NEW YORK – A team from Germany, the US, and Italy has tallied recurrent inversions across dozens of human genomes, providing a look at haplotypes containing these polymorphisms, mechanisms that lead to new inversions, and inversion ties to genetic instability and related microdeletion or duplication conditions. The findings appeared in Cell on Friday.
"We found several haplotypes that are likely to confer risk for particular genomic disorders, and this knowledge is not only relevant for understanding disease mechanisms but could also be used to identify families at risk for genomic disorders," co-senior and co-corresponding author Jan Korbel, a genome biology researcher affiliated with the European Molecular Biology Laboratory and EMBL's European Bioinformatics Institute, said in an email.
Using Pacific Biosciences long-read sequencing, Bionano Genomics optical mapping, and Strand-seq, the researchers profiled genome sequences from 41 unrelated individuals, identifying more than 700 inversions — a set that included 330 inversions smaller than around 2 kilobases and 292 larger inversions that reached the megabase size.
Each of these inversion classes arose through distinct processes, Korbel explained. While some 85 percent of inversions in the smaller set appeared to reflect twin priming events related to L1 retrotransposition activity, for example, many of the larger inversions were balanced events, often stemming from non-allelic homologous recombination events mediated by large segmental duplication repeats.
"These inversions have properties that distinguish them from most classes of genetic variation — in particular, a high abundance of common polymorphisms," Korbel said, further noting that as a consequence, an unexpectedly large number of humans are heterozygous for these inversions.
From there, the team turned to two new computational methods to find and quantify 40 recurrent inversions, leading to sequence "toggling" at sites across some 0.6 percent of the human genome. Along with an overrepresentation for recurrent inversion sites on the X or Y sex chromosomes, the analyses pointed to overlap between inversion-prone sites and genomic conditions such as a chromosome 3q29 microdeletion syndrome.
"Several prior studies described inversions in genomic regions prone to undergo pathogenic microdeletions. … These inversions have been proposed to mediate [formation of microdeletions], but for most loci in which inversions were reported this relationship remained unclear because of difficulties in inversion discovery and breakpoint definition," Korbel explained. "Our study now shows that recurrent inversions — rather than inversions forming once in human history — drive the co-localization with genomic disorder regions."
In addition to analyses aimed at untangling segmental duplication- or retrotransposon-based mechanisms for inversion formation, the team dug into inversion ties to disease-related changes in the genome, particularly microdeletions.
From their results, the researchers speculated that microdeletions may turn up at recurrent inversion sites through homologous recombination suppression in heterozygous individuals carrying one copy of a given inversion, Korbel explained. They suspect that recurrent inversions may also alter the structure of flanking regions, he added, which may be another, non-mutually exclusive mechanism for microdeletions to form.
"These structural differences could make these genomic regions more genetically instable, with certain structural configurations at the flanks of the inversions either predisposing or protecting individuals from recurrent microdeletions," he said.
Based on these and other findings, the authors pointed to the "vast potential of inversions to prime or protect against morbid CNV formation, thereby shaping the human landscape of repeat-mediated mutation that is yet to be fully explored."
The authors cautioned that the current analysis "is limited to 82 unrelated human haplotypes and more genomes will be required to distinguish single-origin inversions from recurrent events" and noted that "[c]haracterization of the complete spectrum of [structural variants] at the sequence level remains an important goal that is likely to be unattainable outside of a multi-platform approach."