Extended Data Fig. 9
From: Somatic evolution following cancer treatment in normal tissue
A. Global dNdScv rations for missense and nonsense mutations in pre-treated samples versus post-treated samples, in lung (yellow) and blood (red) tissues. Vertical lines represent the confidence intervals. When the lower boundary is higher than 1, we deem the set of mutations under positive selection. The plot shows that there is global positive selection in post-treated samples but not pre-treated samples. Lines represent the confidence intervals. B. Gene-wise (n = 30) correlation between dNdScv selection coefficients pre and post treatment, in blood and lung, across single-base substitutions. Lines represent the confidence intervals. Genes are deemed to be under selectionif q < 0.2 in dNdScv. C. Same as B, using indels if q < 0.2 in dNdScv. C. Same as B, using indels. D. Each dot reflects a normal tissue sample, the number of B2M mutations compared relative to expected (based on dNdScv). Samples are grouped according to whether they were administered immunotherapy. An excess of B2M mutations is observed in immunotherapy treated samples (P = 0.028, one-way wilcoxon test). E. Volcano plots showing in the x axis the maximum normalised ratio of dNdS values across samples from patients treated with different immunotherapies versus non immunotherapy treated, using variable_dNdS_twodatasets from Lawson et al.7 (Anti-CTLA4, 5 patients, 28 samples; Anti-PD1, 4 patients, 44 samples; Combination, 4 patients, 36 samples; No immunotherapy, 8 patients, 48 samples). Q-values are represented in the y axis.
