Author ORCID Identifier

Date of Graduation


Document Type

Dissertation (PhD)

Program Affiliation

Cancer Biology

Degree Name

Doctor of Philosophy (PhD)

Advisor/Committee Chair

Traver Hart

Committee Member

Michelle Barton

Committee Member

Jeffrey Chang

Committee Member

Ken Chen

Committee Member

Eduardo Vilar-Sanchez


Adaptation of the bacterial CRISPR-Cas9 system to mammalian cells revolutionized the field of functional genomics, enabling genome-scale genetic perturbations to study essential genes, whose loss of function results in a severe fitness defect. There are two types of essential genes in a cell. Core essential genes are absolutely required for growth and proliferation in every cell type. On the other hand, context-dependent essential genes become essential in an environmental or genetic context. The concept of context-dependent gene essentiality is particularly important in cancer, since killing cancer cells selectively without harming surrounding healthy tissue remains a major challenge. The toxicity of traditional cancer treatment protocols to the normal cells stresses the need for new strategies that can identify and address the weaknesses specific to cancer cells.

Studies showed that CRISPR monogenic knockout screens can identify specific processes that cells rely on for growth and proliferation, which is a crucial step in identifying candidate cancer-specific therapeutic targets. While it is widely accepted that CRISPR screening is both more specific and more sensitive than previously established methods, the limitations of this technology have not been systematically investigated.

In this dissertation, through several lines of integrated analysis of CRISPR screen data in cancer cell lines from the Cancer Dependency Map initiative, I will describe several computational approaches to demonstrate that CRISPR screens are not saturating. In fact, a typical screen has a ~20% false-negative rate, saturating coverage requires multiple repeats and false negatives are more prevalent among moderately expressed genes. I will then introduce a solution to the false negative problem and describe another method that provides a cleaner analysis of the data, rescuing the false negatives observed in these screens. Moreover, I will show that half of all constitutively expressed genes are never observed as essential in any CRISPR screen. Notably, these never-essentials are highly enriched for paralogs, suggesting that functional redundancy masks the detection of a substantial number of genes. Finally, I will describe our efforts to investigate functional buffering among approximately 400 candidate paralog pairs using CRISPR/enCas12a dual-gene knockout screening technology and discuss the paralog synthetic lethal interactions that we have identified, which have escaped detection in monogenic CRISPR-Cas9 knockout screens. Collectively, these observations reveal significant biases and blind-spots in the analysis of CRISPR-based functional genomics approaches and offer new opportunities for the discovery of novel candidate drug targets.


CRISPR-Cas9 screening, functional genomics, systems biology, essential genes, paralogs, synthetic lethality, functional buffering, drug target identification, cancer biology, computational biology