Locating mammalian transcription factor binding sites: a survey of computational and experimental techniques

Functional analysis of transcription factor binding sites in human promoters

Background

The binding of transcription factors to specific locations in the genome is integral to the orchestration of transcriptional regulation in cells. To characterize transcription factor binding site function on a large scale, we predicted and mutagenized 455 binding sites in human promoters. We carried out functional tests on these sites in four different immortalized human cell lines using transient transfections with a luciferase reporter assay, primarily for the transcription factors CTCF, GABP, GATA2, E2F, STAT, and YY1.

Results

In each cell line, between 36% and 49% of binding sites made a functional contribution to the promoter activity; the overall rate for observing function in any of the cell lines was 70%. Transcription factor binding resulted in transcriptional repression in more than a third of functional sites. When compared with predicted binding sites whose function was not experimentally verified, the functional binding sites had higher conservation and were located closer to transcriptional start sites (TSSs). Among functional sites, repressive sites tended to be located further from TSSs than were activating sites. Our data provide significant insight into the functional characteristics of YY1 binding sites, most notably the detection of distinct activating and repressing classes of YY1 binding sites. Repressing sites were located closer to, and often overlapped with, translational start sites and presented a distinctive variation on the canonical YY1 binding motif.

Conclusions

The genomic properties that we found to associate with functional TF binding sites on promoters -- conservation, TSS proximity, motifs and their variations -- point the way to improved accuracy in future TFBS predictions.

Uncovering transcription factor modules using one- and three-dimensional analyses

Transcriptional regulation is a critical mediator of many normal cellular processes, as well as disease progression. Transcription factors (TFs) often co-localize at cis-regulatory elements on the DNA, form protein complexes, and collaboratively regulate gene expression. Machine learning and Bayesian approaches have been used to identify TF modules in a one-dimensional context. However, recent studies using high throughput technologies have shown that TF interactions should also be considered in three-dimensional nuclear space. Here, we describe methods for identifying TF modules and discuss how moving from a one-dimensional to a three-dimensional paradigm, along with integrated experimental and computational approaches, can lead to a better understanding of TF association networks.

Genome-wide epigenetic data facilitate understanding of disease susceptibility association studies

Complex traits such as susceptibility to diseases are determined in part by variants at multiple genetic loci. Genome-wide association studies can identify these loci, but most phenotype-associated variants lie distal to protein-coding regions and are likely involved in regulating gene expression. Understanding how these genetic variants affect complex traits depends on the ability to predict and test the function of the genomic elements harboring them. Community efforts such as the ENCODE Project provide a wealth of data about epigenetic features associated with gene regulation. These data enable the prediction of testable functions for many phenotype-associated variants.

Deciphering the transcriptional regulation of microRNA genes in humans with ACTLocater

Understanding the transcriptional regulation of microRNAs (miRNAs) is extremely important for determining the specific roles they play in signaling cascades. However, precise identification of transcription factor binding sites (TFBSs) orchestrating the expressions of miRNAs remains a challenge. By combining accessible chromatin sequences of 12 cell types released by the ENCODE Project, we found that a significant fraction (∼80%) of such integrated sequences, evolutionary conserved and in regions upstream of human miRNA genes that are independently transcribed, were preserved across cell types. Accordingly, we developed a computational method, Accessible and Conserved TFBSs Locater (ACTLocater), incorporating this chromatin feature and evolutionary conservation to identify the TFBSs associated with human miRNA genes. ACTLocater achieved high positive predictive values, as revealed by the experimental validation of FOXA1 predictions and by the comparison of its predictions of some other transcription factors (TFs) to empirical ChIP-seq data. Most notably, ACTLocater was widely applicable as indicated by the successful prediction of TF→miRNA interactions in cell types whose chromatin accessibility profiles were not incorporated. By applying ACTLocater to TFs with characterized binding specificities, we compiled a novel repository of putative TF→miRNA interactions and displayed it in ACTViewer, providing a promising foundation for future investigations to elucidate the regulatory mechanisms of miRNA transcription in humans.

Genomic approaches towards finding cis-regulatory modules in animals

Differential gene expression is the fundamental mechanism underlying animal development and cell differentiation. However, it is a challenge to identify comprehensively and accurately the DNA sequences that are required to regulate gene expression: namely, cis-regulatory modules (CRMs). Three major features, either singly or in combination, are used to predict CRMs: clusters of transcription factor binding site motifs, non-coding DNA that is under evolutionary constraint and biochemical marks associated with CRMs, such as histone modifications and protein occupancy. The validation rates for predictions indicate that identifying diagnostic biochemical marks is the most reliable method, and understanding is enhanced by the analysis of motifs and conservation patterns within those predicted CRMs.

A niched Pareto genetic algorithm for finding variable length regulatory motifs in DNA sequences

The transcription factor binding sites also called as motifs are short, recurring patterns in DNA sequences that are presumed to have a biological function. Identification of the motifs from the promoter region of the genes is an important and unsolved problem specifically in the eukaryotic genomes. In this paper, we present a niched Pareto genetic algorithm to identify the regulatory motifs. This approach is based on the maximization of two objectives of the problem that is the motif length and the consensus similarity score. A long motif means it is less likely to be a false motif. The similarity score represents a motifs probability of conservation in a given set of sequences. Proposed method can find multiple, variable length motifs. In this method, we represented a candidate motif as a combination of length and starting position of the motif in each sequence of the co-regulated genes. This enables the algorithm to identify multiple motifs of variable length. We applied this approach on various data sets and the results show that it can find multiple motifs of variable length in co-regulated genes.

