Fine-scale characterization of genomic structural variation in the human genome reveals adaptive and biomedically relevant hotspots

Abstract

Genomic structural variants (SVs) are distributed nonrandomly across the human genome. The "hotspots" of SVs have been implicated in evolutionary innovations, as well as medical conditions. However, the evolutionary and biomedical features of these hotspots remain incompletely understood. Here, we analyzed data from 2,504 genomes to construct a refined map of 1,148 SV hotspots in human genomes.We confirmed that segmental duplication-related nonallelic homologous recombination is an important mechanistic driver of SV hotspot formation. However, to our surprise, we also found that a majority of SVs in hotspots do not formthroughsuchrecombination-basedmechanisms,suggestingdiversemechanistic andselective forces shaping hotspots. Indeed, our evolutionary analyses showed that themajority of SV hotspots arewithin gene-poor regions and evolve under relaxed negative selection or neutrality.However,we still found a small subset of SV hotspots harboring genes that are enriched for anthropologically crucial functions and evolve under geography-specific and balancing adaptive forces. These include two independent hotspots on different chromosomes affecting alpha and beta hemoglobin gene clusters. Biomedically, we found that the SV hotspots coincide with breakpoints of clinically relevant, large de novo SVs, significantlymore often than genome-wide expectations. For example,we showed that the breakpoints ofmultiple large SVs,which lead to idiopathic short stature, coincidewith SV hotspots. Therefore, the mutational instability in SV hotpots likely enables chromosomal breaks that lead to pathogenic structural variation formations. Overall, our study contributes to a better understanding of the mutational and adaptive landscape of the genome.