The shaping of cancer genomes with the regional impact of mutation processes

• Published in: Experimental & molecular medicine Vol. 54; no. 7; pp. 1049 - 1060
• Main Authors: Lee, Soo-Youn, Wang, Han, Cho, Hae Jin, Xi, Ruibin, Kim, Tae-Min
• Format: Journal Article
• Language: English
• Published: London Nature Publishing Group UK 01.07.2022; Springer Nature B.V; 생화학분자생물학회
• Subjects: 38/70; 631/67/1857; 631/67/69; Age; Biomedical and Life Sciences; Biomedicine; Cancer; Chromatin; Deoxyribonucleic acid; DNA; DNA repair; Epigenetics; Evolution; Genetic analysis; Genomes; Mathematical models; Medical Biochemistry; Molecular Medicine; Mutagenesis; Mutation; p53 Protein; Stem Cells; Transcription-coupled repair; Ultraviolet radiation; 생화학
• ISSN: 2092-6413; 1226-3613; 2092-6413
• DOI: 10.1038/s12276-022-00808-x

Mutation signature analysis has been used to infer the contributions of various DNA mutagenic-repair events in individual cancer genomes. Here, we build a statistical framework using a multinomial distribution to assign individual mutations to their cognate mutation signatures. We applied it to 47 million somatic mutations in 1925 publicly available cancer genomes to obtain a mutation signature map at the resolution of individual somatic mutations. Based on mutation signature-level genetic-epigenetic correlative analyses, mutations with transcriptional and replicative strand asymmetries show different enrichment patterns across genomes, and “transcribed” chromatin states and gene boundaries are particularly vulnerable to transcription-coupled repair activities. While causative processes of cancer-driving mutations can be diverse, as shown for converging effects of multiple mutational processes on TP53 mutations, the substantial fraction of recurrently mutated amino acids points to specific mutational processes, e.g., age-related C-to-T transition for KRAS p.G12 mutations. Our investigation of evolutionary trajectories with respect to mutation signatures further revealed that candidate pairs of early- vs. late-operative mutation processes in cancer genomes represent evolutionary dynamics of multiple mutational processes in the shaping of cancer genomes. We also observed that the local mutation clusters of kataegis often include mutations arising from multiple mutational processes, suggestive of a locally synchronous impact of multiple mutational processes on cancer genomes. Taken together, our examination of the genome-wide landscape of mutation signatures at the resolution of individual somatic mutations shows the spatially and temporally distinct mutagenesis-repair-replication histories of various mutational processes and their effects on shaping cancer genomes. Cancer: Modeling mutations to understand disease evolution A statistical model that assigns non-hereditary DNA alterations known as somatic mutations to mutation “signatures” (groups of mutations arising from a specific biological process) on cancer genomes provides novel insights into disease evolution. Somatic mutations result from exposure to factors often linked to cancer development, such as tobacco or ultraviolet radiation. However, assigning a somatic mutation to a particular mutation “signature” remains challenging. The model created by Ruibin Xi (Peking University, China) and Tae-Min Kim (Catholic University of Korea, Seoul, South Korea) and co-workers grouped 47 million somatic mutations in 1925 cancer genomes into localized clusters before connecting them with mutation signatures. This strategy highlights the spatial and temporal patterns related to the origins of mutations, how the DNA strands are repaired and replicated, and how this influences the emerging cancer genome.