Position-specific Weight Matrices Research Articles

On the limits of inferring biophysical parameters of RBP-RNA interactions from in vitro RNA Bind’n Seq data

We develop a thermodynamic model describing the binding of RNA binding proteins (RBP) to oligomers in vitro. We apply expectation-maximization to infer the specificity of RBPs, represented as position-specific weight matrices (PWMs), by maximizing the likelihood of RNA Bind’n Seq data from the ENCODE project. Analyzing these public data we find sequence motifs that can partly explain the data for more than half of the studied 111 RBPs, and for 48 of the proteins these motifs are consistent with the known specificity. Our code is publicly available, facilitating analysis of RBP binding data.

On the limits of inferring biophysical parameters of RBP-RNA interactions from in vitro RNA Bind’n Seq data

We develop a thermodynamic model describing the binding of RNA binding proteins (RBP) to oligomers in vitro. We apply expectation-maximization to infer the specificity of RBPs, represented as position-specific weight matrices (PWMs), by maximizing the likelihood of RNA Bind’n Seq data from the ENCODE project. We demonstrate that the model can reproduce known specificities for well-studied proteins and that in some cases we predict novel, longer binding motifs. However, the model does not recover all the motifs that are in principle known, indicating that the data is not well explained by a single underlying biophysical model. Our code is publicly available.

Transcription Factor Binding Affinities and DNA Shape Readout.

SummaryAn essential event in gene regulation is the binding of a transcription factor (TF) to its target DNA. Models considering the interactions between the TF and the DNA geometry proved to be successful approaches to describe this binding event, while conserving data interpretability. However, a direct characterization of the DNA shape contribution to binding is still missing due to the lack of accurate and large-scale binding affinity data. Here, we use a binding assay we recently established to measure with high sensitivity the binding specificities of 13 Drosophila TFs, including dinucleotide dependencies to capture non-independent amino acid-base interactions. Correlating the binding affinities with all DNA shape features, we find that shape readout is widely used by these factors. A shape readout/TF-DNA complex structure analysis validates our approach while providing biological insights such as positively charged or highly polar amino acids often contact nucleotides that exhibit strong shape readout.

MODER2: first-order Markov modeling and discovery of monomeric and dimeric binding motifs.

MotivationPosition-specific probability matrices (PPMs, also called position-specific weight matrices) have been the dominating model for transcription factor (TF)-binding motifs in DNA. There is, however, increasing recent evidence of better performance of higher order models such as Markov models of order one, also called adjacent dinucleotide matrices (ADMs). ADMs can model dependencies between adjacent nucleotides, unlike PPMs. A modeling technique and software tool that would estimate such models simultaneously both for monomers and their dimers have been missing.ResultsWe present an ADM-based mixture model for monomeric and dimeric TF-binding motifs and an expectation maximization algorithm MODER2 for learning such models from training data and seeds. The model is a mixture that includes monomers and dimers, built from the monomers, with a description of the dimeric structure (spacing, orientation). The technique is modular, meaning that the co-operative effect of dimerization is made explicit by evaluating the difference between expected and observed models. The model is validated using HT-SELEX and generated datasets, and by comparing to some earlier PPM and ADM techniques. The ADM models explain data slightly better than PPM models for 314 tested TFs (or their DNA-binding domains) from four families (bHLH, bZIP, ETS and Homeodomain), the ADM mixture models by MODER2 being the best on average.Availability and implementationSoftware implementation is available from https://github.com/jttoivon/moder2.Supplementary information Supplementary data are available at Bioinformatics online.

An Intricate Network of Conserved DNA Upstream Motifs and Associated Transcription Factors Regulate the Expression of Uromodulin Gene

Uromodulin is a kidney specific glycoprotein whose expression can modulate kidney homeostasis. However, the set of sequence specific transcription factors that regulate the uromodulin gene UMOD and their upstream binding locations are not well characterized. We built a high resolution map of its transcriptional regulation. We applied in silico phylogenetic footprinting on the upstream regulatory regions of a diverse set of human UMOD orthologs to identify conserved binding motifs and corresponding position specific weight matrices. We further analyzed the predicted binding motifs by motif comparison, which identified transcription factors likely to bind these discovered motifs. Predicted transcription factors were then integrated with experimentally known protein-protein interactions available from public databases and tissue specific expression resources to delineate important regulators controlling UMOD expression. Analysis allowed the identification of a reliable set of binding motifs inthe upstream regulatory regions of UMOD to build a high confidence compendium of transcription factors that could bind these motifs, such as GATA3, HNF1B, SP1, SMAD3, RUNX2 and KLF4. ENCODE deoxyribonuclease I hypersensitivity sites in the UMOD upstream region of the mouse kidney confirmed that some of these binding motifs were open to binding by predicted transcription factors. The transcription factor-transcription factor network revealed several highly connected transcription factors, such as SP1, SP3, TP53, POU2F1, RARB, RARA and RXRA, as well as the likely protein complexes formed between them. Expression levels of these transcription factors in the kidney suggest their central role in controlling UMOD expression. Our findings will form a map for understanding the regulation ofuromodulin expression in health and disease.

