Fig. 4. Proteogenomic insights into the Metabolic subtype reveal activated oncogenic pathways, enhanced cell cycle signaling and HRD signature.

a Heatmap showing the abundance of key proteins in the MAPK pathway and/or PI3K/Akt/mTOR pathway. For each sample, ssGSEA enrichment score of MAPK- or PI3K-related pathways represents the sample’s own MAPK or PI3K/Akt/mTOR signature score, which is annotated above the heatmap with details shown on the left. b Boxplots showing the abundance of STMN1 S16 and S46 phosphosites across the three proteomic subtypes. For the boxplots: line in the box indicates the median; box borders correspond with the first and third quartiles (25th and 75th percentiles); whiskers extend 1.5 times the interquartile range; outlier data are shown as dots. c Distribution of CCND1 gains (left panel) and CDKN2A/CDKN2B losses (right panel) between the Metabolic subtype and other subtypes. d Boxplots showing the abundance of key cell cycle-related proteins CDK4, CDKN2C, E2F4, PCNA and RB1 across the three proteomic subtypes. P-values are calculated with Student’s t-test. For the boxplots: line in the box indicates the median; box borders correspond with the first and third quartiles (25th and 75th percentiles); whiskers extend 1.5 times the interquartile range; outlier data are shown as dots. e Boxplot comparing the E2F activity score across the three proteomic subtypes. P-values are calculated with Student’s t-test. For the boxplot: line in the box indicates the median; box borders correspond with the first and third quartiles (25th and 75th percentiles); whiskers extend 1.5 times the interquartile range; outlier data are shown as dots. f Histogram comparing the HRD score across the three proteomic subtypes. P-values are calculated with Student’s t-test. g Distribution of copy number losses/deletions of common DDR-related genes in the Metabolic subtype. The 33 Metabolic tumors are arranged in descending order based on HRD score. See also Supplementary Fig. S5 and Table S8.