Determining the impact of uncharacterized inversions in the human genome by droplet digital PCR

Abstract

Despite the interest in characterizing genomic variation, the presence of large repeats at the breakpoints hinders the analysis of many structural variants. This is especially problematic for inversions, since there is typically no gain or loss of DNA. Here, we tested novel linkage-based droplet digital PCR (ddPCR) assays to study 20 inversions ranging from 3.1 to 742 kb flanked by inverted repeats (IRs) up to 134 kb long. Of those, we validated 13 inversions predicted by different genome-wide techniques. In addition, we obtained new experimental human population information across 95 African, European, and East Asian individuals for 16 inversions, including four already validated variants without high-throughput genotyping methods. Through comparison with previous data, independent replicates and both inversion breakpoints, we demonstrate that the technique is highly accurate and reproducible. Most studied inversions are widespread across continents, and their frequency is negatively correlated with genetic length. Moreover, all except two show clear signs of being recurrent, and we could better define the factors affecting recurrence levels and estimate the inversion rate across the genome. Finally, the generated genotypes have allowed us to check inversion functional effects, validating gene expression differences reported before for two inversions and finding new candidate associations. Therefore, the developed methodology makes it possible to screen these and other complex genomic variants quickly in a large number of samples for the first time, highlighting the importance of direct genotyping to assess their potential consequences and clinical implications.

Haplotype-resolved inversion landscape reveals hotspots of mutational recurrence associated with genomic disorders

“…We next focused on likely novel sites of inversion. Our integrated callset contains 100 putatively novel balanced inversions with none or less than 10% reciprocal overlap with events from the prior reports (Sudmant et al 2015;Sanders et al 2016;Audano et al 2019;Chaisson et al 2019;Giner-Delgado et al 2019;Puig et al 2020). These inversions span ~39 Mbp of the genome (Table S7), and five are >1 Mbp in size, including a ~23.2 Mbp pericentromeric inversion on chromosome 2 originating from a Mexican donor (NA19650), with its distal breakpoint lying near (~2 Mbp) the ancestral chromosome 2 fusion point specific to humans (Fig.…”

“…Previously, we and others reported that in the course of primate evolution regions corresponding to disease-associated microdeletions and microduplications at chromosomal regions 15q25,15q13.3,16p12.2,16p11.2,8p23.1,Xq22,17q21.31, as well as at the hemophilia A locus (Xq28) changed their orientation multiple times between a direct and inverted state (Lozier et al 2002;Zody et al 2008;Antonacci et al 2014;Catacchio et al 2018;Maggiolini et al 2019;Maggiolini et al 2020;Puig et al 2020). Based on the presence of inverted repeats with high identity at the boundaries of these regions, we hypothesized that non-allelic homologous recombination (NAHR) increases the probability of recurrent mutations during evolution, a phenomenon that we termed "inversion toggling" (Zody et al 2008).…”

Single-cell strand sequencing of a macaque genome reveals multiple nested inversions and breakpoint reuse during primate evolution

“…Conversely, 327 regions (87%), ranging in size from 859 bp to 9 Mbp and amounting to 55.6 Mbp of inverted DNA, were novel and described here for the first time as human/macaque inversions. However, 58 of these 327 regions were previously found to be inverted in Hominidae in a study in which Strand-seq was applied to discover inversions between humans and great apes (Porubsky et al 2020b), and 11 out 327 have been described to be polymorphic inversions in human (Antonacci et al 2009;Chaisson et al 2019;Giner-Delgado et al 2019;Puig et al 2020).…”

InvertypeR: Bayesian inversion genotyping with Strand-seq data

“…1 , Supplemental Figure S1, Additional file 1 ). This is suggestive of genotyping errors because de novo large inversions are thought to arise very infrequently (10 − 5 –10 − 4 /generation [ 18 , 19 ]). We reasoned that two innovations might enable more accurate inversion genotyping.…”