Genome-wide maps of enhancer-gene regulation to link variants to functions

The National Institutes of Health National Human Genome Research Institute awarded The Leland Stanford Junior University $721,450 on March 1, 2026, under the Human Genome Research program (CFDA 93.172) to develop genome-wide maps of enhancer-gene regulatory interactions and link noncoding variants to disease-associated target genes.

The project extends computational and experimental tools to map enhancer-gene regulatory interactions using single-cell multiomic datasets across hundreds of human cell types and states. Work includes optimizing and deploying a computational model to predict enhancer-gene interactions from single-cell data in complex tissues; building a genome-wide resource of enhancer regulatory interactions across diverse cell types while exploring variation across disease state, biological sex, and age; and functionally characterizing genetic variants associated with cardiovascular disease and traits to validate the utility of these regulatory maps. The deliverables include a predictive model for mapping enhancer-gene interactions across multiple cell types and states and a genome-wide resource of regulatory interactions. Performance occurs in California with a completion date of February 28, 2030.

Abstract

The human genome encodes over 2 million DNA regulatory elements called enhancers that control gene expression in specific cell types and states. Enhancers harbor tens of thousands of genetic variants that influence risk for common diseases and traits. Each of these enhancer variants could reveal insights into the molecular mechanisms of human diseases. Yet, we have lacked tools to systematically map which enhancers regulate which genes in each of the thousands of cell types in the human body. To address this challenge, we have recently applied crispr tools to experimentally test thousands of enhancers in parallel, developed simple computational models that can predict enhancer-gene regulatory interactions from chromatin state, and demonstrated that these predicted enhancer maps can link noncoding variants to novel target genes. These technologies suggest a new strategy to map enhancers across many cell types to connect noncoding variants to target genes. Here, we will extend these technologies to map and predict enhancer-gene regulatory interactions using single-cell multiomic input datasets, and thereby enable applications to many additional cell types and states. First, we will optimize, stress-test, and deploy a computational model to predict e-g regulatory interactions from single-cell datasets in complex tissues. Second, we will build a genome-wide map of enhancer regulation across 100S of human cell types, and explore variation across disease state, biological SEX, and age. Third, we will apply these tools to functionally characterize genetic variants associated with cardiovascular diseases and traits, as an exemplary system to validate the utility of these maps in linking variants to functions. Together, this proposal will deliver (i) a predictive model to map e-g interactions across many cell types and states; (II) a genome-wide resource of e-g regulatory interactions across human cell types; (III) insights into how this regulatory wiring differs across cell types and key biological variables; and (IV) novel enhancers, genes, and cell types that affect risk for heart structure and function. These methods will enable future studies to study the impacts of variants and elements on genome function across a wide range of biological contexts and diseases.