Fig. 3

Gene exomic location distributions of germline variants within selected cancer predisposition genes. Empirical probability density functions (PDF) of germline variant coordinates, plotted for four selected genes (BRCA1, BRCA2, RET and KRAS). Each line represents the density function of one of the six cancer types, colored by their corresponding sequencing center: BI (SKCM, STAD and THCA) in blue, WUGSC (BRCA and UCEC) in orange, and BCM (LIHC) in green. The genes BRCA1, RET and KRAS display distinct distributions per sequencing batch, while BRCA2 displays a relatively cohesive distribution. Exons are colored by alternating gray and white backgrounds to enhance the visibility of exon boundaries (introns, for which we have no data, are omitted). The schemas of the transcripts (including the non-coding 5′-UTR and 3′-UTR parts) are shown below each figure. For visibility, the graphs are smoothed by kernel density estimation (KDE), using a window size of 100 nt