Fig. 2: Prevalence of driver gene mutations in localized and metastatic prostate cancer.

A The proportion of tumors harboring each driver mutation in localized prostate cancer or mCRPC (‘Observed Δ Proportion’), as described. Dot size indicates −log₁₀ q-value; dot color indicates driver mutation type. Specific genes of interest are labeled. B Comparison of driver gene mutation prevalence in localized disease and mCRPC. Differences in proportion of localized and metastatic cases harboring each driver mutation (‘Observed Δ Proportion’) were subtracted from the difference in proportions resulting from 100,000 simulations, per gene, per sample, where the probability of observing a mutation in a given sample was weighted by the global mutational burden (i.e. SNVs per Mb or PGA) in that sample (‘Expected Δ Proportion’) to generate the Adjusted Δ Proportion. A two-sided p-value was calculated as the proportion of simulated proportions that were as extreme or more extreme than the observed Δ proportion. q-values were then derived using the False Discovery Rate method. C Positive adjusted Δ proportion indicates higher than expected prevalence in mCRPC, while negative adjusted Δ proportion indicates lower than expected prevalence. Mutations are ordered from top to bottom by adjusted Δ proportion, as shown by the height of each bar. Error bars represent 95% confidence intervals; centre of the error bars represents the adjusted Δ proportion. Statistical significance was tested using adjusted Fisher’s Exact tests with correction for multiple testing using the False Discovery Rate method. Mutations with q-value < 0.05 (red line) were considered statistically significant. Source data are provided as a Source Data file.