Skip to main content


Figure 3 | Genome Biology

Figure 3

From: EMu: probabilistic inference of mutational processes and their localization in the cancer genome

Figure 3

Clustering of mutations within the cancer genome. (A) The index of dispersion (corrected for the heterogeneity in spectrum-opportunity overlap and ordered by process B value) for all 21 tumors and the elementary processes (logarithmic scale). Process B shows a tendency to be over-dispersed in some tumors (values larger than 1) indicating local clustering of mutations in the corresponding cancer genomes. (B) Local assignment of mutations to processes for sample PD4107a by chromosome. (C) The corresponding P value with respect to the null hypothesis that the mutations are randomly distributed in the genome (positive values signify a surplus of mutation, negative values a deficit; values are capped at 20) by chromosome. The kataegis events in chromosomes 6 and 12 can clearly be attributed to process B. (D) and (E) show the same information for sample PD4103a, which harbors many more kataegis-like events, albeit of different magnitude. Green: process A; pink: process B; blue: process C; yellow: process D.

Back to article page