Fig. 2

Genes with altered expression associated with nearby SV breakpoints across 327 cancer cell lines. a For each of the indicated genomic region windows examined, numbers of significant genes (FDR < 10%) showing a correlation between expression and associated SV event across 327 cancer cell lines with WGS and expression data [10]. Numbers above and below the zero point of the y-axis denote positively and negatively correlated genes, respectively. Linear regression models evaluated significant associations when correcting for cancer type (gray) and for both cancer type and gene-level CNA (black). For the 1 Mb region window, the model weights the relative gene distances of the breakpoints [2]. b Heat map of significance patterns for 1249 genes significant for any region window (FDR < 10%, correcting for both cancer type and CNA). Red denotes significant positive correlation; blue, significant negative correlation. Genes listed are cancer-associated [23]. c Significance of genes in cancer cell lines, as plotted (Y-axis) versus the number of cell lines impacted (expression > 0.4SD from sample median) by nearby SV breakpoint (within 1 Mb). d Significance of genes in combined PCAWG-TCGA cohort (2334 patients, x-axis) [2], as compared to their significance in the cancer cell line cohort (327 cell lines, y-axis) [10]. Genes in the upper left quadrant reached significance only in the 327-cancer cell line dataset. For parts c and d, significant genes are defined by 1 Mb region window, correcting for tumor type and CNA, and “cancer-related” is by COSMIC [23]. SV, Structural Variant; FDR, False Discovery Rate; CCLE, Cancer Cell Line Encyclopedia; PCAWG, Pan-Cancer Analysis of Whole Genomes; TCGA, The Cancer Genome Atlas