Genome-Wide Characterisation of Somatic Mutations Across COSMIC Cancer Cell Lines Enables Gene-Centric Experimental Model Selection

Cancer cell lines are widely used to investigate cancer-associated mutations and evaluate candidate therapeutic compounds, but selecting an appropriate model carrying a specific mutation can be challenging. A genome-wide analysis of somatic mutations across 1020 human cancer cell lines from the COSMIC Cell Lines Project was performed to establish a gene- and variant-centric approach for experimental model selection. Transcript-specific annotations were normalised to gene symbols, and redundant transcript-derived records were removed. The resulting dataset contained 3,937,879 unique somatic mutation events affecting 18,885 genes, including 95,307 mutation events across 579 Cancer Gene Census Tier 1 genes. Substantial variation in mutation burden was observed among cell lines, while recurrent cancer-associated genes, including TP53, KRAS, BRAF, EGFR and PIK3CA, were represented across multiple models. Importantly, the cell line containing the greatest number of alterations in a cancer-associated gene did not necessarily carry the specific driver mutation required for a particular experimental question. Representative driver variants were therefore used to demonstrate a variant-centric selection strategy in which cell lines carrying the specific alteration of interest are identified rather than models being selected solely according to gene-specific mutation count. This approach distinguishes descriptive gene-level mutation ranking from variant-specific experimental model selection and provides a systematic framework for identifying genetically appropriate candidate cell lines for functional studies and compound testing.