Identifying novel transcriptional components controlling energy metabolism

The investigation of metabolic regulation at the transcriptional level presents different challenges than those encountered in the study of other important problems like development or cancer. Levels of key components like glucose, insulin, and lipids can be modulated but rarely change in an all-or-none fashion, necessitating quantitative techniques that can be applied to multiple tissues and systems. This review examines recent advances in methods for studying transcriptional regulation, with special emphasis on metabolic science. We compare these methods for investigators trying to decide on the best approach for their particular physiological paradigm or model system.

Assessing the effects of symmetry on motif discovery and modeling

Background

Identifying the DNA binding sites for transcription factors is a key task in modeling the gene regulatory network of a cell. Predicting DNA binding sites computationally suffers from high false positives and false negatives due to various contributing factors, including the inaccurate models for transcription factor specificity. One source of inaccuracy in the specificity models is the assumption of asymmetry for symmetric models.

Methodology/Principal Findings

Using simulation studies, so that the correct binding site model is known and various parameters of the process can be systematically controlled, we test different motif finding algorithms on both symmetric and asymmetric binding site data. We show that if the true binding site is asymmetric the results are unambiguous and the asymmetric model is clearly superior to the symmetric model. But if the true binding specificity is symmetric commonly used methods can infer, incorrectly, that the motif is asymmetric. The resulting inaccurate motifs lead to lower sensitivity and specificity than would the correct, symmetric models. We also show how the correct model can be obtained by the use of appropriate measures of statistical significance.

Conclusions/Significance

This study demonstrates that the most commonly used motif-finding approaches usually model symmetric motifs incorrectly, which leads to higher than necessary false prediction errors. It also demonstrates how alternative motif-finding methods can correct the problem, providing more accurate motif models and reducing the errors. Furthermore, it provides criteria for determining whether a symmetric or asymmetric model is the most appropriate for any experimental dataset.

Probabilistic models for semisupervised discriminative motif discovery in DNA sequences

Methods for discriminative motif discovery in DNA sequences identify transcription factor binding sites (TFBSs), searching only for patterns that differentiate two sets (positive and negative sets) of sequences. On one hand, discriminative methods increase the sensitivity and specificity of motif discovery, compared to generative models. On the other hand, generative models can easily exploit unlabeled sequences to better detect functional motifs when labeled training samples are limited. In this paper, we develop a hybrid generative/discriminative model which enables us to make use of unlabeled sequences in the framework of discriminative motif discovery, leading to semisupervised discriminative motif discovery. Numerical experiments on yeast ChIP-chip data for discovering DNA motifs demonstrate that the best performance is obtained between the purely-generative and the purely-discriminative and the semisupervised learning improves the performance when labeled sequences are limited.

Identifying functional single nucleotide polymorphisms in the human CArGome

Regulatory SNPs (rSNPs) reside primarily within the nonprotein coding genome and are thought to disturb normal patterns of gene expression by altering DNA binding of transcription factors. Nevertheless, despite the explosive rise in SNP association studies, there is little information as to the function of rSNPs in human disease. Serum response factor (SRF) is a widely expressed DNA-binding transcription factor that has variable affinity to at least 1,216 permutations of a 10 bp transcription factor binding site (TFBS) known as the CArG box. We developed a robust in silico bioinformatics screening method to evaluate sequences around RefSeq genes for conserved CArG boxes. Utilizing a predetermined phastCons threshold score, we identified 8,252 strand-specific CArGs within an 8 kb window around the transcription start site of 5,213 genes, including all previously defined SRF target genes. We then interrogated this CArG dataset for the presence of previously annotated common polymorphisms. We found a total of 118 unique CArG boxes harboring a SNP within the 10 bp CArG sequence and 1,130 CArG boxes with SNPs located just outside the CArG element. Gel shift and luciferase reporter assays validated SRF binding and functional activity of several new CArG boxes. Importantly, SNPs within or just outside the CArG box often resulted in altered SRF binding and activity. Collectively, these findings demonstrate a powerful approach to computationally define rSNPs in the human CArGome and provide a foundation for similar analyses of other TFBS. Such information may find utility in genetic association studies of human disease where little insight is known regarding the functionality of rSNPs.

High resolution detection and analysis of CpG dinucleotides methylation using MBD-Seq technology

Methyl-CpG binding domain protein sequencing (MBD-seq) is widely used to survey DNA methylation patterns. However, the optimal experimental parameters for MBD-seq remain unclear and the data analysis remains challenging. In this study, we generated high depth MBD-seq data in MCF-7 cell and developed a bi-asymmetric-Laplace model (BALM) to perform data analysis. We found that optimal efficiency of MBD-seq experiments was achieved by sequencing ∼100 million unique mapped tags from a combination of 500 mM and 1000 mM salt concentration elution in MCF-7 cells. Clonal bisulfite sequencing results showed that the methylation status of each CpG dinucleotides in the tested regions was accurately detected with high resolution using the proposed model. These results demonstrated the combination of MBD-seq and BALM could serve as a useful tool to investigate DNA methylome due to its low cost, high specificity, efficiency and resolution.

