Computational tools for studying gene regulation in the 3-dimensional genome

Abstract

Determining the 3-dimensional structure of the genome and its impact on gene expression has been a long-standing question in cell biology. Recent development in mapping technologies for chromatin interactions has led to a rapid increase in this kind of interaction data, revealing a hierarchical organization of the 3D genome, from large compartments spanning multiple chromosomes, to mega-base-sized topological associated chromatin domains, to individual long-range chromatin loops mediating enhancer-promoter interactions. With the explosion of chromatin interaction data, there is a pressing need for analytical tools. In this talk, I will describe two computational algorithms for analyzing chromatin interaction data at different scales. I will first present a fast algorithm for identifying large-scale, hierarchical chromatin domains. I will demonstrate how the algorithm enables studies of chromatin subdomains in gene regulation. Accurate knowledge of enhancer-promoter interactions is a pre-requisite to understanding regulatory output of enhancers. I will present an algorithm for predicting enhancer-promoter interactions by integrating genomic, transcriptomic, and epigenomic data. Using data from multiple human cell types, I will demonstrate how the algorithm can help decipher the mechanisms underlying enhancer-promoter communication.