Predicting the disruption of 3D genome folding in congenital heart defects
Maureen Pittman, BS, UCSF & Gladstone Institutes, PhD Student with Katie Pollard, PhD
Abstract: Many disease-associated genetic variants are located in putative regulatory elements, suggesting a major role for transcriptional misregulation in disease. This is especially relevant to developmental disorders, in which the disruption of finely-tuned transcriptional networks is likely to have acute consequences to cell differentiation and morphogenesis. The fidelity of gene regulation is enforced partly by the three-dimensional organization of the genome, which is arranged such that regulatory elements and their target promoters are able to make contact in the appropriate contexts. Structural variants have the potential to disrupt that system by changing the contact frequency of promoters to their regulatory elements. Here, we hypothesize that such variants may contribute to the formation of congenital heart defects. Using structural variants identified by the Pediatric Cardiac Genomics Consortium, we use a convolutional neural network to predict the resulting change in chromatin contact frequency and prioritize variants for experimental validation.
Genome-wide maps of enhancer regulation connect risk variants to disease genes
Jesse Engreitz, PhD, Stanford University, Assistant Professor
Abstract: Enhancers harbor thousands of variants associated with common diseases and traits. Each of these variants could reveal insights into disease mechanisms or therapeutic targets. Yet, it has proven difficult to connect these variants to their molecular functions because we have lacked tools to systematically map which enhancers regulate which genes in which cell types. Here we use new CRISPR methods to perturb >4,500 enhancer-gene connections in several cell types. We show that an Activity-by-Contact (ABC) Model of enhancer function — involving multiplying enhancer activity by enhancer-promoter contact — can accurately predict these experimental perturbations based on easily obtained maps of chromatin state. We apply this ABC Model to create enhancer-gene maps in 131 cell types and tissues, and use these maps to interpret the functions of genetic variants associated with common diseases and complex traits. We find that over half of causal noncoding GWAS variants likely act via effects on enhancers. Variants associated with cardiovascular diseases connect to genes in endothelial cells and cardiomyocytes. These ABC maps provide a generalizable strategy to connect common disease risk variants in enhancers to target genes, and will help to understand the genetic etiology of congenital heart defects.