Flexible and rapid validation of structural variation using adaptive sampling

Identification of genomic rearrangements by microarrays or short-read sequencing frequently lacks information about the exact architecture and breakpoints of variants due to technical limitations. Independent verification of complex structural variants (SVs) is often performed using custom targeted assays, making confirmation of clinically relevant findings time consuming and laborious. In this study we evaluate Oxford Nanopore long-read adaptive sampling for flexible and rapid confirmation and characterization of complex genomic rearrangements and structural variants. Adaptive sampling is an in silico target enrichment, where continued sequencing or ejection of a fragment is based on whether it matches a defined reference sequence. Using adaptive sampling, we targeted 10 regions with different structural variant types, including deletions, translocations, and complex rearrangements. Each sample was analyzed on a MinION or PromethION flow-cell, and sequencing resulted in between 14.1-18.3 Gb of data per sample, with mean autosomal on-target coverage of 28.4x and off-target read depth coverage of 5.3x. We were able to verify all 10 rearrangements, with breakpoint spanning reads for nine of the ten regions, and fully resolved the architecture of nine regions. We also show that background reads can be used to detect structural variants in non-targeted regions of the genome. Our results show that adaptive sampling represents a flexible and rapid strategy for confirmation and characterization of clinically relevant genomic rearrangements in clinical samples. By providing sequence information, read depth, and methylation data, nanopore adaptive sampling has advantages over other assays for variant confirmation used in diagnostic laboratories today.