Statistical Analysis of Single Cell CRISPR Screens

• Activity Number: 86 - New Topics and Methodological Developments for Single-Cell Data Science
• Type: Invited
• Date/Time: Monday, August 8, 2022 : 8:30 AM to 10:20 AM
• Sponsor: Section on Statistics in Genomics and Genetics
• Author(s): Eugene Katsevich* and Timothy Barry and Kathryn Roeder and Xuran Wang and John Morris
• Companies: University of Pennsylvania and Carnegie Mellon University and Carnegie Mellon University and Carnegie Mellon University and New York Genome Center
• Keywords: single cell CRISPR screen; conditional randomization test; errors-in-variables; gene regulation; differential expression; doubly robust inference

Abstract

Single cell CRISPR screens, pairing CRISPR perturbations of regulatory elements with single cell RNA-sequencing, are an emerging biotechnology promising unprecedented insights into gene regulation. However, the analysis of these screens presents significant statistical challenges. Some of these challenges are inherited from single cell RNA-seq, such as cell type heterogeneity and noisy expression measurements; others are specific to CRISPR experiments, such as the fact that the identities of the CRISPR perturbations a given cell receives are subject to measurement error. In this talk, I will discuss these challenges and present a recent line of work whose goal is to construct reliable hypothesis tests and confidence intervals in this setting.