Functional pathway mapping analysis for hypoxia-inducible factors

Background

Hypoxia-inducible factors (HIFs) are transcription factors that play a crucial role in response to hypoxic stress in living organisms. The HIF pathway is activated by changes in cellular oxygen levels and has significant impacts on the regulation of gene expression patterns in cancer cells. Identifying functional conservation across species and discovering conserved regulatory motifs can facilitate the selection of reference species for empirical tests. This paper describes a cross-species functional pathway mapping strategy based on evidence of homologous relationships that employs matrix-based searching techniques for identifying transcription factor-binding sites on all retrieved HIF target genes.

Results

HIF-related orthologous and paralogous genes were mapped onto the conserved pathways to indicate functional conservation across species. Quantitatively measured HIF pathways are depicted in order to illustrate the extent of functional conservation. The results show that in spite of the evolutionary process of speciation, distantly related species may exhibit functional conservation owing to conservative pathways. The novel terms OrthRate and ParaRate are proposed to quantitatively indicate the flexibility of a homologous pathway and reveal the alternative regulation of functional genes.

Conclusion

The developed functional pathway mapping strategy provides a bioinformatics approach for constructing biological pathways by highlighting the homologous relationships between various model species. The mapped HIF pathways were quantitatively illustrated and evaluated by statistically analyzing their conserved transcription factor-binding elements.

Keywords

hypoxia-inducible factor (HIF), hypoxia-response element (HRE), transcription factor (TF), transcription factor binding site (TFBS), KEGG (Kyoto Encyclopedia of Genes and Genomes), cross-species comparison, orthology, paralogy, functional pathway

In silico characterization of the neural alpha tubulin gene promoter of the sea urchin embryo Paracentrotus lividus by phylogenetic footprinting

During Paracentrotus lividus sea urchin embryo development one alpha and one beta tubulin genes are expressed specifically in the neural cells and they are early end output of the gene regulatory network that specifies the neural commitment. In this paper we have used a comparative genomics approach to identify conserved regulatory elements in the P. lividus neural alpha tubulin gene. To this purpose, we have first isolated a genomic clone containing the entire gene plus 4.5 Kb of 5' upstream sequences. Then, we have shown by gene transfer experiments that its non-coding region drives the spatio-temporal gene expression corresponding substantially to that of the endogenous gene. In addition, we have identified by genome and EST sequence analysis the S. purpuratus alpha tubulin orthologous gene and we propose a revised annotation of some tubulin family members. Moreover, by computational techniques we delineate at least three putative regulatory regions located both in the upstream region and in the first intron containing putative binding sites for Forkhead and Nkx transcription factor families.

On weight matrix and free energy models for sequence motif detection

The problem of motif detection can be formulated as the construction of a discriminant function to separate sequences of a specific pattern from background. In computational biology, motif detection is used to predict DNA binding sites of a transcription factor (TF), mostly based on the weight matrix (WM) model or the Gibbs free energy (FE) model. However, despite the wide applications, theoretical analysis of these two models and their predictions is still lacking. We derive asymptotic error rates of prediction procedures based on these models under different data generation assumptions. This allows a theoretical comparison between the WM-based and the FE-based predictions in terms of asymptotic efficiency. Applications of the theoretical results are demonstrated with empirical studies on ChIP-seq data and protein binding microarray data. We find that, irrespective of underlying data generation mechanisms, the FE approach shows higher or comparable predictive power relative to the WM approach when the number of observed binding sites used for constructing a discriminant decision is not too small.

Inflammatory gene regulatory networks in amnion cells following cytokine stimulation: translational systems approach to modeling human parturition

A majority of the studies examining the molecular regulation of human labor have been conducted using single gene approaches. While the technology to produce multi-dimensional datasets is readily available, the means for facile analysis of such data are limited. The objective of this study was to develop a systems approach to infer regulatory mechanisms governing global gene expression in cytokine-challenged cells in vitro, and to apply these methods to predict gene regulatory networks (GRNs) in intrauterine tissues during term parturition. To this end, microarray analysis was applied to human amnion mesenchymal cells (AMCs) stimulated with interleukin-1β, and differentially expressed transcripts were subjected to hierarchical clustering, temporal expression profiling, and motif enrichment analysis, from which a GRN was constructed. These methods were then applied to fetal membrane specimens collected in the absence or presence of spontaneous term labor. Analysis of cytokine-responsive genes in AMCs revealed a sterile immune response signature, with promoters enriched in response elements for several inflammation-associated transcription factors. In comparison to the fetal membrane dataset, there were 34 genes commonly upregulated, many of which were part of an acute inflammation gene expression signature. Binding motifs for nuclear factor-κB were prominent in the gene interaction and regulatory networks for both datasets; however, we found little evidence to support the utilization of pathogen-associated molecular pattern (PAMP) signaling. The tissue specimens were also enriched for transcripts governed by hypoxia-inducible factor. The approach presented here provides an uncomplicated means to infer global relationships among gene clusters involved in cellular responses to labor-associated signals.

To understand the whole, you must know the parts: unraveling the roles of protein-DNA interactions in genome regulation

The regulation of gene transcription is fundamental to the existence of complex multicellular organisms such as humans. This process dictates which genes are expressed in which tissues, and controls how various cell types grow, differentiate, and respond to their environments. Although the deciphering of the human genome sequence has given us the "source code" for life, we still know far too little about the mechanisms that control which sets of genes are active in which tissues, and how their expression is regulated. It is clear, however, that much of this system depends upon the sequence-specific interactions of regulatory proteins with particular genetic loci. To be able to unravel the details of these interactions on a genome-wide basis, it is necessary to know what proteins are bound to the DNA where in the genome, and to be able to monitor how those proteins change over time and in response to external stimuli. Developing a new technology to provide this information constitutes a "Grand Challenge" for Analytical Chemistry. In this brief article we outline the nature of this challenge, and propose one strategy to address it.

