FIGURE 10.

Mutation rates are similar within pathways. (A) Dot plot of TCGA samples ranked by the number of simple mutations in coding gene regions. Horizontal reference lines, cutoffs for number of mutations for “low,” “medium,” and “high” mutation load groups (MLGs); vertical reference lines, ranking position cutoffs. (B) Dot plot of ranked normalized numbers of coding gene mutations in the “low” (black), “medium” (dark gray), and “high” (light gray) MLGs. Color, ranking position of commonly mutated oncogenes. (C) Heat map of Gene Ontology (GO) terms for genes ranked in (B) for the three MLGs. Ranked genes in (B) were split into 35 bins of increasing mutational loads (x-axis) so that each bin contained ∼400 genes, irrespective of MLG. Each set of 400 genes was then used to conduct a GSEA using DAVID (https://david.ncifcrf.gov) and the strongest Benjamini–Hochberg-corrected p-value for each GO term within each MLG (red value in parenthesis) was then used to rank the results. For all three MLGs, the most mutated genes (bins ∼30–35) were the olfactory receptor genes (top 2 GO terms) and plasma membrane molecules involved in homophilic cell–cell adhesion (GO terms 3–5). Color-coding and size scale, Benjamini–Hochberg-corrected p-values within individual bins. Pink rectangle, no significant enrichment.