Fig. 2

Applying SigProfilerSimulator to three distinct cancer genomics problems. a Distribution of the expected number of doublet base substitutions (DBSs) due to the adjacent single base substitutions (SBSs) observed by chance for the PCAWG sample SP99325. The distributions represent the results from 1000 simulations of the mutational pattern of SP99325 treating mutations as statistically independent events (blue) and 1000 simulations of the mutational pattern of SP99325 treating mutations as dependent events. b The fold increase of DBSs observed in the original PCAWG samples and the average number of DBSs observed in our simulations. The mutational pattern of each sample was generated 1000 times considering somatic mutations as statistically independent events. c Comparing the similarities of mutational patterns at ± 2 bp context (SBS-1536) between real and simulated PCAWG samples. Simulations were performed at SBS-6 and SBS-96 resolutions. d Comparing the similarities of mutational patterns at ± 3 bp context (SBS-24576) between real and simulated PCAWG samples. Simulations were performed at SBS-6, SBS-96, and SBS-1536 resolutions. e Evaluating the false-positive rates of MutSigCV1.41, MutSigCV2, and dNdScv driver detection tools using SigProfilerSimulator. All TCGA breast cancer WES samples were simulated 100 times and examined for driver mutations using both MutSigCV and dNdScv. The average number of significant driver genes are plotted using a recommended q-value cutoff of 0.10