Figure 2. Gene Set Enrichment Analysis and Immune Deconvolution in the Validation Cohort.
Heat map representing enrichment and depletion of multiple gene sets and immune compartments. Single-sample gene set enrichment analysis (A) and immune deconvolution (B) are shown across molecular subtypes. Statistical analysis of enrichment scores and immune deconvolution consisted of t tests for pairwise comparison between subtypes, with P values adjusted by controlling the false discovery rate (<.05) to account for multiple comparisons. If a row is annotated with a single subtype label (canonical [C], immune [I], or stromal [S]), that molecular subtype was enriched for the corresponding pathway with adjusted P < .05 (compared with each of the other subtypes). If a row is annotated with 2 subtype labels, both subtypes were enriched with adjusted P < .05 compared with the other subtype. An additional detailed view is presented in eFigure 7 in Supplement 2. IFN indicates interferon; TGF-β, transforming growth factor β; TNF, tumor necrosis factor.