BACKGROUND: Elucidating the Transcription Factors (TFs) that drive the gene expression changes in a given experiment is a common question asked by researchers. The existing methods rely on the predicted Transcription Factor Binding Site (TFBS) to model the changes in the motif activity. Such methods only work for TFs that have a motif and assume the TF binding profile is the same in all cell types. RESULTS: Given the wealth of the ChIP-seq data available for a wide range of the TFs in various cell types, we propose that gene expression modeling can be done using ChIP-seq "signatures" directly, effectively skipping the motif finding and TFBS prediction steps. We present xcore, an R package that allows TF activity modeling based on ChIP-seq signatures and the user's gene expression data. We also provide xcoredata a companion data package that provides a collection of preprocessed ChIP-seq signatures. We demonstrate that xcore leads to biologically relevant predictions using transforming growth factor beta induced epithelial-mesenchymal transition time-courses, rinderpest infection time-courses, and embryonic stem cells differentiated to cardiomyocytes time-course profiled with Cap Analysis Gene Expression. CONCLUSIONS: xcore provides a simple analytical framework for gene expression modeling using linear models that can be easily incorporated into differential expression analysis pipelines. Taking advantage of public ChIP-seq databases, xcore can identify meaningful molecular signatures and relevant ChIP-seq experiments.

Drug treatment induces cell type specific transcriptional programs, and as the number of combinations of drugs and cell types grows, the cost for exhaustive screens measuring the transcriptional drug response becomes intractable. We developed DeepCellState, a deep learning autoencoder-based framework, for predicting the induced transcriptional state in a cell type after drug treatment, based on the drug response in another cell type. Training the method on a large collection of transcriptional drug perturbation profiles, prediction accuracy improves significantly over baseline and alternative deep learning approaches when applying the method to two cell types, with improved accuracy when generalizing the framework to additional cell types. Treatments with drugs or whole drug families not seen during training are predicted with similar accuracy, and the same framework can be used for predicting the results from other interventions, such as gene knock-downs. Finally, analysis of the trained model shows that the internal representation is able to learn regulatory relationships between genes in a fully data-driven manner.

Regulatory elements control gene expression through transcription initiation (promoters) and by enhancing transcription at distant regions (enhancers). Accurate identification of regulatory elements is fundamental for annotating genomes and understanding gene expression patterns. While there are many attempts to develop computational promoter and enhancer identification methods, reliable tools to analyze long genomic sequences are still lacking. Prediction methods often perform poorly on the genome-wide scale because the number of negatives is much higher than that in the training sets. To address this issue, we propose a dynamic negative set updating scheme with a two-model approach, using one model for scanning the genome and the other one for testing candidate positions. The developed method achieves good genome-level performance and maintains robust performance when applied to other vertebrate species, without re-training. Moreover, the unannotated predicted regulatory regions made on the human genome are enriched for disease-associated variants, suggesting them to be potentially true regulatory elements rather than false positives. We validated high scoring "false positive" predictions using reporter assay and all tested candidates were successfully validated, demonstrating the ability of our method to discover novel human regulatory regions.

MOTIVATION: Drug combination therapy for treatment of cancers and other multifactorial diseases has the potential of increasing the therapeutic effect, while reducing the likelihood of drug resistance. In order to reduce time and cost spent in comprehensive screens, methods are needed which can model additive effects of possible drug combinations. RESULTS: We here show that the transcriptional response to combinatorial drug treatment at promoters, as measured by single molecule CAGE technology, is accurately described by a linear combination of the responses of the individual drugs at a genome wide scale. We also find that the same linear relationship holds for transcription at enhancer elements. We conclude that the described approach is promising for eliciting the transcriptional response to multidrug treatment at promoters and enhancers in an unbiased genome wide way, which may minimize the need for exhaustive combinatorial screens. AVAILABILITY AND IMPLEMENTATION: The CAGE sequence data used in this study is available in the DDBJ Sequence Read Archive (http://trace.ddbj.nig.ac.jp/index_e.html), accession number DRP001113. CONTACT: [email protected] or [email protected]. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

