Scalable chemical-genetic screen platform for predicting compound mode-of-action

Abstract

Screening chemical compounds against a collection of defined gene mutants can identify mutations that confer sensitivity or resistance to a compound’s effect. Advances in CRISPR-Cas9 gene editing technology have enabled these chemical-genetic screens in human cells, allowing for the construction of high-dimensional functional profiles that can be used to predict compound mode-of-action. However, scalability of these screens has proven to be difficult due to the cost- and time-intensive nature of these screens. My contributions to the field of drug discovery through chemical-genetic screens can be divided into two major components. First, I designed an efficient scalable chemical screen platform for human cell lines and describe its advantages relative to traditional genome-wide screens. Through five proof-of-principal chemical screens, I demonstrate the ability of these scalable screens to recapitulate similar functional information relative to genome-wide screens, as well as the ability to provide temporal resolution on novel biology. Second, I adapted a computational tool called CG-TARGET to utilize a reference set of genetic interaction profiles to interpret chemical-genetic interaction profiles and predict the bioprocesses perturbed by these screened compounds. The insights derived from the work described in this dissertation have the potential to move us closer to functionally annotating compounds at scale and discovering novel therapeutics for clinical use.