Distinct interfacial biclique patterns between ssDNA-binding proteins and those with dsDNAs

We introduce a new motif called interfacial biclique pattern to study the difference between double-stranded DNA-binding proteins (DSBs, most of them also known to play the role as transcriptional factors) and single-stranded DNA-binding proteins (SSBs) which are found to involve in many applications recently. An interfacial biclique pattern in a protein-DNA complex usually consists of a group of residues and a group of nucleotides such that every residue has a contact to all of the bases. The proposal of this idea is based on a biological redundancy mechanism that: a site mutation has little influence on the other residues to recognize the target nucleotides and vice versa. The distribution of the residues on the interfacial motifs is investigated to identify distinct stable preferred residues, stable un-preferred residues and unstable preferred residues between SSBs and DSBs. We also examine residue co-occurrence and residue-base association rules in the interfacial motifs to uncover the different choices of residue combinations by SSBs and DSBs that have contacts with one or more bases. We found that DSBs and SSBs have their own right residues at the right places for the binding preference and association with nucleotides. Some of our results can be supported by literature work.

A ChIP-Seq benchmark shows that sequence conservation mainly improves detection of strong transcription factor binding sites

BACKGROUND

Transcription factors are important controllers of gene expression and mapping transcription factor binding sites (TFBS) is key to inferring transcription factor regulatory networks. Several methods for predicting TFBS exist, but there are no standard genome-wide datasets on which to assess the performance of these prediction methods. Also, it is believed that information about sequence conservation across different genomes can generally improve accuracy of motif-based predictors, but it is not clear under what circumstances use of conservation is most beneficial.

RESULTS

Here we use published ChIP-seq data and an improved peak detection method to create comprehensive benchmark datasets for prediction methods which use known descriptors or binding motifs to detect TFBS in genomic sequences. We use this benchmark to assess the performance of five different prediction methods and find that the methods that use information about sequence conservation generally perform better than simpler motif-scanning methods. The difference is greater on high-affinity peaks and when using short and information-poor motifs. However, if the motifs are specific and information-rich, we find that simple motif-scanning methods can perform better than conservation-based methods.

CONCLUSIONS

Our benchmark provides a comprehensive test that can be used to rank the relative performance of transcription factor binding site prediction methods. Moreover, our results show that, contrary to previous reports, sequence conservation is better suited for predicting strong than weak transcription factor binding sites.

Novel computational analysis of protein binding array data identifies direct targets of Nkx2.2 in the pancreas

Background

The creation of a complete genome-wide map of transcription factor binding sites is essential for understanding gene regulatory networks in vivo. However, current prediction methods generally rely on statistical models that imperfectly model transcription factor binding. Generation of new prediction methods that are based on protein binding data, but do not rely on these models may improve prediction sensitivity and specificity.

Results

We propose a method for predicting transcription factor binding sites in the genome by directly mapping data generated from protein binding microarrays (PBM) to the genome and calculating a moving average of several overlapping octamers. Using this unique algorithm, we predicted binding sites for the essential pancreatic islet transcription factor Nkx2.2 in the mouse genome and confirmed >90% of the tested sites by EMSA and ChIP. Scores generated from this method more accurately predicted relative binding affinity than PWM based methods. We have also identified an alternative core sequence recognized by the Nkx2.2 homeodomain. Furthermore, we have shown that this method correctly identified binding sites in the promoters of two critical pancreatic islet β-cell genes, NeuroD1 and insulin2, that were not predicted by traditional methods. Finally, we show evidence that the algorithm can also be applied to predict binding sites for the nuclear receptor Hnf4α.

Conclusions

PBM-mapping is an accurate method for predicting Nkx2.2 binding sites and may be widely applicable for the creation of genome-wide maps of transcription factor binding sites.

In silico prediction of transcriptional factor-binding sites

The recognition of transcription factor binding sites (TFBSs) is the first step on the way to deciphering the DNA regulatory code. A large variety of computational approaches and corresponding in silico tools for TFBS recognition are available, each having their own advantages and shortcomings. This chapter provides a brief tutorial to assist end users in the application of these tools for functional characterization of genes.

When needles look like hay: how to find tissue-specific enhancers in model organism genomes

A major prerequisite for the investigation of tissue-specific processes is the identification of cis-regulatory elements. No generally applicable technique is available to distinguish them from any other type of genomic non-coding sequence. Therefore, researchers often have to identify these elements by elaborate in vivo screens, testing individual regions until the right one is found. Here, based on many examples from the literature, we summarize how functional enhancers have been isolated from other elements in the genome and how they have been characterized in transgenic animals. Covering computational and experimental studies, we provide an overview of the global properties of cis-regulatory elements, like their specific interactions with promoters and target gene distances. We describe conserved non-coding elements (CNEs) and their internal structure, nucleotide composition, binding site clustering and overlap, with a special focus on developmental enhancers. Conflicting data and unresolved questions on the nature of these elements are highlighted. Our comprehensive overview of the experimental shortcuts that have been found in the different model organism communities and the new field of high-throughput assays should help during the preparation phase of a screen for enhancers. The review is accompanied by a list of general guidelines for such a project.

