Mapping and Prediction of Epistatic Enhancer Interactions in Gene Regulation

Fred Hutchinson Cancer Center, Seattle WA

Investigators

Abstract

Project Summary Gene regulation is a fundamental biological process through which noncoding regulatory elements work interactively in a holistic model to determine traits and diseases. Regulatory elements (e.g., enhancers) recruit transcriptional factors (TFs) collectively and explicitly control gene expression to define phenotypic states. The epistatic interaction, in which the function of one enhancer depends on the presence or absence of another enhancer, has been recognized as fundamentally important to understand the complexity of gene regulation. Elucidating the principles of how enhancers coordinate to control gene transcription, and the TFs they utilize to get there, is a major challenge in the field. Perturbing combinatorial enhancers and measuring their epistatic effects on the gene regulatory contributions to traits and diseases is necessary. Recent advances in CRISPR-based genome engineering technologies have enabled new approaches for the multiplexed perturbation of regulatory elements. We recently demonstrated that multiplexed CRISPR interference (CRISPRi) systems enable the robust and accurate interrogation of the function of enhancer epistasis networks in gene regulation. Furthermore, the emergence of new technologies and datasets, including single-cell RNA sequencing, non-linear deep learning models, and large-scale genome-wide association studies (GWAS) datasets, provides unprecedented opportunities to uncover the mechanisms underlying enhancer epistatic interactions and how they relate to gene expression. My research program aims to exploit this enormous potential by developing analytic approaches that probe the enhancer epistatic interactions in gene regulation and model the interactive consequence of noncoding variants for functionality and ultimately pathogenicity. We will (i) map the epistatic interactions by establishing best practices in multiplexed perturbation assay design and analysis; (ii) dissect how inter-molecular TF interactions determine enhancer epistasis; (iii) develop novel mechanistic computational approaches to predict the functional consequence of interactive enhancer variants. Our studies will lead to novel insights into fully understanding the regulatory complexity that controls the expression of human genes. They will shift the paradigm in the interpretation of millions of noncoding disease variants from a locus-by-locus model to an epistasis-aware model. We anticipate that ability will be particularly powerful for translating genetic associations into disease mechanisms, thus creating a windfall of new knowledge about how genomic regulatory elements contribute to disease, and how to monitor and manipulate enhancers for diagnostic and therapeutic benefit. 1