Year: 2026
Type: preprint
Abstract: We present FRACTEL, a statistical framework for region-level analysis of CRISPR perturbation screens. FRACTEL aggregates gRNA p -values using a bounded minimum across order statistics, preserving scale and enabling adaptive sensitivity to sparse or diffuse effects. Region-level null distributions are estimated via simulation, ensuring precise type I error control. Simulations and real CRISPRi/a datasets demonstrate improved power and replication rate over gRNA-level analyses. FRACTEL also informs experimental design, revealing trade-offs between gRNA redundancy and efficacy and identifying inherent limits in single-cell repression screens of lowly expressed genes. The method integrates with existing pipelines and supports diverse CRISPR screening applications.
Institution: Duke University
Language: English
Subfield: Molecular Biology
Domain: Life Sciences
Open Access status: green
OpenAlex ID: W7167675929