Sparse dictionary learning recovers pleiotropy from human cell fitness screens

Joshua Pan,Jason J. Kwon,Jessica A. Talamas,Ashir A. Borah,Francisca Vazquez,Jesse S. Boehm,Aviad Tsherniak,Marinka Zitnik,James M. McFarland,William C. Hahn

DOI: https://doi.org/10.48550/arXiv.2111.06247

2021-11-11

Abstract:In high-throughput functional genomic screens, each gene product is commonly assumed to exhibit a singular biological function within a defined protein complex or pathway. In practice, a single gene perturbation may induce multiple cascading functional outcomes, a genetic principle known as pleiotropy. Here, we model pleiotropy in fitness screen collections by representing each gene perturbation as the sum of multiple perturbations of biological functions, each harboring independent fitness effects inferred empirically from the data. Our approach ('Webster') recovered pleiotropic functions for DNA damage proteins from genotoxic fitness screens, untangled distinct signaling pathways upstream of shared effector proteins from cancer cell fitness screens, and learned aspects of the cellular hierarchy in an unsupervised manner. Modeling compound sensitivity profiles in terms of genetically defined functions recovered compound mechanisms of action. Our approach establishes a sparse approximation mechanism for unraveling complex genetic architectures underlying high-dimensional gene perturbation readouts.

Quantitative Methods,Genomics,Molecular Networks