Resolving missing human polymorphic inversions and other complex variants from ultra-long read data

Identification of inversion alleles from ultra-long Oxford Nanopore reads and adaptive sampling, in regions flanked by inverted repeats.

Inversions are balanced structural variants that escape most analyses because of the large repeats that flank them. The authors assembled a catalogue of 612 candidate inversions ranging from 197 bp to 4.4 Mb, flanked by inverted repeats shorter than 190 kb, and developed a bioinformatic package dedicated to reliably calling inversion alleles from long-read data. Combining several DNA extraction, library preparation and Oxford Nanopore sequencing protocols, they show that ultra-long reads (50-100 kb) together with adaptive sampling are an efficient way to detect most human inversions. Across 54 diverse individuals, 87-99% of inversions could be genotyped per sample, depending mainly on read length, inverted-repeat length and genome coverage. Both orientations were observed for 155 regions (frequency 0.01-0.49), tripling the number of polymorphic inverted-repeat-mediated inversions studied in detail so far, and more than 300 additional independent structural variants were found in these regions.

Synthesis written by Geno'X. For the full original abstract, please refer to the source publication.

These polymorphic inversions and balanced rearrangements are exactly the kind of events that conventional karyotyping could see, at least the larger ones, and that the move to short-read sequencing pushed out of view: a balanced rearrangement does not change read depth and gets lost in the repeats that flank it. Ultra-long reads bring them back into molecular analysis, with a base-pair resolution that conventional cytogenetics never had. The immediate value for a laboratory is a reference set of allele frequencies: without it, an inversion found in a patient remains uninterpretable, because there is no way to tell whether it is common in the population.

Analysis by Dr Thibaut Benquey