3DTF: a web server for predicting transcription factor PWMs using 3D structure-based energy calculations

We present the webserver 3D transcription factor (3DTF) to compute position-specific weight matrices (PWMs) of transcription factors using a knowledge-based statistical potential derived from crystallographic data on protein–DNA complexes. Analysis of available structures that can be used to construct PWMs shows that there are hundreds of 3D structures from which PWMs could be derived, as well as thousands of proteins homologous to these. Therefore, we created 3DTF, which delivers binding matrices given the experimental or modeled protein–DNA complex. The webserver can be used by biologists to derive novel PWMs for transcription factors lacking known binding sites and is freely accessible at http://www.gene-regulation.com/pub/programs/3dtf/.

Importance Sampling of Word Patterns in DNA and Protein Sequences

Monte Carlo methods can provide accurate p-value estimates of word counting test statistics and are easy to implement. They are especially attractive when an asymptotic theory is absent or when either the search sequence or the word pattern is too short for the application of asymptotic formulae. Naive direct Monte Carlo is undesirable for the estimation of small probabilities because the associated rare events of interest are seldom generated. We propose instead efficient importance sampling algorithms that use controlled insertion of the desired word patterns on randomly generated sequences. The implementation is illustrated on word patterns of biological interest: palindromes and inverted repeats, patterns arising from position-specific weight matrices (PSWMs), and co-occurrences of pairs of motifs.

Using Sequence-Specific Chemical and Structural Properties of DNA to Predict Transcription Factor Binding Sites

An important step in understanding gene regulation is to identify the DNA binding sites recognized by each transcription factor (TF). Conventional approaches to prediction of TF binding sites involve the definition of consensus sequences or position-specific weight matrices and rely on statistical analysis of DNA sequences of known binding sites. Here, we present a method called SiteSleuth in which DNA structure prediction, computational chemistry, and machine learning are applied to develop models for TF binding sites. In this approach, binary classifiers are trained to discriminate between true and false binding sites based on the sequence-specific chemical and structural features of DNA. These features are determined via molecular dynamics calculations in which we consider each base in different local neighborhoods. For each of 54 TFs in Escherichia coli, for which at least five DNA binding sites are documented in RegulonDB, the TF binding sites and portions of the non-coding genome sequence are mapped to feature vectors and used in training. According to cross-validation analysis and a comparison of computational predictions against ChIP-chip data available for the TF Fis, SiteSleuth outperforms three conventional approaches: Match, MATRIX SEARCH, and the method of Berg and von Hippel. SiteSleuth also outperforms QPMEME, a method similar to SiteSleuth in that it involves a learning algorithm. The main advantage of SiteSleuth is a lower false positive rate.

Inclusion of neighboring base interdependencies substantially improves genome-wide prokaryotic transcription factor binding site prediction

Prediction of transcription factor binding sites is an important challenge in genome analysis. The advent of next generation genome sequencing technologies makes the development of effective computational approaches particularly imperative. We have developed a novel training-based methodology intended for prokaryotic transcription factor binding site prediction. Our methodology extends existing models by taking into account base interdependencies between neighbouring positions using conditional probabilities and includes genomic background weighting. This has been tested against other existing and novel methodologies including position-specific weight matrices, first-order Hidden Markov Models and joint probability models. We have also tested the use of gapped and ungapped alignments and the inclusion or exclusion of background weighting. We show that our best method enhances binding site prediction for all of the 22 Escherichia coli transcription factors with at least 20 known binding sites, with many showing substantial improvements. We highlight the advantage of using block alignments of binding sites over gapped alignments to capture neighbouring position interdependencies. We also show that combining these methods with ChIP-on-chip data has the potential to further improve binding site prediction. Finally we have developed the ungapped likelihood under positional background platform: a user friendly website that gives access to the prediction method devised in this work.

Predicted transcription factor binding sites as predictors of operons in Escherichia coli and Streptomyces coelicolor.

