CROTONdb: genome-wide variant effect prediction database for CRISPR/Cas9 editing outcomes
CRISPR/Cas9 is a genome editing tool widely used in biological research and clinical therapeutics. Natural human genetic variations, through altering the sequence context of CRISPR/Cas9 target regions, can significantly affect its DNA repair outcomes and ultimately lead to different editing efficiencies. However, these effects have not been systematically studied, even as CRISPR/Cas9 is broadly applied to primary cells and patient samples that harbor such genetic diversity. Here, we present comprehensive investigations of natural genetic variations on CRISPR/Cas9 outcomes across the human genome. The utility of our analysis is illustrated in two case studies, on both preclinical discoveries of CD33 knockout in Chimeric Antigen Receptor (CAR)-T cell therapy, and clinical applications of TTR inactivation for treating ATTR amyloidosis. We further expand our analysis to genome scale, population stratified common variants that may lead to gene editing disparity. Our analyses demonstrate pitfalls of failing to account for the widespread genetic variations in Cas9 target selection, and how they can be effectively examined and avoided using our method. To facilitate broad access to our analysis, a web platform CROTONdb is developed, which provides predictions for all possible CRISPR/Cas9 target sites in the coding region, spanning over 5.38 million gRNA targets and 90.82 million estimated variant effects. We anticipate CROTONdb having broad clinical utilities in gene and cellular therapies.