Fig. 2

RD signal computed from entire-fragment counting approach recapitulates WGS-derived coverage signal. a The coverage plot (left) at 5-kb bin and the RD frequency distribution (right) of MCF7 chr16:56-88 Mb region computed from Hi-C data using exact-cut (top), midpoint (middle), and HiCNAtra’s entire-fragment (bottom) counting approaches. Each red dot represents the copy number per bin. Focal alterations (e.g. FA1 and FA2) and LCVs are shown. b Box plot of centralization scores (5-kb bin) and zero scores (5-kb bin) computed from the Hi-C-derived RD signal of 6 cell lines (MCF7, LNCaP, PC3, GM12878, IMR90, and PrEC) using exact-cut, midpoint, and entire-fragment counting approaches at 5-kb and 100-kb bin sizes. c HiCNAtra gain (in percentage) of centralization and zero scores by utilizing the entire fragment counting approach, compared to exact-cut and midpoint approaches. d The genome-wide coverage plot (chr 1–chr X) of MCF7 breast cancer cell line is computed from Hi-C (top) and WGS (bottom) data. Each dot represents the RD signal per bin computed from Hi-C data (red) or WGS data (grey). WGS data is obtained from the ‘input’ (control) data of MCF7 ChIP-seq experiment