Genome-wide modeling of transcription kinetics reveals patterns of RNA production delays.

Genes with similar transcriptional activation kinetics can display very different temporal mRNA profiles because of differences in transcription time, degradation rate, and RNA-processing kinetics. Recent studies have shown that a splicing-associated RNA production delay can be significant. To investigate this issue more generally, it is useful to develop methods applicable to genome-wide datasets. We introduce a joint model of transcriptional activation and mRNA accumulation that can be used for inference of transcription rate, RNA production delay, and degradation rate given data from high-throughput sequencing time course experiments. We combine a mechanistic differential equation model with a nonparametric statistical modeling approach allowing us to capture a broad range of activation kinetics, and we use Bayesian parameter estimation to quantify the uncertainty in estimates of the kinetic parameters. We apply the model to data from estrogen receptor α activation in the MCF-7 breast cancer cell line. We use RNA polymerase II ChIP-Seq time course data to characterize transcriptional activation and mRNA-Seq time course data to quantify mature transcripts. We find that 11% of genes with a good signal in the data display a delay of more than 20 min between completing transcription and mature mRNA production. The genes displaying these long delays are significantly more likely to be short. We also find a statistical association between high delay and late intron retention in pre-mRNA data, indicating significant splicing-associated production delays in many genes.

Inference of RNA polymerase II transcription dynamics from chromatin immunoprecipitation time course data.

Gene transcription mediated by RNA polymerase II (pol-II) is a key step in gene expression. The dynamics of pol-II moving along the transcribed region influence the rate and timing of gene expression. In this work, we present a probabilistic model of transcription dynamics which is fitted to pol-II occupancy time course data measured using ChIP-Seq. The model can be used to estimate transcription speed and to infer the temporal pol-II activity profile at the gene promoter. Model parameters are estimated using either maximum likelihood estimation or via Bayesian inference using Markov chain Monte Carlo sampling. The Bayesian approach provides confidence intervals for parameter estimates and allows the use of priors that capture domain knowledge, e.g. the expected range of transcription speeds, based on previous experiments. The model describes the movement of pol-II down the gene body and can be used to identify the time of induction for transcriptionally engaged genes. By clustering the inferred promoter activity time profiles, we are able to determine which genes respond quickly to stimuli and group genes that share activity profiles and may therefore be co-regulated. We apply our methodology to biological data obtained using ChIP-seq to measure pol-II occupancy genome-wide when MCF-7 human breast cancer cells are treated with estradiol (E2). The transcription speeds we obtain agree with those obtained previously for smaller numbers of genes with the advantage that our approach can be applied genome-wide. We validate the biological significance of the pol-II promoter activity clusters by investigating cluster-specific transcription factor binding patterns and determining canonical pathway enrichment. We find that rapidly induced genes are enriched for both estrogen receptor alpha (ERα) and FOXA1 binding in their proximal promoter regions.

Integrative analysis of deep sequencing data identifies estrogen receptor early response genes and links ATAD3B to poor survival in breast cancer.

Identification of responsive genes to an extra-cellular cue enables characterization of pathophysiologically crucial biological processes. Deep sequencing technologies provide a powerful means to identify responsive genes, which creates a need for computational methods able to analyze dynamic and multi-level deep sequencing data. To answer this need we introduce here a data-driven algorithm, SPINLONG, which is designed to search for genes that match the user-defined hypotheses or models. SPINLONG is applicable to various experimental setups measuring several molecular markers in parallel. To demonstrate the SPINLONG approach, we analyzed ChIP-seq data reporting PolII, estrogen receptor α (ERα), H3K4me3 and H2A.Z occupancy at five time points in the MCF-7 breast cancer cell line after estradiol stimulus. We obtained 777 ERa early responsive genes and compared the biological functions of the genes having ERα binding within 20 kb of the transcription start site (TSS) to genes without such binding site. Our results show that the non-genomic action of ERα via the MAPK pathway, instead of direct ERa binding, may be responsible for early cell responses to ERα activation. Our results also indicate that the ERα responsive genes triggered by the genomic pathway are transcribed faster than those without ERα binding sites. The survival analysis of the 777 ERα responsive genes with 150 primary breast cancer tumors and in two independent validation cohorts indicated the ATAD3B gene, which does not have ERα binding site within 20 kb of its TSS, to be significantly associated with poor patient survival.

