Modeling cis-regulation with a compendium of genome-wide histone H3K27ac profiles

• DOI: 10.1101/gr.201574.115
• Publication Date: 2016-07-28T02:59:27Z
• Authors: Su Wang; Chongzhi Zang; Tengfei Xiao; Jingyu Fan; Shenglin Mei; Qian Qin; Qiu Wu; Xujuan Li; Kexin Xu; Housheng Hansen He; Myles Brown; Clifford A. Meyer; X. Shirley Liu
• Keywords: Epigenomics

Model-based analysis of regulation gene expression (MARGE) is a framework for interpreting the relationship between H3K27ac chromatin environment and differentially expressed sets. The has three main functions: MARGE-potential, MARGE-express, MARGE-cistrome. MARGE-potential defines regulatory potential (RP) each as sum ChIP-seq signals weighted by function genomic distance from transcription start site. MARGE includes compendium RPs derived 365 human 267 mouse data Relative RPs, scaled using this compendium, are superior to superenhancers in predicting BET (bromodomain extraterminal domain) -inhibitor repressed genes. which uses logistic regression retrieve relevant profiles accurately model query set genes, was tested on 671 diverse sets MSigDB. MARGE-cistrome adopts novel semisupervised learning approach identify cis -regulatory elements regulating set. exploits information signal at DNase I hypersensitive sites identified published DNase-seq data. We newly generated RNA-seq upon siRNA silencing multiple transcriptional epigenetic regulators prostate cancer cell line, LNCaP-abl. can predict binding silenced factors without matched Even when matching available, leverages public enhance these This study demonstrates advantage integrating large historical studies regulation.