The immediate-early response mediates cell fate in response to a variety of extracellular stimuli and is dysregulated in many cancers. However, the specificity of the response across stimuli and cell types, and the roles of non-coding RNAs are not well understood. Using a large collection of densely-sampled time series expression data we have examined the induction of the immediate-early response in unparalleled detail, across cell types and stimuli. We exploit cap analysis of gene expression (CAGE) time series datasets to directly measure promoter activities over time. Using a novel analysis method for time series data we identify transcripts with expression patterns that closely resemble the dynamics of known immediate-early genes (IEGs) and this enables a comprehensive comparative study of these genes and their chromatin state. Surprisingly, these data suggest that the earliest transcriptional responses often involve promoters generating non-coding RNAs, many of which are produced in advance of canonical protein-coding IEGs. IEGs are known to be capable of induction without de novo protein synthesis. Consistent with this, we find that the response of both protein-coding and non-coding RNA IEGs can be explained by their transcriptionally poised, permissive chromatin state prior to stimulation. We also explore the function of non-coding RNAs in the attenuation of the immediate early response in a small RNA sequencing dataset matched to the CAGE data: We identify a novel set of microRNAs responsible for the attenuation of the IEG response in an estrogen receptor positive cancer cell line. Our computational statistical method is well suited to meta-analyses as there is no requirement for transcripts to pass thresholds for significant differential expression between time points, and it is agnostic to the number of time points per dataset.

SUMMARY: High-throughput sequencing technologies introduce novel demands on tools available for data analysis. We have developed NGSView (Next Generation Sequence View), a generally applicable, flexible and extensible next-generation sequence alignment editor. The software allows for visualization and manipulation of millions of sequences simultaneously on a desktop computer, through a graphical interface. NGSView is available under an open source license and can be extended through a well documented API. AVAILABILITY: http://ngsview.sourceforge.net

BACKGROUND: Many genome projects are left unfinished due to complex, repeated regions. Finishing is the most time consuming step in sequencing and current finishing tools are not designed with particular attention to the repeat problem. RESULTS: We have developed DNPTrapper, a shotgun sequence finishing tool, specifically designed to address the problems posed by the presence of repeated regions in the target sequence. The program detects and visualizes single base differences between nearly identical repeat copies, and offers the overview and flexibility needed to rapidly resolve complex regions within a working session. The use of a database allows large amounts of data to be stored and handled, and allows viewing of mammalian size genomes. The program is available under an Open Source license. CONCLUSION: With DNPTrapper, it is possible to separate repeated regions that previously were considered impossible to resolve, and finishing tasks that previously took days or weeks can be resolved within hours or even minutes.

UNLABELLED: Finishing, i.e. gap closure and editing, is the most time-consuming part of genome sequencing. Repeated sequences together with sequencing errors complicate the assembly and often result in misassemblies that are difficult to correct. Repeat Discrepancy Tagger (ReDiT) is a tool designed to aid in the finishing step. This software processes assembly results produced by any fragment assembly program that outputs ace files. The input sequences are analyzed to determine possible differences between repeated sequences. The output is written as tags in an ace file that can be viewed by, e.g. the Consed sequence editor. AVAILABILITY: The ReDiT program is freely available at http://web.cgb.ki.se/redit

An increasingly important problem in genome sequencing is the failure of the commonly used shotgun assembly programs to correctly assemble repetitive sequences. The assembly of non-repetitive regions or regions containing repeats considerably shorter than the average read length is in practice easy to solve, while longer repeats have been a difficult problem. We here present a statistical method to separate arbitrarily long, almost identical repeats, which makes it possible to correctly assemble complex repetitive sequence regions. The differences between repeat units may be as low as 1% and the sequencing error may be up to ten times higher. The method is based on the realization that a comparison of only a part of all overlapping sequences at a time in a data set does not generate enough information for a conclusive analysis. Our method uses optimal multi-alignments consisting of all the overlaps of each read. This makes it possible to determine defined nucleotide positions, DNPs, which constitute the differences between the repeat units. Differences between repeats are distinguished from sequencing errors using statistical methods, where the probabilities of obtaining certain combinations of candidate DNPs are calculated using the information from the multi-alignments. The use of DNPs and combinations of DNPs will allow for optimal and rapid assemblies of repeated regions. This method can solve repeats that differ in only two positions in a read length, which is the theoretical limit for repeat separation. We predict that this method will be highly useful in shotgun sequencing in the future.