Detection of genome structural variation in normal cells and tissues by single molecule sequencing

Abstract

Rare genome structural-variants (SVs) in somatic cells and tissues are difficult to measure in bulk DNA because library preparation can create chimeric junctions that mimic authentic events. Based on our previous Single-Molecule Mutation sequencing (SMM-seq) assay, we developed a new method, SMM-SV-seq, which combines a chimera-suppressing library preparation with duplex confirmation of breakpoint-supporting molecules. Using libraries from mixtures of human and Drosophila DNA, we show a remaining artifact rate of 1.52 +/- 0.53 detectable cross-species chimeras per billion bases, providing an empirical background for correcting breakpoint frequencies. In a benchmark of evaluable homozygous germline deletions in the GM24385 cell line, SMM-SV-seq achieved 81.4% precision and 85.7% recall, exceeding the precision of established short-read callers while maintaining comparable recall. The assay detected dose-associated increases in SV-breakpoint frequencies in cells treated with bleomycin, a known clastogen. Biological relevance was demonstrated by a significantly greater irradiation-associated increase in translocation-like breakpoint frequencies in the BRCA1 185delAG-heterozygous mammary epithelial clone than in its isogenic wild-type control. These results establish SMM-SV-seq as a practical assay for calibrated, duplex-confirmed detection of rare somatic SV breakpoints directly in bulk DNA from non-clonal cell populations.