LUMPY: A probabilistic framework for structural variant discovery
Layer, Chiang, et al. Genome Biology
University of Virginia researchers describe an algorithm for finding new structural variants in whole-genome sequencing data. The breakpoint prediction approach, called LUMPY, integrates several types of structural variant information in read data — from split reads and read pairs in alignment data to read depth — in multiple sample datasets, the study's authors say. Their comparison suggests LUMPY compares favorably with other structural variant detection methods. In a previously sequenced human genome, for example, the method identified more than 4,500 new or known structural variant breakpoints.