Gene Set Variation Analysis

To calculate single-sample gene set enrichment, we used the gene set variation analysis (GSVA) program (Hanzelmann et al., 2013) to derive the absolute enrichment scores of previous literature reported DNA damage repair (DDR) (Knijnenburg et al., 2018) gene signatures as follows: (1) Base Excision Repair (BER), (2) Nucleotide Excision Repair (NER; including TC-NER and GC-NER), (3) Mismatch Repair (MMR), (4) Fanconi Anemia (FA), (5) Homologous Recombination (HR), (6) Non-Homologous End Joining (NHEJ), (7) Direct Repair (DR), (8) Translesion Synthesis (TLS), (9) Damage Sensor, etc., and oncogenic signaling pathway (Sanchezvega et al., 2018) gene signatures as follows: (1) cell cycle, (2) Hippo signaling, (3) Myc signaling, (4) Notch signaling, (5) oxidative stress response/Nrf2, (6) phosphatidylinositol 3-kinase (PI3K) signaling, (7) receptor tyrosine kinase (RTK)/RAS/mitogen-activated protein (MAP) kinase signaling, (8) TGF-β signaling, (9) tumor protein (TP)53 signaling, (10) b-catenin/Wnt signaling, and (11) Erb-B2 receptor tyrosine kinase (ERBB) signaling. To make a more comprehensive analysis of the functional modules, we further evaluated the activity of Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways (gene sets) (Minoru and Susumu, 2000) within immune subtypes using the single-sample GSVA (ssGSVA). This method quantifies gene set enrichment in individual samples rather than at the group level.