Extended Data Fig. 2: Function scores based on cell fitness and drug response enhances classification accuracy of ESC-based BRCA2 SGE.
From: Saturation genome editing-based clinical classification of BRCA2 variants
a, The Pearson correlation of Function Scores (FS) for 6551SNVs between cell fitness (DMSO treated) and response to drugs (cisplatin treated). Additionally, we assessed the correlation between cell fitness and olaparib treatment, as well as the correlation between cisplatin and olaparib responses. b, Schematic showing our filtering strategy to integrate classification scores based on cell fitness and drug response data to accurately categorize SNVs. 6551 SNVs were categorized using Mixture-modelling based on their survival at day 14, and integrated FS values were computed by combining cell fitness and drug response scores. c, By employing GMM, we categorized variants based on the integrated values, revealing that 6209 (94.7%) SNVs exhibited strong concordance, while only 353 (5.4%) SNVs showed an opposite classification with cell fitness data. d, Heatmap represents the percentage of variant recovered and then finally classified by integrating the scores and excluding the uncertain variants. The value in the boxes represent the percentages ranging from 82% (light blue) to 100% (dark blue).
