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. These "hotspots" have been implicated in critical evolutionary innovations, as well as serious medical conditions. However, the evolutionary and biomedical features of these hotspots remain incompletely understood. In this study, we analyzed data from 2,504 genomes from the 1000 Genomes Project Consortium and constructed a refined map of 1,148 SV hotspots in human genomes. By studying the genomic architecture of these hotspots, we found that both nonallelic homologous recombination and non-homologous mechanisms act as mechanistic drivers of SV formation. We found that the majority of SV hotspots are within gene-poor regions and evolve under relaxed negative selection or neutrality. However, we found that a small subset of SV hotspots harbor genes that are enriched for anthropologically crucial functions, including blood oxygen transport, olfaction, synapse assembly, and antigen binding. We provide evidence that balancing selection may have maintained these SV hotspots, which 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, significantly more often than genome-wide expectations. As an example, we showed that the breakpoints of multiple large de novo SVs, which lead to idiopathic short stature, coincide with SV hotspots. As such, the mutational instability in SV hotpots likely enables chromosomal breaks that lead to pathogenic structural variation formations. Our study contributes to a better understanding of the mutational landscape of the genome and implicates both mechanistic and adaptive forces in the formation and maintenance of SV hotspots.

Genetic-variant hotspots and hotspot clusters in the human genome facilitating adaptation while increasing instability

“…In general, the level of positive selection can be assessed based on (i) the allele-frequency spectrum, Table S3B), significantly more than the simulated regions (p = 0.0001 for both hotspots and clusters, 10,000 simulations). These three different It has been proposed that balancing selection represents as an essential adaptive force acting on structural variation hotspots harbouring genes involved in immune functions or anthropologically crucial functions [24,25,35]. This is confirmed by the present findings that 4.7% of the GV hotspots and 6.0% of the hotspot clusters overlap with the hotspots of balancing selection in the human genome (Supplementary Table S3B).…”

“…It is noteworthy that nucleotide differences between highly similar sequences (e.g. SDPs) located at different genomic positions may be falsely identified as SNPs due to mismapping of short-reads [25],…”

“…Previous studies have indicated that genetic variants are not evenly distributed in the human genome [21,22], with density enhancements giving rise to genetic-variant hotspots in some genomic regions, and identified homologous recombination as one of the major mechanisms for the formation of GV hotspots [23][24][25]. Since transposable elements constitute highly abundant repeat sequences in the genome, they can serve as homologous templates in recombination events that produce GVs [26][27][28].…”

Clinical Interpretation and Management of Genetic Variants

“…Accurate identification of SVs in the genome has been difficult in part because of the shortcomings of short-read sequencing and the reference genome, as discussed earlier (reviewed in Eichler [ 21 ]). The existing data suggest the presence of several thousand SVs in each genome, which are nonrandomly distributed across the genome ( 37 ). The majority of the SVs are smaller than then 10 kilobases, whereas a few could encompass several million nucleotides; collectively, SVs involve more nucleotides than SNVs ( 38 ).…”

Archaic hominin introgression into modern human genomes

“…These comparisons are conducted in segments of the DNA where the sequence alignment and thus variation calling is robust and have low false positive rates. A good portion of the genome is terra incognita for such short‐read sequence‐based variation analyses due to complications stemming from simple repeats, segmental duplications, structural variations, retrotranspositions, and assembly errors, etc (Chaisson et al, ; Kronenberg et al, ; Lin & Gokcumen, ; Miller et al, ; Sedlazeck, Lee, Darby, & Schatz, ). It is likely that this limitation in our approach may hinder our ability to investigate important variation between and within species that may explain some of the phenotypic variations.…”