Enhancing statistical accuracy in gene perturbation studies

Accurately analysing gene expression changes in high-throughput perturbation studies remains a challenge due to confounding technical factors. This paper evaluates and extends the SCEPTRE (Single-Cell PerTurbation screens via Conditional REsampling) framework, originally introduced by Barry et al. (2021), demonstrating its applicability to high-multiplicity-of-infection (MOI) CRISPR screens. By leveraging a resampling-based methodology, our approach effectively adjusts for sequencing biases, reducing false discoveries while maintaining statistical power.