| Sumario: | Simple Summary: Women who inherit harmful variants in the BRCA1 tumour suppressor gene have an elevated risk of developing multiple cancers, and may benefit from increased screening and preventative surgeries. Unfortunately, ~40% of BRCA1 variants have an unknown effect on cancer risk, leaving their carriers with inconclusive information to guide care decisions. This study provides evidence for the classification of variant effects within BRCA1's RING region using two complementary approaches: (1) an artificial intelligence algorithm designed to predict the severity of variant impacts, and (2) a cell-based assay that evaluates whether variant BRCA1 RING proteins retain an interaction required for normal functioning. The results were generally consistent, and confidently clarified the effect of four variants on cancer risk. This study both aids the preventative care decisions for carriers of these variants and underscores the potential of protein-specific approaches for variant characterization. The BRCA1 protein serves an essential function in maintaining genomic integrity, to the extent that up to 80% of women carrying a pathogenic BRCA1 variant develop breast cancer (BC). Most of these carriers would benefit from prophylactic care, but genetic screens that uncover variants of uncertain significance (VUSs) do not provide insight on disease risk or clinical decision-making. In accordance with guidelines established by The American College of Molecular Genetics (ACMG) and Association for Molecular Pathology (AMP), this study produced computational and functional evidence to inform the reclassification of BRCA1 VUSs as pathogenic or benign, with a specific focus on the abundant subset of missense variants within the RING domain. A six-feature linear support vector machine (LSVM) specifically trained on BRCA1 RING variants performed well (84% accurate in predicting in vitro binding loss) and provided supporting classification evidence for 322 VUS. A mammalian cell co-immunoprecipitation (co-IP) assay that quantified the binding between variant BRCA RING constructs and endogenous BARD1 provided corroborating strong evidence for nine VUSs and correlated with a homology-directed repair (HDR) assay by Starita et al. (p = 0.04). The combined evidence warrants the reclassification of three VUSs as likely benign (N16S, A17D, and E100D) and one as likely pathogenic (H41P), and underscores the promise of domain-specific approaches for missense VUS reclassification.
|
|---|