Fig. 5

The relations among the overall number of recorded cancer-linked somatic mutations, transcript length and number of splice variants. a Linear correlation between the overall number of cancer-linked mutations in different genes and their transcript length. The cancer-linked somatic mutations (as deposited in the COSMIC database) are counted from both the coding and non-coding regions, for the longest transcript of each gene, and plotted against the L^tr length of that transcript. The correlation coefficient (top-left corner) and the linear model fit (red line) are shown. b Relation between the overall number of cancer-linked mutations and the number of N^tr transcript variants in different genes. The individual boxplots (for each N^tr category) describe the distribution of the overall number of mutations found in the longest transcript of each gene, selected from the genes with fixed N^tr number of available transcripts