Fig. 3

Landscape of ITH in ccRCC. A Oncoprint of key ccRCC driver mutations and copy number alterations for all regions of all 29 patients in this cohort. Margin shows comparison between mutation frequency observed in this cohort and TRACERx Renal. B We have performed unsupervised hierarchical clustering of genomic features including patient level presence or absence of a small variant in VHL, PBRM1, SETD2, BAP1 (most commonly mutated genes) as well as loss of heterozygosity in HLA genes as well as 9p (which includes CDKN2A/B) SCNA which are known to affect ICI response. Heatmap shows ITH high vs low classification across data type. Annotation illustrates evolutionary subtypes and treatment status of patients. A patient is annotated as wildtype if all regions are wild type for that alteration. Cases where ITH score could not be calculated due to lack of sufficient number of biopsies are shown in gray pixels. CIN: chromosome instability. C Association between antigen presentation machinery (APM), effector T cell (Teff) and myeloid gene signatures, and ITH. Wilcox P, false discovery rate (FDR), and linear mixed effect (LME) P shown. D Intratumoral heterogeneity and myeloid score are associated with CDKN2A/B loss in TCGA KIRC cohort. E ITH low patients show a significantly higher TCR diversity, richness, and clone count