Föreläsare: Karin Forsberg Nilsson, Uppsala universitet
Värd: Silvia Remeseiro, WCMM
Föreläsningen är på engelska
Cancer genomics has been biased against non-coding mutations (98% of all mutations), and driver mutations outside exons remain underrepresented. We hypothesized that a position under evolutionary constraint in the genome could be used to identify functional cancer mutations. In our previous work, we have successfully detected functionally active non-coding constraint mutations (NCCMs) in glioblastoma. The Zoonomia Consortium has since published single base (phyloP) constraint scores from whole genome alignments of 240 mammals. Based on a comparison to previous large-scale genome annotations, and data sets such as ClinVar we provide proof-of-concept that evolutionary constrained positions are enriched for variants behind human disease, e.g. medulloblastoma.
Next, we examined if NCCMs and their associated enriched genes exert functional effects across multiple tumor types. To investigate this, we annotated evolutionary constraint and mapped to somatic single‑nucleotide variants (SNVs) and indels across >2,500 tumors representing 19 tumor types in the ICGC‑PCAWG dataset. This identified 56 pan-cancer and 132 cancer-type-specific putative driver genes enriched for NCCMs. Notably, genes with a high NCCM load rarely harbor protein-coding mutations, suggesting that distinct selective pressures preserve protein integrity while altering expression regulation.