ChIP-Seq of transcription factors predicts absolute and differential gene expression in embryonic stem cells

Abstract

Next-generation sequencing has greatly increased the scope and the resolution of transcriptional regulation study. RNA sequencing (RNA-Seq) and ChIP-Seq experiments are now generating comprehensive data on transcript abundance and on regulator-DNA interactions. We propose an approach for an integrated analysis of these data based on feature extraction of ChIP-Seq signals, principal component analysis, and regression-based component selection. Compared with traditional methods, our approach not only offers higher power in predicting gene expression from ChIP-Seq data but also provides a way to capture cooperation among regulators. In mouse embryonic stem cells (ESCs), we find that a remarkably high proportion of variation in gene expression (65%) can be explained by the binding signals of 12 transcription factors (TFs). Two groups of TFs are identified. Whereas the first group (E2f1, Myc, Mycn, and Zfx) act as activators in general, the second group (Oct4, Nanog, Sox2, Smad1, Stat3, Tcfcp2l1, and Esrrb) may serve as either activator or repressor depending on the target. The two groups of TFs cooperate tightly to activate genes that are differentially up-regulated in ESCs. In the absence of binding by the first group, the binding of the second group is associated with genes that are repressed in ESCs and derepressed upon early differentiation.ChIP binding ͉ pluripotency ͉ RNA-Seq ͉ transcription regulation

Inference of transcriptional regulation in cancers

“…edu). As a key distinction from other algorithms integrating gene expression and TF ChIP-seq data (15)(16)(17)(18), RABIT better captures the properties of cancer cells, such as CNA and DNA methylation, that shape tumor gene expression independently from TF regulation. Additionally, somatic mutations of the TF-coding region can perturb transcriptional regulation.…”

“…Notably, there are abundant previous works on integrating ChIP-seq and gene expression data to understand gene regulatory mechanisms (15). For example, ChIP-seq profiles of 12 TFs and RNA-seq expression profiles in mouse embryonic stem cells have been analyzed together, using the regression method (16)(17)(18). However, these previous studies were conducted when ChIP-seq and expressionprofiling data were generated in the same condition, without further requirement of removing any background confounding effect.…”

The TAGteam motif facilitates binding of 21 sequence-specific transcription factors in the Drosophila embryo

“…Supporting this hypothesis, clusters of transcription factor binding sites have been used to identify novel enhancers (Berman et al 2002;Markstein et al 2002), a method that is particularly powerful when combined with sequence conservation (Kazemian et al 2010), and which has proved useful even in the absence of experimentally determined transcription factor binding specificities ). Many, although not all, regions bound in vivo by multiple transcription factors confer enhancer activity and drive gene expression (Chen et al 2008;Ouyang et al 2009;Zinzen et al 2009;Negre et al 2011). However, the mechanisms underlying these overlapping patterns of binding remain incompletely understood.…”

Combining transcription factor binding affinities with open-chromatin data for accurate gene expression prediction

“…27 Several general approaches have been proposed to identify TFs acting as key players 28 in gene regulation depending on the available data: Coexpression analysis combined 29 with computational predictions of TF sequence binding can be used to identify key 30 TFs [18]. Genome-wide TF binding data, as produced by TF ChIP-seq, is widely used 31 to identify important TFs: ChIP-seq data was incorporated into coexpression 32 analysis [43], was combined with transcriptome data [49,66], used for the construction 33 of Gene Regulatory Networks (GRNs) [8], and used together with Hi-C data [40].…”

“…It was shown that using 47 DNaseI-seq data alone can lead to highly accurate TF binding predictions [13,51], 48 therefore we mainly focus on open-chromatin data in this article. 49 There are two general classes of methods to predict TF binding: site-centric 50 methods [13,32,42,51,57,69], and segmentation-based methods [3,7,24,25,27,48,50,58]. 51 Site-centric methods require the identification of putative TF binding sites (TFBS) 52 using TF binding motifs represented with position weight matrices (pwms).…”