Detecting and visualizing gene fusions.

In recent years, gene fusions have gained significant recognition as biomarkers. They can assist treatment decisions, are seldom found in normal tissue and are detectable through Next-generation sequencing (NGS) of the transcriptome (RNA-seq). To transform the data provided by the sequencer into robust gene fusion detection several analysis steps are needed. Usually the first step is to map the sequenced transcript fragments (RNA-seq) to a reference genome. One standard application of this approach is to estimate expression and detect variants within known genes, e.g. SNPs and indels. In case of gene fusions, however, completely novel gene structures have to be detected. Here, we describe the detection of such gene fusion events based on our comprehensive transcript annotation (ElDorado). To demonstrate the utility of our approach, we extract gene fusion candidates from eight breast cancer cell lines, which we compare to experimentally verified gene fusions. We discuss several gene fusion events, like BCAS3-BCAS4 that was only detected in the breast cancer cell line MCF7. As supporting evidence we show that gene fusions occur more frequently in copy number enriched regions (CNV analysis). In addition, we present the Transcriptome Viewer (TViewer) a tool that allows to interactively visualize gene fusions. Finally, we support detected gene fusions through literature mining based annotations and network analyses. In conclusion, we present a platform that allows detecting gene fusions and supporting them through literature knowledge as well as rich visualization capabilities. This enables scientists to better understand molecular processes, biological functions and disease associations, which will ultimately lead to better biomedical knowledge for the development of biomarkers for diagnostics and therapies.

LitMiner and WikiGene: identifying problem-related key players of gene regulation using publication abstracts.

The LitMiner software is a literature data-mining tool that facilitates the identification of major gene regulation key players related to a user-defined field of interest in PubMed abstracts. The prediction of gene-regulatory relationships is based on co-occurrence analysis of key terms within the abstracts. LitMiner predicts relationships between key terms from the biomedical domain in four categories (genes, chemical compounds, diseases and tissues). Owing to the limitations (no direction, unverified automatic prediction) of the co-occurrence approach, the primary data in the LitMiner database represent postulated basic gene-gene relationships. The usefulness of the LitMiner system has been demonstrated recently in a study that reconstructed disease-related regulatory networks by promoter modelling that was initiated by a LitMiner generated primary gene list. To overcome the limitations and to verify and improve the data, we developed WikiGene, a Wiki-based curation tool that allows revision of the data by expert users over the Internet. LitMiner (http://andromeda.gsf.de/litminer) and WikiGene (http://andromeda.gsf.de/wiki) can be used unrestricted with any Internet browser.

Predicting stimulation-dependent enhancer-promoter interactions from ChIP-Seq time course data.

We have developed a machine learning approach to predict stimulation-dependent enhancer-promoter interactions using evidence from changes in genomic protein occupancy over time. The occupancy of estrogen receptor alpha (ERα), RNA polymerase (Pol II) and histone marks H2AZ and H3K4me3 were measured over time using ChIP-Seq experiments in MCF7 cells stimulated with estrogen. A Bayesian classifier was developed which uses the correlation of temporal binding patterns at enhancers and promoters and genomic proximity as features to predict interactions. This method was trained using experimentally determined interactions from the same system and was shown to achieve much higher precision than predictions based on the genomic proximity of nearest ERα binding. We use the method to identify a genome-wide confident set of ERα target genes and their regulatory enhancers genome-wide. Validation with publicly available GRO-Seq data demonstrates that our predicted targets are much more likely to show early nascent transcription than predictions based on genomic ERα binding proximity alone.

A novel genomic alteration of LSAMP associates with aggressive prostate cancer in African American men.

Evaluation of cancer genomes in global context is of great interest in light of changing ethnic distribution of the world population. We focused our study on men of African ancestry because of their disproportionately higher rate of prostate cancer (CaP) incidence and mortality. We present a systematic whole genome analyses, revealing alterations that differentiate African American (AA) and Caucasian American (CA) CaP genomes. We discovered a recurrent deletion on chromosome 3q13.31 centering on the LSAMP locus that was prevalent in tumors from AA men (cumulative analyses of 435 patients: whole genome sequence, 14; FISH evaluations, 101; and SNP array, 320 patients). Notably, carriers of this deletion experienced more rapid disease progression. In contrast, PTEN and ERG common driver alterations in CaP were significantly lower in AA prostate tumors compared to prostate tumors from CA. Moreover, the frequency of inter-chromosomal rearrangements was significantly higher in AA than CA tumors. These findings reveal differentially distributed somatic mutations in CaP across ancestral groups, which have implications for precision medicine strategies.

