Fig. 2

Converting copy number segmentation profiles into copy number mutational vectors. The CNV classification schema converts a sample’s segmentation profile (a, c, e) into a count vector of 48 mutually exclusive components (b, d, f). These components are based on segment size, heterozygosity status, and total copy number. A breast cancer sample with many highly amplified segments, possibly due to the presence of extrachromosomal DNA, is shown in (a, b). This sample’s count vector is characterized by peaks in the 5–8 and 9+ total copy number categories. A gastric cancer sample with extensive loss of heterozygosity is shown in (c, d). This sample’s count vector is characterized by peaks in the LOH category, specifically with a total copy number of 1 indicating a loss of an allele. A sarcoma sample with a whole-genome duplication event, characterized by peaks in the 3–4 total copy number category and the 40 + Mb size bin, is shown in (e, f)