Transcription factor binding variation in the evolution of gene regulation

Transcription factor interactions with DNA are one of the primary mechanisms by which expression is modulated, yet their evolution remains poorly understood. Chromatin immunoprecipitation followed by microarray (ChIP-chip) or sequencing (ChIP-Seq) has revolutionized the study of protein-DNA interactions. However, only recently has attention focused on determining to what extent these regulatory interactions vary between species across entire genomes. A series of recent studies have compared in vivo binding data across a range of evolutionary distances. Binding events diverge rapidly, indicating gene regulation is an evolutionarily flexible process.

Experimental strategies for studying transcription factor-DNA binding specificities

Specific binding of transcription factors (TFs) determines in a large part the connectivity of gene regulatory networks as well as the quantitative level of gene expression. A multiplicity of both experimental and computational methods is currently used to discover and characterize the underlying TF-DNA interactions. Experimental methods can be further subdivided into in vitro- and in vivo-based approaches, each accenting different aspects of TF-binding events. In this review we summarize the flexibility and performance of a selection of both types of experimental methods. In conclusion, we argue that a serial combination of methods with different throughput and data type constitutes an optimal experimental strategy.

CircuitsDB: a database of mixed microRNA/transcription factor feed-forward regulatory circuits in human and mouse

BACKGROUND

Transcription Factors (TFs) and microRNAs (miRNAs) are key players for gene expression regulation in higher eukaryotes. In the last years, a large amount of bioinformatic studies were devoted to the elucidation of transcriptional and post-transcriptional (mostly miRNA-mediated) regulatory interactions, but little is known about the interplay between them.

DESCRIPTION

Here we describe a dynamic web-accessible database, CircuitsDB, supporting a genome-wide transcriptional and post-transcriptional regulatory network integration, for the human and mouse genomes, based on a bioinformatic sequence-analysis approach. In particular, CircuitsDB is currently focused on the study of mixed miRNA/TF Feed-Forward regulatory Loops (FFLs), i.e. elementary circuits in which a master TF regulates an miRNA and together with it a set of Joint Target protein-coding genes. The database was constructed using an ab-initio oligo analysis procedure for the identification of the transcriptional and post-transcriptional interactions. Several external sources of information were then pooled together to obtain the functional annotation of the proposed interactions. Results for human and mouse genomes are presented in an integrated web tool, that allows users to explore the circuits, investigate their sequence and functional properties and thus suggest possible biological experiments.

CONCLUSIONS

We present CircuitsDB, a web-server devoted to the study of human and mouse mixed miRNA/TF Feed-Forward regulatory circuits, freely available at: http://biocluster.di.unito.it/circuits/

The current state of chromatin immunoprecipitation

The biological significance of interactions of nuclear proteins with DNA in the context of gene expression, cell differentiation, or disease has immensely been enhanced by the advent of chromatin immunoprecipitation (ChIP). ChIP is a technique whereby a protein of interest is selectively immunoprecipitated from a chromatin preparation to determine the DNA sequences associated with it. ChIP has been widely used to map the localization of post-translationally modified histones, histone variants, transcription factors, or chromatin modifying enzymes on the genome or on a given locus. In spite of its power, ChIP has for a long time remained a cumbersome procedure requiring large numbers of cells. These limitations have sparked the development of modifications to shorten the procedure, simplify sample handling and make ChIP amenable to small numbers of cells. In addition, the combination of ChIP with DNA microarray and high-throughput sequencing technologies has in recent years enabled the profiling of histone modification, histone variants, and transcription factor occupancy on a genome-wide scale. This review highlights the variations on the theme of the ChIP assay, the various detection methods applied downstream of ChIP, and examples of their application.

Comprehensive analysis of the palindromic motif TCTCGCGAGA: a regulatory element of the HNRNPK promoter

Definitive identification of promoters, their cis-regulatory motifs, and their trans-acting proteins requires experimental analysis. To define the HNRNPK promoter and its cognate DNA–protein interactions, we performed a comprehensive study combining experimental approaches, including luciferase reporter gene assays, chromatin immunoprecipitations (ChIP), electrophoretic mobility shift assays (EMSA), and mass spectrometry (MS). We discovered that out of the four potential HNRNPK promoters tested, the one containing the palindromic motif TCTCGCGAGA exhibited the highest activity in a reporter system assay. Although further EMSA and MS analyses, performed to uncover the identity of the palindrome-binding transcription factor, did identify a complex of DNA-binding proteins, neither method unambiguously identified the pertinent direct trans-acting protein(s). ChIP revealed similar chromatin states at the promoters with the palindromic motif and at housekeeping gene promoters. A ChIP survey showed significantly higher recruitment of PARP1, a protein identified by MS as ubiquitously attached to DNA probes, within heterochromatin sites. Computational analyses indicated that this palindrome displays features that mark nucleosome boundaries, causing the surrounding DNA landscape to be constitutively open. Our strategy of diverse approaches facilitated the direct characterization of various molecular properties of HNRNPK promoter bearing the palindromic motif TCTCGCGAGA, despite the obstacles that accompany in vitro methods.

