癌症基因组中的突变特征改变参与细胞信号网络的短线性蛋白基序

Cancer genomes accumulate somatic mutations from diverse mutational processes, driven by extrinsic carcinogen exposures and intrinsic DNA repair deficiencies, leaving distinct genomic imprints referred to as mutational signatures. Mutational signatures are inferred computationally from large cancer genomics datasets and validated in experiments, however their functional impacts on protein-coding sequences and protein function remain largely unexplored. Leveraging machine learning and statistics, we aimed to discern the impact of single base substitution (SBS) mutational signatures in rewiring conserved short linear protein motifs (SLiMs). These are peptide sequences that mediate key protein-protein interactions and post-translational modifications. We performed a proteogenomic analysis of 1.1 million missense single-nucleotide variants (msSNVs) in 12,000 genomes across 19 cancer types. We assessed the impact of msSNVs on 150 classes of SLiMs recognized by kinases and other enzymes involved in cellular signaling networks. Distinct signatures were significantly associated with msSNVs that disrupted, created or switched between specific SLiMs. For example, UV light-associated signature SBS7b disrupted proline-directed kinase motifs, while signatures arising from aberrant APOBEC cytidine deaminase activity, SBS2 and SBS13, disrupted caspase-cleavage motifs, vital for apoptosis, as well as motifs involved in the DNA damage response. These differential signature impacts stem from their trinucleotide-context biases, which steer specific codon changes, inducing amino acid substitutions in distinct SLiMs. At the gene level, recurrent SLiM-rewiring variants clustered in 22 canonical cancer drivers such as BRAF , CTNNB1 and U2AF1 , and 35 candidate genes revealing mechanistic connections between mutational processes and oncogenic signaling pathways. UV light-induced V600E substitutions in BRAF create a novel polo-like kinase (PLK1)-binding motif, putatively contributing to vemurafenib resistance in melanoma patients. Recurrent S34F substitutions in the splicing factor U2AF1 are driven by APOBEC activity and disrupt a regulatory phosphorylation site critical in splice-site fidelity, suggesting a link between APOBEC-driven mutagenesis and RNA splicing dysregulation in lung cancer. Lastly, motif-rewiring alterations in cellular signaling networks strongly correlated with patient attributes such as smoking status and functional characteristics such as APOBEC gene expression, revealing how lifestyle variables and intrinsic mutagenic programs can potentially reconfigure cellular signaling and protein-protein interactions. Together, our results uncover process-specific proteomic consequences of mutational processes, offering mechanistic insights into cancer etiology and exposing potential therapeutic vulnerabilities.