Basal core promoters control the equilibrium between negative cofactor 2 and preinitiation complexes in human cells.

BACKGROUND: The general transcription factor TFIIB and its antagonist negative cofactor 2 (NC2) are hallmarks of RNA polymerase II (RNAPII) transcription. Both factors bind TATA box-binding protein (TBP) at promoters in a mutually exclusive manner. Dissociation of NC2 is thought to be followed by TFIIB association and subsequent preinitiation complex formation. TFIIB dissociates upon RNAPII promoter clearance, thereby providing a specific measure for steady-state preinitiation complex levels. As yet, genome-scale promoter mapping of human TFIIB has not been reported. It thus remains elusive how human core promoters contribute to preinitiation complex formation in vivo. RESULTS: We compare target genes of TFIIB and NC2 in human B cells and analyze associated core promoter architectures. TFIIB occupancy is positively correlated with gene expression, with the vast majority of promoters being GC-rich and lacking defined core promoter elements. TATA elements, but not the previously in vitro defined TFIIB recognition elements, are enriched in some 4 to 5% of the genes. NC2 binds to a highly related target gene set. Nonetheless, subpopulations show strong variations in factor ratios: whereas high TFIIB/NC2 ratios select for promoters with focused start sites and conserved core elements, high NC2/TFIIB ratios correlate to multiple start-site promoters lacking defined core elements. CONCLUSIONS: TFIIB and NC2 are global players that occupy active genes. Preinitiation complex formation is independent of core elements at the majority of genes. TATA and TATA-like elements dictate TFIIB occupancy at a subset of genes. Biochemical data support a model in which preinitiation complex but not TBP-NC2 complex formation is regulated.

Global distribution of negative cofactor 2 subunit-alpha on human promoters.

Negative cofactor 2 (NC2) forms a stable complex with TATA-binding protein (TBP) on promoters in vitro. Its association with TBP prevents the binding of TFIIB and leads to inhibition of preinitiation complex formation. Here, we investigate the association of NC2 subunit-alpha with human RNA polymerase II promoter regions by using gene-specific ChIP and genome-wide promoter ChIPchip analyses. We find NC2alpha associated with a large number of human promoters, where it peaks close to the core regions. NC2 occupancy in vivo positively correlates with mRNA levels, which perhaps reflects its capacity to stabilize TBP on promoter regions. In single gene analyses, we confirm core promoter binding and in addition map the NC2 complex to enhancer proximal regions. High-occupancy histone genes display a stable NC2/TFIIB ratio during the cell cycle, which otherwise varies markedly from one gene to another. The latter is at least in part explained by an observed negative correlation of NC2 occupancy with the presence of the TFIIB recognition element in core promoter regions. Our data establish the genome-wide basis for general and gene-specific functions of NC2 in mammalian cells.

Genome-wide in silico mapping of scaffold/matrix attachment regions in Arabidopsis suggests correlation of intragenic scaffold/matrix attachment regions with gene expression.

We carried out a genome-wide prediction of scaffold/matrix attachment regions (S/MARs) in Arabidopsis. Results indicate no uneven distribution on the chromosomal level but a clear underrepresentation of S/MARs inside genes. In cases where S/MARs were predicted within genes, these intragenic S/MARs were preferentially located within the 5'-half, most prominently within introns 1 and 2. Using Arabidopsis whole-genome expression data generated by the massively parallel signature sequencing methodology, we found a negative correlation between S/MAR-containing genes and transcriptional abundance. Expressed sequence tag data correlated the same way with S/MAR-containing genes. Thus, intragenic S/MARs show a negative correlation with transcription level. For various genes it has been shown experimentally that S/MARs can function as transcriptional regulators and that they have an implication in stabilizing expression levels within transgenic plants. On the basis of a genome-wide in silico S/MAR analysis, we found a significant correlation between the presence of intragenic S/MARs and transcriptional down-regulation.