Fig. 1

Use of chimeric and split reads to detect structural variation. Structural variation in the sample is depicted in box 1a as a fusion between genomic regions A (green) and B (blue). Sequence differences in the sample come from structural variation, repetitive sequence (orange), and base substitutions due to sequencing errors and SNPs (black). In box 1b, each group of partially aligned reads, or “stack,” corresponds to a candidate breakpoint located at shared end (left: orange, black, and blue; right: green, yellow, and black). Pairwise combinations of breakpoints form a library of candidate junctions (box 1c). All stacked reads are aligned to the library and are used to assess their support for the candidate junctions. A read aligned to a candidate provides support equal to the product of the length of the “tail” and total alignment quality. The total support for each candidate junction (box 1d) is the sum of supports from the stacked reads aligned to it