Detailed analysis of inversions predicted between two human genomes: errors, real polymorphisms, and their origin and population distribution

Abstract

The growing catalogue of structural variants in humans often overlooks inversions as one of the most difficult types of variation to study, even though they affect phenotypic traits in diverse organisms. Here, we have analysed in detail 90 inversions predicted from the comparison of two independently assembled human genomes: the reference genome (NCBI36/HG18) and HuRef. Surprisingly, we found that two thirds of these predictions (62) represent errors either in assembly comparison or in one of the assemblies, including 27 misassembled regions in HG18. Next, we validated 22 of the remaining 28 potential polymorphic inversions using different PCR techniques and characterized their breakpoints and ancestral state. In addition, we determined experimentally the derived allele frequency in Europeans for 17 inversions (DAF = 0.01-0.80), as well as the distribution in 14 worldwide populations for 12 of them based on the 1000 Genomes Project data. Among the validated inversions, nine have inverted repeats (IRs) at their breakpoints, and two show nucleotide variation patterns consistent with a recurrent origin. Conversely, inversions without IRs have a unique origin and almost all of them show deletions or insertions at the breakpoints in the derived allele mediated by microhomology sequences, which highlights the importance of mechanisms like FoSTeS/MMBIR in the generation of complex rearrangements in the human genome. Finally, we found several inversions located within genes and at least one candidate to be positively selected in Africa. Thus, our study emphasizes the importance of careful analysis and validation of large-scale genomic predictions to extract reliable biological conclusions.

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

“…To test the ddPCR application for inversion genotyping, we analyzed a representative sample of 20 inversions mediated by IRs of different sizes from the InvFEST database . Due to the high error rate of inversion detection in repetitive sequences (Vicente-Salvador et al 2017;Chaisson et al 2019;Giner-Delgado et al 2019), we 6 selected well-supported inversions, excluding predictions on very complex regions full of segmental duplications (SDs) or with genome assembly gaps. In particular, 14 were initially predicted from fosmid PEM data in nine individuals (Kidd et al 2008), although five had additional supporting evidence (Supplemental Table S1).…”

“…Moreover, the great majority of inversions mediated by IRs have been shown to occur recurrently several times both within the human lineage and in non-human primates (Aguado et al 2014;Vicente-Salvador et al 2017;Giner-Delgado et al 2019;Antonacci et al 2009).…”

Functional and evolutionary impact of polymorphic inversions in the human genome

“…Y), with sizes ranging from 83 bp to 415 kb (median of 4.1 kb). In addition, 24 (53%) have been generated by non-allelic homologous recombination (NAHR) between >90% identical IRs (from 654 bp to 24.2 kb, with a median of 5.9 kb), whereas the rest (47%) were probably generated by non-homologous mechanisms (NH), including three with clean breakpoints and 18 with small deletions or insertions in the derived allele that may have been created in a single complex FoSTeS/MMBIR event (Sudmant et al, 2015;Vicente-Salvador et al, 2017). Three of those (HsInv0031, HsInv0045 and HsInv0098) have also short low-identity IRs (249-297 bp, 83.2-86.2% identity) in the ancestral orientation that are partially deleted in the derived orientation ( Figure S1).…”

“…The fact that they are balanced changes, together with the highly-identical inverted repeats (IRs) often found at their breakpoints, makes inversion detection very difficult, even with the newest sequencing methods based on short or long reads (Huddleston et al, 2017;Lucas-Lledó and Cáceres, 2013;Vicente-Salvador et al, 2017). Other techniques, like BioNano optical maps (Li et al, 2017) or Strandseq (Sanders et al, 2016) promise to help in inversion discovery, but they are not suitable for highthroughput genotyping.…”

“…Also, although inversion genotypes might be predicted based on SNP data (Cáceres and González, 2015;Cáceres et al, 2012;Ma and Amos, 2012;Salm et al, 2012), these methods can only detect certain types of inversions and the error rate can be high. Currently, a small number of inversions have been genotyped by FISH (Antonacci et al, 2009;Salm et al, 2012) and regular or inverse PCR (iPCR) (Aguado et al, 2014;Pang et al, 2013;Vicente-Salvador et al, 2017), which are labour intensive and are limited by the characteristics of the inversion IRs. Therefore, although inversions affect a significant fraction of the human genome , only a few have been characterized in detail (Aguado et al, 2014;Antonacci et al, 2009;Pang et al, 2013;Puig et al, 2015b;Salm et al, 2012;Stefansson et al, 2005;Vicente-Salvador et al, 2017) and very little is known about the global frequency and distribution of inversions in human populations.…”