BackgroundAs a polycistronic transcriptional unit of one or more adjacent genes, operons play a key role in regulation and function in prokaryotic biology, and a better understanding of how they are constituted and controlled is needed. Recent efforts have attempted to predict operonic status in sequenced genomes using a variety of techniques and data sources. To date, non-homology based operon prediction strategies have mainly used predicted promoters and terminators present at the extremities of transcriptional unit as predictors, with reasonable success. However, transcription factor binding sites (TFBSs), typically found upstream of the first gene in an operon, have not yet been evaluated.ResultsHere we apply a method originally developed for the prediction of TFBSs in Escherichia coli that minimises the need for prior knowledge and tests its ability to predict operons in E. coli and the 'more complex', pharmaceutically important, Streptomyces coelicolor. We demonstrate that through building genome specific TFBS position-specific-weight-matrices (PSWMs) it is possible to predict operons in E. coli and S. coelicolor with 83% and 93% accuracy respectively, using only TFBS as delimiters of operons. Additionally, the 'palindromicity' of TFBS footprint data of E. coli is characterised.ConclusionTFBS are proposed as novel independent features for use in prokaryotic operon prediction (whether alone or as part of a set of features) given their efficacy as operon predictors in E. coli and S. coelicolor. We also show that TFBS footprint data in E. coli generally contains inverted repeats with significantly (p < 0.05) greater palindromicity than random sequences. Consequently, the palindromicity of putative TFBSs predicted can also enhance operon predictions.

Structure and Function of the Smallest Vertebrate Telomerase RNA from Teleost Fish

Telomerase extends chromosome ends by copying a short template sequence within its intrinsic RNA component. Telomerase RNA (TR) from different groups of species varies dramatically in sequence and size. We report here the bioinformatic identification, secondary structure comparison, and functional analysis of the smallest known vertebrate TRs from five teleost fishes. The teleost TRs (312-348 nucleotides) are significantly smaller than the cartilaginous fish TRs (478-559 nucleotides) and tetrapod TRs. This remarkable length reduction of teleost fish TRs correlates positively with the genome size, reflecting an unusual structural plasticity of TR during evolution. The teleost TR consists of a compact three-domain structure, lacking most of the sequences in regions that are variable in other vertebrate TR structures. The medaka and fugu TRs, when assembled with their telomerase reverse transcriptase (TERT) protein counterparts, reconstituted active and processive telomerase enzymes. Titration analysis of individual RNA domains suggests that the efficient assembly of the telomerase complex is influenced more by the telomerase reverse transcriptase (TERT) binding of the CR4-CR5 domain than the pseudoknot domain of TR. The remarkably small teleost fish TR further expands our understanding about the evolutionary divergence of vertebrate TR.

Finding regulatory elements and regulatory motifs: a general probabilistic framework

Over the last two decades a large number of algorithms has been developed for regulatory motif finding. Here we show how many of these algorithms, especially those that model binding specificities of regulatory factors with position specific weight matrices (WMs), naturally arise within a general Bayesian probabilistic framework. We discuss how WMs are constructed from sets of regulatory sites, how sites for a given WM can be discovered by scanning of large sequences, how to cluster WMs, and more generally how to cluster large sets of sites from different WMs into clusters. We discuss how 'regulatory modules', clusters of sites for subsets of WMs, can be found in large intergenic sequences, and we discuss different methods for ab initio motif finding, including expectation maximization (EM) algorithms, and motif sampling algorithms. Finally, we extensively discuss how module finding methods and ab initio motif finding methods can be extended to take phylogenetic relations between the input sequences into account, i.e. we show how motif finding and phylogenetic footprinting can be integrated in a rigorous probabilistic framework. The article is intended for readers with a solid background in applied mathematics, and preferably with some knowledge of general Bayesian probabilistic methods. The main purpose of the article is to elucidate that all these methods are not a disconnected set of individual algorithmic recipes, but that they are just different facets of a single integrated probabilistic theory.

Protein-DNA binding specificity predictions with structural models

Protein–DNA interactions play a central role in transcriptional regulation and other biological processes. Investigating the mechanism of binding affinity and specificity in protein–DNA complexes is thus an important goal. Here we develop a simple physical energy function, which uses electrostatics, solvation, hydrogen bonds and atom-packing terms to model direct readout and sequence-specific DNA conformational energy to model indirect readout of DNA sequence by the bound protein. The predictive capability of the model is tested against another model based only on the knowledge of the consensus sequence and the number of contacts between amino acids and DNA bases. Both models are used to carry out predictions of protein–DNA binding affinities which are then compared with experimental measurements. The nearly additive nature of protein–DNA interaction energies in our model allows us to construct position-specific weight matrices by computing base pair probabilities independently for each position in the binding site. Our approach is less data intensive than knowledge-based models of protein–DNA interactions, and is not limited to any specific family of transcription factors. However, native structures of protein–DNA complexes or their close homologs are required as input to the model. Use of homology modeling can significantly increase the extent of our approach, making it a useful tool for studying regulatory pathways in many organisms and cell types.