Using the Tet-On system to develop a procedure for extracting transcription factor activation dynamics

The regulation of gene expression by transcription factors through different expression and activation dynamics is an important aspect of genomics and systems biology. Reporter systems using green fluorescent protein (GFP) or luciferase are often used to infer transcription factor dynamics. We recently used an inverse problem solution of GFP reporter profiles to demonstrate that the activation dynamics of a model transcription actor (NF-kappaB) can be reconstructed from GFP data. This approach assumes that the general nature of the transcription factor dynamics is known; however, it is non-trivial to determine the exact nature of the transcription factor dynamics as it often depends upon the cell type and the stimulus used to activate the transcription factor. This, in turn, limits the determination of accurate transcription factor dynamics from reporter data, especially since the model used for solution of an inverse problem needs to be verified. To address this point, we developed a reporter cell line for expressing GFP using an inducible, artificial transcription factor (tTA) and minimal promoter system. The artificial transcription factor can be activated independent of the cellular regulatory machinery through addition of doxycycline. This allows us to directly control the dynamics of the artificial transcription factor, and thereby, develop a model describing its activation dynamics from reporter data. Our experimental data and model predictions are in good agreement, and illustrate the utility of our approach. Future work will focus on using the developed approach, i.e. solution of an inverse problem involving the model describing expression of GFP, to extract the dynamics of transcription factors that are currently uncharacterized.

Genome-wide identification of hypoxia-inducible factor binding sites and target genes by a probabilistic model integrating transcription-profiling data and in silico binding site prediction

The transcriptional response driven by Hypoxia-inducible factor (HIF) is central to the adaptation to oxygen restriction. Hence, the complete identification of HIF targets is essential for understanding the cellular responses to hypoxia. Herein we describe a computational strategy based on the combination of phylogenetic footprinting and transcription profiling meta-analysis for the identification of HIF-target genes. Comparison of the resulting candidates with published HIF1a genome-wide chromatin immunoprecipitation indicates a high sensitivity (78%) and specificity (97.8%). To validate our strategy, we performed HIF1a chromatin immunoprecipitation on a set of putative targets. Our results confirm the robustness of the computational strategy in predicting HIF-binding sites and reveal several novel HIF targets, including RE1-silencing transcription factor co-repressor (RCOR2). In addition, mapping of described polymorphisms to the predicted HIF-binding sites identified several single-nucleotide polymorphisms (SNPs) that could alter HIF binding. As a proof of principle, we demonstrate that SNP rs17004038, mapping to a functional hypoxia response element in the macrophage migration inhibitory factor (MIF) locus, prevents induction of this gene by hypoxia. Altogether, our results show that the proposed strategy is a powerful tool for the identification of HIF direct targets that expands our knowledge of the cellular adaptation to hypoxia and provides cues on the inter-individual variation in this response.

Gene expression mining in type 2 diabetes research

Microarray analysis has become a core part of biomedical research and its value can be seen in thousands of research papers. A successful microarray experiment needs to be augmented by specialized data mining techniques if the data are to be fully exploited. Here, tools that concentrate on three areas--gene enrichment analysis, literature mining, and transcription factor binding site analysis--are described for the novice user of microarray technology. The focus of this chapter is on free, publicly available, web-based tools.

Scoring overlapping and adjacent signals from genome-wide ChIP and DamID assays

Much of the research utilising genome-wide ChIP and DamID assays aims to understand the combinatorial feature of transcription factor binding and the chromatin modification code. With these experimental methods becoming more affordable and widespread, the focus of research is shifting to making sense of the data. Amongst the many challenges arising from data analyses, we are concerned with identifying biologically meaningful co-occurrences of transcription factor binding or chromatin modifications, using genome-wide profiles generated from ChIP and DamID assays. Co-occurrences are reflected in overlapping and adjacent signals in multiple ChIP or DamID profiles. We review existing quantitative methods to score overlaps and to cluster binding events in ChIP and DamID profiles. For pairwise comparison, existing methods either are based on a single score at the genome level or take a genomic, region-specific view. To draw inference from many profiles simultaneously, methods exist to cluster regions by their regulatory importance or to infer cis-regulatory modules for a particular region. We provide a simple guide to some of the statistical tools used by these methods.

Seed storage protein gene promoters contain conserved DNA motifs in Brassicaceae, Fabaceae and Poaceae

BACKGROUND

Accurate computational identification of cis-regulatory motifs is difficult, particularly in eukaryotic promoters, which typically contain multiple short and degenerate DNA sequences bound by several interacting factors. Enrichment in combinations of rare motifs in the promoter sequence of functionally or evolutionarily related genes among several species is an indicator of conserved transcriptional regulatory mechanisms. This provides a basis for the computational identification of cis-regulatory motifs.

RESULTS

We have used a discriminative seeding DNA motif discovery algorithm for an in-depth analysis of 54 seed storage protein (SSP) gene promoters from three plant families, namely Brassicaceae (mustards), Fabaceae (legumes) and Poaceae (grasses) using backgrounds based on complete sets of promoters from a representative species in each family, namely Arabidopsis (Arabidopsis thaliana (L.) Heynh.), soybean (Glycine max (L.) Merr.) and rice (Oryza sativa L.) respectively. We have identified three conserved motifs (two RY-like and one ACGT-like) in Brassicaceae and Fabaceae SSP gene promoters that are similar to experimentally characterized seed-specific cis-regulatory elements. Fabaceae SSP gene promoter sequences are also enriched in a novel, seed-specific E2Fb-like motif. Conserved motifs identified in Poaceae SSP gene promoters include a GCN4-like motif, two prolamin-box-like motifs and an Skn-1-like motif. Evidence of the presence of a variant of the TATA-box is found in the SSP gene promoters from the three plant families. Motifs discovered in SSP gene promoters were used to score whole-genome sets of promoters from Arabidopsis, soybean and rice. The highest-scoring promoters are associated with genes coding for different subunits or precursors of seed storage proteins.

