Leveraging Multi-Omic Data to Characterize Cis-Regulatory Elements and Investigate Their Roles in Gene Regulation

Abstract

Regulatory elements are non–coding genomic regions that interact with transcription factors to govern when, where, and how much of each gene is expressed. Understanding regulatory elements is essential to understanding mammalian gene regulation; however, our understanding of regulatory syntax is incomplete. Here, we have leveraged multi-omic data to characterize regulatory elements and investigate their roles in gene regulation. We examined a subset of CG-rich promoters exhibiting ubiquitous chromatin accessibility. While most promoters are cell-type specific, these promoters are enriched in cell-essential genes. To maintain universal transcription, they recruit distinct TFs. Furthermore, ubiquitous and cell-type specific promoters are enriched in different sets of Mendelian disease genes, suggesting different contributions to disease susceptibility. We investigated DNA methylation patterns among different categories of regulatory elements during mouse embryonic development and identified a novel class of highly conserved bivalent enhancers that are hypermethylated in cancers, suggesting a unique underlying mechanism involved in cancer. We developed a method to annotate transcription-initiating enhancers. We found that transcribed enhancers mark core enhancer regions, but flanking enhancers also contribute to overall activity. Next, we examined the evolutionary conservation pattern of regulatory elements across 240 mammal species. While conserved regulatory elements are associated with essential genes, recently evolving elements are involved in a species’ interaction with its surroundings. Finally, we performed genome-wide annotation of chromatin states across hundreds of samples, building a regulatory landscape in human and mouse genomes. Together, these works contribute to the characterization of regulatory elements and demonstrate distinct mechanisms of different regulatory subclasses.