High-throughput methods for elucidating the control of chromatin accessibility

点击翻译按钮获取中文摘要

 DESCRIPTION (provided by applicant): We will develop the first validated predictive model of how transcription factors dynamically determine genome-wide chromatin accessibility that is generalizable across biological systems. We will accomplish this goal with three specific aims. We will develop novel Genome Syntax to Regulation (GSR) models that accurately learn a genomic regulatory vocabulary and predict how phrases in this vocabulary control chromatin accessibility (Aim 1). As part of this aim we will identify transcription factor binding motifs tha are in the discovered regulatory vocabulary. We will validate and refine the causality of these models by testing whether they accurately predict the chromatin accessibility of thousands of synthetic DNA "phrases" that have been engineered into specific genomic locations and measured in the context of transcription factor gain-of-function and loss-of-function studies. The phrases will be designed to elucidate both the factors and grammar that control chromatin opening in several distinct cellular states (Aim 2). We will use our predictive models to assign importance scores to individual genome bases and to predict how selected factors alter chromatin accessibility genome wide (Aim 3). We will test the ability of our importance scores to identify regulatory SNPs in the context of human genome-wide association study (GWAS) data, and we will validate model predictions of changes in whole genome chromatin accessibility in response to ectopic factor expression. Through computational modeling of the effect of such ectopic factor expression, we will develop a predictive understanding of how transcription factors alter chromatin state, laying the groundwork for a novel regenerative medicine paradigm of predictive cellular programming.