Discrimination of yeast genes involved in methionine and phosphate metabolism on the basis of upstream motifs

In yeast, methionine and phosphate metabolism are regulated by the complexes Met4p/Met28p/Cbf1p and Pho4p, respectively. The binding sites for these factors share a common core CACGTG. We evaluate our capability to discriminate phosphate- and methionine-responding genes on the basis of putative regulatory elements, despite the similarity between Met4p/Met28p/Cbf1p and Pho4p consensus. We scanned upstream regions of methionine, phosphate and control genes with position-specific weight matrices for Pho4p, Met4p/Met28p/Cbf1p and Met31p/Met32p, and applied discriminant analysis to classify genes according to matrix matching scores. This analysis showed that matrix scores provided a good discrimination between phosphate, methionine and control genes. The optimal parameters have then been used to predict phosphate and methionine regulation at a genome scale. The genome-scale analysis predicts 37 genes as methionine-regulated and 40 as phosphate-regulated. We compare the predictive results with high throughput data and discuss the difference. The programs for sequence retrieval and analysis, as well as the complete data and results, are available on the website on regulatory sequence analysis tools (http://rsat.scmbb.ulb.ac.be/rsat/). jvanheld@scmbb.ulb.ac.be The complete datasets and results are available at http://rsat.scmbb.ulb.ac.be/rsat/data/published_data/Gonze_MET_PHO/

Improvement of TRANSFAC matrices using multiple local alignment of transcription factor binding site sequences.

This paper describes a novel approach to constructing Position-Specific Weight Matrices (PWMs) based on the transcription factor binding site (TFBS) data provide by the TRANSFAC database and comparison of the newly generated PWMs with the original TRANSFAC matrices. Multiple local sequence alignment was performed on the TFBSs of each transcription factor. Several different alignment programs were tested and their matrices were compared to the original TRANSFAC matrices. One of the alignment programs, GLAM, produced comparable matrices in terms of the average ranking of true positive sites across the whole test set of sequences.

Bioinformatic identification of novel regulatory DNA sequence motifs in Streptomyces coelicolor

BackgroundStreptomyces coelicolor is a bacterium with a vast repertoire of metabolic functions and complex systems of cellular development. Its genome sequence is rich in genes that encode regulatory proteins to control these processes in response to its changing environment. We wished to apply a recently published bioinformatic method for identifying novel regulatory sequence signals to gain new insights into regulation in S. coelicolor.ResultsThe method involved production of position-specific weight matrices from alignments of over-represented words of DNA sequence. We generated 2497 weight matrices, each representing a candidate regulatory DNA sequence motif. We scanned the genome sequence of S. coelicolor against each of these matrices. A DNA sequence motif represented by one of the matrices was found preferentially in non-coding sequences immediately upstream of genes involved in polysaccharide degradation, including several that encode chitinases. This motif (TGGTCTAGACCA) was also found upstream of genes encoding components of the phosphoenolpyruvate phosphotransfer system (PTS). We hypothesise that this DNA sequence motif represents a regulatory element that is responsive to availability of carbon-sources.Other motifs of potential biological significance were found upstream of genes implicated in secondary metabolism (TTAGGTtAGgCTaACCTAA), sigma factors (TGACN19TGAC), DNA replication and repair (ttgtCAGTGN13TGGA), nucleotide conversions (CTACgcNCGTAG), and ArsR (TCAGN12TCAG). A motif found upstream of genes involved in chromosome replication (TGTCagtgcN7Tagg) was similar to a previously described motif found in UV-responsive promoters.ConclusionsWe successfully applied a recently published in silico method to identify conserved sequence motifs in S. coelicolor that may be biologically significant as regulatory elements. Our data are broadly consistent with and further extend data from previously published studies. We invite experimental testing of our hypotheses in vitro and in vivo.

Identification of the binding sites of regulatory proteins in bacterial genomes.

We present an algorithm that extracts the binding sites (represented by position-specific weight matrices) for many different transcription factors from the regulatory regions of a genome, without the need for delineating groups of coregulated genes. The algorithm uses the fact that many DNA-binding proteins in bacteria bind to a bipartite motif with two short segments more conserved than the intervening region. It identifies all statistically significant patterns of the form W(1)N(x)W(2), where W(1) and W(2) are two short oligonucleotides separated by x arbitrary bases, and groups them into clusters of similar patterns. These clusters are then used to derive quantitative recognition profiles of putative regulatory proteins. For a given cluster, the algorithm finds the matching sequences plus the flanking regions in the genome and performs a multiple sequence alignment to derive position-specific weight matrices. We have analyzed the Escherichia coli genome with this algorithm and found approximately 1,500 significant patterns, which give rise to approximately 160 distinct position-specific weight matrices. A fraction of these matrices match the binding sites of one-third of the approximately 60 characterized transcription factors with high statistical significance. Many of the remaining matrices are likely to describe binding sites and regulons of uncharacterized transcription factors. The significance of these matrices was evaluated by their specificity, the location of the predicted sites, and the biological functions of the corresponding regulons, allowing us to suggest putative regulatory functions. The algorithm is efficient for analyzing newly sequenced bacterial genomes for which little is known about transcriptional regulation.