ChIP-chip/seq in combination with transcriptome profiling has greatly helped our understanding of the molecular mechanisms underlying many physiological and pathological processes. It has also left unanswered questions on the combinatorial and context-specific nature of mammalian transcription regulation, and created challenges for computational data integration and modeling. To address these challenges, we propose to: 1) develop the computational framework for constructing condition-specific combinatorial and probabilistic transcription regulatory modules in mammalian genomes by integrating transcription factor ChIP-chip/seq, cis-element epigenome and transcriptome data;2) apply the model in 1) to construct a comprehensive probabilistic nuclear receptor regulatory network, experimentally validate the predictions, and use the results to refine the model;3) develop and maintain an open source publicly available integrated ChIP-chip/seq data analysis pipeline Cistrome. With rapid growth of transcription factor ChIP-chip/seq, cis-element epigenome, and transcriptome datasets, our methods will integrate the available datasets, infer the important missing data, and extract maximum knowledge from individual datasets. Our resulting nuclear receptor regulatory network and computational tools will also be a good resource for the community.
The proposed study will lead to a suite of powerful and user-friendly computational tools for integrative analysis of ever-increasing amount of and diverse sources of genomic data in understanding gene regulation in mammals. These tools will allow biologists to perform discovery- based computational analyses using state-of-the-art probabilistic data mining methods. It will also build a nuclear receptor transcription regulatory network, which will provide important insights into identifying new therapeutic targets and designing novel therapeutic strategies, especially combination therapies for nuclear receptor related diseases such as atherosclerosis, diabetes and cancer.
Showing the most recent 10 out of 67 publications
