Summary CRISPR-StAR (stochastic activation by recombination) is an inducible pooled screening system that activates gene knockout after tumor engraftment and provides matched internal controls for guide-level normalization. In this study, we explore the scalability and reproducibility of this approach for in vivo cancer screens. Through barcode-embedded sequencing and the development of a Bayesian analysis pipeline, we screened a 30,000-sgRNA library in A549 xenografts, achieving reproducible dropout and enrichment phenotypes using just ∼30 tumors. Across additional xenograft models, single tumors yielded reliable functional annotation for ∼1,000 genes. Comparing in vivo and in vitro screens uncovered tumor suppressor effects detectable only in vivo; for example, KMT2C and KMT2D knockouts produced contrasting growth and transcriptional programs. Together with our R analysis package, we show that CRISPR-StAR enables scalable in vivo dependency mapping that complements in vitro resources and reduces animal use by up to 7-fold versus conventional dropout screens, improving methodological rigor at genome-scale clonal resolution.
Fenoglio et al. (Mon,) studied this question.