Fig. 2: Multi-organ temporal mutational patterns of RAM tumors.
a, Phylogenetic relationship of multi-organ metastases in 10 RAM cases organized by treatment histories. Each tumorʼs somatic mutations (SNVs and IDs) were used to construct a maximally parsimonious phylogenetic tree. Some branches, as indicated, are not shown to scale owing to extensive lengths. Each evolutionary trajectory is annotated by selected cancer genes and their mutations. AMP, copy number amplification; DEL, copy number deletion; Gb, gallbladder; NRM (blue node), normal tissue; OM, omentum. b, Spectra of mutational signatures among early, intermediate and late mutations, based respectively on shared, semi-private and private SBSs in a across RAM cases, organ sites and treatment histories. With WES of two RAM14001 tumors available, we identified only early and late mutations for signature detection. c, Unsupervised clustering of treatment histories, RAM cases and organ sites based on the proportions or compositions of SBS mutational signatures present in early, intermediate and late mutations in b. d, Analysis of SBS mutational signatures in a clinical cohort of patient-matched MAPKi-sensitive (referred to as baseline) and MAPKi-acquired resistant (referred to as disease progression (DP)) cutaneous BRAFMUT melanoma tumors (n = 88 tumors). MAPKi-sensitive tumors represent ICB-naive, pre-MAPKi-treatment clinical tumors (n = 28 patients). e, As in d, except all samples are PDX tumors from sex-matched NSG mice and consisting of patient-matched (n = 8 models), vehicle-treated and MAPKi-sensitive tumors (n = 8) and acquired MAPKi-resistant BRAF-mutant or NRAS-mutant tumors (n = 21). f, As in d, except all samples are clinical tumors (n = 14) consisting of patient-matched MAPKi-naive (except one), pre-ICB baseline tumors and acquired ICB-resistant tumors (n = 7 patients).