CONCLUSION

Seed storage protein gene promoter motifs are conserved in diverse species, and different plant families are characterized by a distinct combination of conserved motifs. The majority of discovered motifs match experimentally characterized cis-regulatory elements. These results provide a good starting point for further experimental analysis of plant seed-specific promoters and our methodology can be used to unravel more transcriptional regulatory mechanisms in plants and other eukaryotes.

Comparison of melanoblast expression patterns identifies distinct classes of genes

A full understanding of transcriptional regulation requires integration of information obtained from multiple experimental datasets. These include datasets annotating gene expression within the context of an entire organism under normal and genetically perturbed conditions. Here we describe an expression dataset annotating pigment cell-expressed genes of the developing melanocyte and retinal pigmented epithelium lineages. Expression images are annotated and available at http://research.nhgri.nih.gov/manuscripts/Loftus/March2009/. Data are also summarized in a standardized manner using a universal melanoblast scoring scale that accounts for the embryonic location of cells and regional cell density. This approach allowed us to classify 14 pigment genes into four groupings classified by cell lineage expression, temporal-spatial context, and differential alteration in response to altered MITF and SOX10 status. Significant differences in regional populations were also observed across inbred strain backgrounds, highlighting the value of this approach to identify modifier allele influences on melanoblast number and distributions. This analysis revealed novel features of in vivo expression patterns that are not measurable by in vitro-based assays, providing data that in combination with genomic analyses will allow modeling of pigment cell gene expression in development and disease.

Prediction and experimental validation of novel STAT3 target genes in human cancer cells

The comprehensive identification of functional transcription factor binding sites (TFBSs) is an important step in understanding complex transcriptional regulatory networks. This study presents a motif-based comparative approach, STAT-Finder, for identifying functional DNA binding sites of STAT3 transcription factor. STAT-Finder combines STAT-Scanner, which was designed to predict functional STAT TFBSs with improved sensitivity, and a motif-based alignment to minimize false positive prediction rates. Using two reference sets containing promoter sequences of known STAT3 target genes, STAT-Finder identified functional STAT3 TFBSs with enhanced prediction efficiency and sensitivity relative to other conventional TFBS prediction tools. In addition, STAT-Finder identified novel STAT3 target genes among a group of genes that are over-expressed in human cancer cells. The binding of STAT3 to the predicted TFBSs was also experimentally confirmed through chromatin immunoprecipitation. Our proposed method provides a systematic approach to the prediction of functional TFBSs that can be applied to other TFs.

Modeling the quantitative specificity of DNA-binding proteins from example binding sites

BACKGROUND

The binding of transcription factors to their respective DNA sites is a key component of every regulatory network. Predictions of transcription factor binding sites are usually based on models for transcription factor specificity. These models, in turn, are often based on examples of known binding sites.

METHODOLOGY/PRINCIPAL FINDINGS

Collections of binding sites are obtained in simulation experiments where the true model for the transcription factor is known and various sampling procedures are employed. We compare the accuracies of three different and commonly used methods for predicting the specificity of the transcription factor based on example binding sites. Different methods for constructing the models can lead to significant differences in the accuracy of the predictions and we show that commonly used methods can be positively misleading, even at large sample sizes and using noise-free data. Methods that minimize the number of predicted binding sequences are often significantly more accurate than the other methods tested.

CONCLUSIONS/SIGNIFICANCE

Different methods for generating motifs from example binding sites can have significantly different numbers of false positive and false negative predictions. For many different sampling procedures models based on quadratic programming are the most accurate.

Insights from genomic profiling of transcription factors

A crucial question in the field of gene regulation is whether the location at which a transcription factor binds influences its effectiveness or the mechanism by which it regulates transcription. Comprehensive transcription factor binding maps are needed to address these issues, and genome-wide mapping is now possible thanks to the technological advances of ChIP-chip and ChIP-seq. This Review discusses how recent genomic profiling of transcription factors gives insight into how binding specificity is achieved and what features of chromatin influence the ability of transcription factors to interact with the genome. It also suggests future experiments that may further our understanding of the causes and consequences of transcription factor-genome interactions.

Genome-wide survey of microRNA-transcription factor feed-forward regulatory circuits in human

In this work, we describe a computational framework for the genome-wide identification and characterization of mixed transcriptional/post-transcriptional regulatory circuits in humans.

In this work, we describe a computational framework for the genome-wide identification and characterization of mixed transcriptional/post-transcriptional regulatory circuits in humans. We concentrated in particular on feed-forward loops (FFL), in which a master transcription factor regulates a microRNA, and together with it, a set of joint target protein coding genes. The circuits were assembled with a two step procedure. We first constructed separately the transcriptional and post-transcriptional components of the human regulatory network by looking for conserved over-represented motifs in human and mouse promoters, and 3′-UTRs. Then, we combined the two subnetworks looking for mixed feed-forward regulatory interactions, finding a total of 638 putative (merged) FFLs. In order to investigate their biological relevance, we filtered these circuits using three selection criteria: (I) GeneOntology enrichment among the joint targets of the FFL, (II) independent computational evidence for the regulatory interactions of the FFL, extracted from external databases, and (III) relevance of the FFL in cancer. Most of the selected FFLs seem to be involved in various aspects of organism development and differentiation. We finally discuss a few of the most interesting cases in detail.

