In whole genome sequence data, large deletions are readily detected by the same methods as moderately sized deletions, that is, paired-end methods such as BreakDancer or split-read methods such as Pindel (25,33). Large deletions are easier to detect than smaller indels using paired-end methods, as they are easily distinguished from normal variation in the insert size. Large duplications are more difficult to detect, as there is no single read or read pair spanning the insertion. One software tool capable of detecting large insertions is Pindel, which uses a pattern-growth approach to detect breakpoints, at which the sequenced genome diverges from the reference. If two such breakpoints occur in the same chromosome with an appropriate orientation, Pindel can piece them together to discover a large insertion event.
