Systematic identification of cooperation between DNA binding proteins in 3 D space

Cooperation between DNA-binding proteins (DBPs) such as transcription factors and chromatin remodeling enzymes plays a pivotal role in regulating gene expression and other biological processes. Such cooperation is often via interaction between DBPs that bind to loci located distal in the linear genome but close in the 3D space, referred as trans-cooperation. Due to the lack of 3D chromosomal structure, identification of DBP cooperation has been limited to those binding to neighbor regions in the linear genome, referred as cis-cooperation. Here we present the first study that integrates protein ChIPseq and Hi-C data to systematically identify both cisand trans-cooperation between DBPs. We developed a new network model that allows identification of cooperation between multiple DBPs and reveals cell type specific or independent regulations. Particularly interesting, we have retrieved many known and previously unknown transcooperation between DBPs in the chromosomal loops that may be a key factor for influencing 3D chromosomal structure. The software is available at http://wanglab.ucsd.edu/star/DBPnet/index.html. Introduction DNA binding proteins (DBPs) such as transcription factors (TFs), insulators and chromatin remodeling enzymes play key roles in many important biological processes. These proteins rarely function alone but rather cooperate with one another to regulate gene expression, epigenetic modifications and formation of 3D interactions between distal genomic loci[1, 2]. Identification of DBP cooperation is thus critical for understanding the mechanisms regulating these crucial molecular and cellular functions. Previous studies have focused on identifying DBPs binding to neighbor genomic regions[3-7], which is hereinafter referred as cis-cooperation. Despite the great insight of these studies provided into revealing combinatorial regulation of DBPs, they missed cooperation between DBPs binding to distal genomic loci but localized in spatial proximity to form so called trans-cooperation. Trans-cooperation of DBPs either enhances the existing 3D contacts or creates new ones to bring functional elements such as enhancers to their target loci such as promoters. However, no study has thoroughly investigated the trans-cooperation and its relationship with cis-cooperation. The ENCODE project has generated hundreds of ChIP-seq data to map binding sites of DBPs in multiple cell lines[8-10]. Recently, kilobase-resolution Hi-C data were available in two of these cell lines, GM12878 and K562[11]. These data provide an unprecedented opportunity to systematically map both cisand trans-cooperation between DBPs. However, it is a great challenge to analyze this large amount of data and extract the cooperation among multiple rather than a pair of DBPs. To tackle this challenge and comprehensively catalog DBP cooperation, we present here a new model to constructing networks that represent both cisand trans-association peer-reviewed) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (which was not . http://dx.doi.org/10.1101/036145 doi: bioRxiv preprint first posted online Jan. 7, 2016;