Extended Data Fig. 3: Complex SV detection based on Hi-C maps.
a, Illustration of how we re-construct Hi-C map surrounding breakpoints. There are four orientation types of inter-chromosomal translocations. b, Overall workflow of the complex SV assembling module in NeoLoopFinder. c, In the first step of the pipeline, we determine the whole rearranged fragments (green box) of the input SV breakpoints within the checking window (gray box, by default 5Mb extended from the breakpoints). d, The algorithm for determining rearranged fragments. First, correlation matrices are calculated by rows (top) and columns (bottom) of the contact matrix separately within the checking window; then the rearranged fragment boundaries are determined by checking the first principal component profile (PC1) of the correlation matrices. e,f, Determination of the R2 cutoff for SV filtering. e, The distribution of R2 across all large SVs (intra-chromosomal rearrangements larger than 1Mb and inter-chromosomal translocations). The number was summarized from all 50 cancer samples. f, The fraction of SVs after filtering as a function of R2 cutoff. Data were merged from eight cancer cell lines: A549, Caki2, K562, LNCaP, NCI-H460, PANC-1, SK-N-MC, and T47D.
