A high-resolution study of the chromatin environment around regulatory elements

[摘要] Chemical modifications to histones, the proteins around which DNA wraps, are believed to play an important role in gene regulation. These modifications, along with others, make up a cell;;s ;;epigenome.;; It is known that the presence of a particular combination of these modifications at a region of a cell;;s genome determines, for that region, a state that carries functional significance. This work seeks to better understand the importance of not just presence, but also distribution of modifications within regulatory regions. One approach aimed at improving our understanding is to cluster regulatory regions based on information contained in signals that describe, at a high-resolution, the distribution of these modifications. In this thesis we develop a tool, called ChromSMS, to perform this clustering in a biologically meaningful and efficient way that is versatile in handling the underlying complexities of these signals. We apply the tool to data from the NIH;;s Roadmap Epigenomics Project to analyze ChromSMS and to better understand the mechanisms behind the patterns we observe. We find that ChromSMS produces meaningful clusters that are different from each other at a statistically significant level. Using ChromSMS to conduct analyses of epigenomic data, we discover strong relations between GC-content and the distribution of particular modifications. Furthermore, we uncover a small number of patterns that display high functional enrichment, and we begin to study the possible role and significance of motifs in driving these patterns. We conclude that ChromSMS can serve as a useful tool in examining regulatory regions at a high-resolution.

[发布机构] Massachusetts Institute of Technology