DISCOVER: a feature-based discriminative method for motif search in complex genomes

MOTIVATION

Identifying transcription factor binding sites (TFBSs) encoding complex regulatory signals in metazoan genomes remains a challenging problem in computational genomics. Due to degeneracy of nucleotide content among binding site instances or motifs, and intricate 'grammatical organization' of motifs within cis-regulatory modules (CRMs), extant pattern matching-based in silico motif search methods often suffer from impractically high false positive rates, especially in the context of analyzing large genomic datasets, and noisy position weight matrices which characterize binding sites. Here, we try to address this problem by using a framework to maximally utilize the information content of the genomic DNA in the region of query, taking cues from values of various biologically meaningful genetic and epigenetic factors in the query region such as clade-specific evolutionary parameters, presence/absence of nearby coding regions, etc. We present a new method for TFBS prediction in metazoan genomes that utilizes both the CRM architecture of sequences and a variety of features of individual motifs. Our proposed approach is based on a discriminative probabilistic model known as conditional random fields that explicitly optimizes the predictive probability of motif presence in large sequences, based on the joint effect of all such features.

RESULTS

This model overcomes weaknesses in earlier methods based on less effective statistical formalisms that are sensitive to spurious signals in the data. We evaluate our method on both simulated CRMs and real Drosophila sequences in comparison with a wide spectrum of existing models, and outperform the state of the art by 22% in F1 score.

AVAILABILITY AND IMPLEMENTATION

The code is publicly available at http://www.sailing.cs.cmu.edu/discover.html.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

SNPinfo: integrating GWAS and candidate gene information into functional SNP selection for genetic association studies

We have developed a set of web-based SNP selection tools (freely available at http://www.niehs.nih.gov/snpinfo) where investigators can specify genes or linkage regions and select SNPs based on GWAS results, linkage disequilibrium (LD), and predicted functional characteristics of both coding and non-coding SNPs. The algorithm uses GWAS SNP P-value data and finds all SNPs in high LD with GWAS SNPs, so that selection is from a much larger set of SNPs than the GWAS itself. The program can also identify and choose tag SNPs for SNPs not in high LD with any GWAS SNP. We incorporate functional predictions of protein structure, gene regulation, splicing and miRNA binding, and consider whether the alternative alleles of a SNP are likely to have differential effects on function. Users can assign weights for different functional categories of SNPs to further tailor SNP selection. The program accounts for LD structure of different populations so that a GWAS study from one ethnic group can be used to choose SNPs for one or more other ethnic groups. Finally, we provide an example using prostate cancer and demonstrate that this algorithm can select a small panel of SNPs that include many of the recently validated prostate cancer SNPs.

Lighting up individual DNA binding proteins with quantum dots

The ability to determine the precise loci and occupancy of DNA-binding proteins is instrumental to our understanding of cellular processes like gene expression and regulation. We propose a single-molecule approach for the direct visualization of proteins bound to their template DNA. Fluorescent quantum dots (QD) are used to label proteins bound to DNA, allowing multicolor, nanometer-resolution localization. Protein-DNA complexes are linearly extended and imaged to determine the precise location of the protein binding sites. The method is demonstrated by detecting individual QD-labeled T7-RNA polymerases on the T7 bacteriophage genome. This work demonstrates the potential of this approach to precisely read protein binding position or, alternatively, "write" such information on extended DNA with QDs via sequence-specific molecular recognition.

Enhancement of lipid production using biochemical, genetic and transcription factor engineering approaches

This paper compares three possible strategies for enhanced lipid overproduction in microalgae: the biochemical engineering (BE) approaches, the genetic engineering (GE) approaches, and the transcription factor engineering (TFE) approaches. The BE strategy relies on creating a physiological stress such as nutrient-starvation or high salinity to channel metabolic fluxes to lipid accumulation. The GE strategy exploits our understanding to the lipid metabolic pathway, especially the rate-limiting enzymes, to create a channelling of metabolites to lipid biosynthesis by overexpressing one or more key enzymes in recombinant microalgal strains. The TFE strategy is an emerging technology aiming at enhancing the production of a particular metabolite by means of overexpressing TFs regulating the metabolic pathways involved in the accumulation of target metabolites. Currently, BE approaches are the most established in microalgal lipid production. The TFE is a very promising strategy because it may avoid the inhibitive effects of the BE approaches and the limitation of "secondary bottlenecks" as commonly observed in the GE approaches. However, it is still a novel concept to be investigated systematically.

ChIP-seq: using high-throughput sequencing to discover protein-DNA interactions

Chromatin immunoprecipitation (ChIP) allows specific protein-DNA interactions to be isolated. Combining ChIP with high-throughput sequencing reveals the DNA sequence involved in these interactions. Here, we describe how to perform ChIP-seq starting with whole tissues or cell lines, and ending with millions of short sequencing tags that can be aligned to the reference genome of the species under investigation. We also outline additional procedures to recover ChIP-chip libraries for ChIP-seq and discuss contemporary issues in data analysis.