Fig. 2

Genetic analysis of gene pairs that affect cancer cell response to T cell killing. A–D Tukey box plots (IQR boxes with 1.5 × IQR whiskers) overlaid on dot plots of sgRNA pair abundances for each DKO, SKO, and DNTCs for pre-T cell treatment controls (also labeled as “cell”) and post-selection co-culture samples with reads pooled from samples across screens, E:T ratios, and technical replicates. Count distributions are shown for gene pairs Jak1_Kmt2d, Jak2_Kmt2d, Jak1_Trp53, and Jak2_Trp53. E Boxplots of KMT2D and JAK1 expression in RNA-seq samples from the TCGA SKCM melanoma cohort and paired normal samples (461 tumor samples, 558 normal samples). *, q-value < 0.01 and |log2FC|> 0.5. F Bar plots of top enriched pathways identified by DAVID biological processes analyses of genes negatively (left) and positively (right) correlated with KMT2D expression in 473 melanoma RNA-seq samples from TCGA. G Boxplots of effector T-cell gene signatures (CX3CR1, FGFBP2, FCGR3A) and exhausted T-cell gene signatures (HAVCR2, TIGIT, LAG3, PDCD1, CXCL13, LAYN) in RNA-seq samples from the TCGA SKCM melanoma cohort and paired normal samples (461 tumor samples, 558 normal samples). *, q-value < 0.01 and |log2FC|> 0.5. H Boxplots of normalized cell type proportions from CIBERSORT deconvolution analyses of TCGA-SKCM and GTEx normal skin RNA-seq samples for T cells. Statistics shown on plots. I Kaplan–Meier curves showing the survival of melanoma patients from the GSE22153 cohort based on the expression status of KMT2D and linked to estimated cytotoxic T lymphocyte (CTL) levels. Analyses performed with the Tumor Immune Dysfunction and Exclusion (TIDE) algorithm. Statistics shown on plot