polymorphic, and regulatory interactions have been gained and lost within human populations. On average, humans are heterozygous at more functional cis-regulatory sites (>16,000) than at amino acid positions (<13,000), in part because of an overrepresentation among the former in multiallelic tandem repeat variation, especially (AC)n dinucleotide microsatellites. The role of microsatellites in gene expression variation may provide a larger store of heritable phenotypic variation, and a more rapid mutational input of such variation, than has been realized. Finally, we outline the distinctive consequences of cis-regulatory variation for the genotype-phenotype relationship, including ubiquitous epistasis and genotype-by-environment interactions, as well as underappreciated modes of pleiotropy and overdominance. Ordinary small-scale mutations contribute to pervasive variation in transcription rates and consequently to patterns of human phenotypic variation.