Identifying tissue-selective transcription factor binding sites in vertebrate promoters

Abstract

We present a computational method aimed at systematically identifying tissue-selective transcription factor binding sites. Our method focuses on the differences between sets of promoters that are associated with differentially expressed genes, and it is effective at identifying the highly degenerate motifs that characterize vertebrate transcription factor binding sites. Results on simulated data indicate that our method detects motifs with greater accuracy than the leading methods, and its detection of strongly overrepresented motifs is nearly perfect. We present motifs identified by our method as the most overrepresented in promoters of liver-and muscle-selective genes, demonstrating that our method accurately identifies known transcription factor binding sites and previously uncharacterized motifs.bioinformatics ͉ motif discovery D issecting the transcription-regulation networks in higher eukaryotes is an immediate challenge for systems biology. Techniques like microarray analysis and chromatin immunoprecipitation have produced a significant volume of expression and localization data that can be used to investigate this machinery. Transcription factors play a prominent role in transcription regulation; identifying and characterizing their binding sites is central to annotating genomic regulatory regions and understanding gene-regulatory networks.Computational methods that use both sequence and expression data to identify transcription factor binding sites (TFBS) are becoming increasingly accurate (1), but binding-site identification in vertebrates remains a difficult problem. Tissue-selective transcription regulation requires more complex regulatory machinery, contributing to less predictable binding-site location (2) and a greater role for combinatorial control (3). Fortunately, knowledge of gene expression in different tissues can facilitate the detection of tissue-selective regulatory elements through comparative analysis of regulatory sequences.Tools for discovering binding sites associated with specific tissues need to be able to identify highly degenerate motifs that are overrepresented in one set of promoters relative to another. Motif-discovery algorithms, such as CONSENSUS (4), MEME (5), and GIBBS MOTIF SAMPLER (6) represent motifs as positionweight matrices and can express sufficient degeneracy, but none of these algorithms focus on relative overrepresentation between two sets of sequences. DMOTIFS (7) identifies the motifs that best discriminate between two sets of sequences, but an initial, and often prohibitively large, set of candidate motifs must be provided. Other methods that allow the user to give a background set (1, 8) use the background to fit a statistical model, which is then used to determine overrepresentation.We describe a general method for discovering TFBSs by identifying motifs based on a relative overrepresentation between two sets of promoters. Our method, DME (discriminating matrix enumerator), uses an enumerative algorithm to exhaustively and efficiently search a discrete space...

Open Chromatin Profiling in Mice Livers Reveals Unique Chromatin Variations Induced by High Fat Diet

“…Transcript abundance was quantified by fragments per kilobase of exon per million fragments mapped (FPKM) using Cufflinks (23). Cuffdiff (24) (25). For both strains, we performed motif discovery on the top 1000 HF-specific FAIRE peaks with 10,000 random selected regions of length 500 bp used as background in DME.…”

Site identification in high-throughput RNA–protein interaction data

“…To be agnostic of existing characterizations of an RBPs motif, we perform de novo motif discovery for each dataset and consider the top enriched motif to be the correct one. For motif discovery, we use the DME algorithm (Smith et al, 2005). Full details of the scoring method used are given in Supplementary Material.…”

Sequence signatures of genes with accompanying antisense transcripts in Saccharomyces cerevisiae

“…Motifclass DME (discriminating matrix enumerator) is a widely used tool for identifying motifs overrepresented in the set of foreground sequences relative to a set of background sequences [12]. It uses an enumerative algorithm to exhaustively and efficiently search a discrete space of matrices, scoring each matrix according to its relative overrepresentation in the foreground.…”