Histone modifications (HMs) are pivotal in various biological processes, including transcription, replication, and DNA repair, significantly impacting chromatin structure. These modifications underpin the molecular mechanisms of cell-type-specific gene expression and complex diseases. However, annotating HMs across different cell types solely using experimental approaches is impractical due to cost and time constraints. Herein, we present dHICA (deep histone imputation using chromatin accessibility), a novel deep learning framework that integrates DNA sequences and chromatin accessibility data to predict multiple HM tracks. Employing the transformer architecture alongside dilated convolutions, dHICA boasts an extensive receptive field and captures more cell-type-specific information. dHICA outperforms state-of-the-art baselines and achieves superior performance in cell-type-specific loci and gene elements, aligning with biological expectations. Furthermore, dHICA's imputations hold significant potential for downstream applications, including chromatin state segmentation and elucidating the functional implications of SNPs (Single Nucleotide Polymorphisms). In conclusion, dHICA serves as a valuable tool for advancing the understanding of chromatin dynamics, offering enhanced predictive capabilities and interpretability.

MOTIVATION: Quantification of isoform abundance has been extensively studied at the mature RNA level using RNA-seq but not at the level of precursor RNAs using nascent RNA sequencing. RESULTS: We address this problem with a new computational method called Deconvolution of Expression for Nascent RNA-sequencing data (DENR), which models nascent RNA-sequencing read-counts as a mixture of user-provided isoforms. The baseline algorithm is enhanced by machine-learning predictions of active transcription start sites and an adjustment for the typical 'shape profile' of read-counts along a transcription unit. We show that DENR outperforms simple read-count-based methods for estimating gene and isoform abundances, and that transcription of multiple pre-RNA isoforms per gene is widespread, with frequent differences between cell types. In addition, we provide evidence that a majority of human isoform diversity derives from primary transcription rather than from post-transcriptional processes. AVAILABILITY AND IMPLEMENTATION: DENR and nascentRNASim are freely available at https://github.com/CshlSiepelLab/DENR (version v1.0.0) and https://github.com/CshlSiepelLab/nascentRNASim (version v0.3.0). SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

UNLABELLED: Transcription factors (TFs) regulate complex programs of gene transcription by binding to short DNA sequence motifs. Here, we introduce rtfbsdb, a unified framework that integrates a database of more than 65 000 TF binding motifs with tools to easily and efficiently scan target genome sequences. Rtfbsdb clusters motifs with similar DNA sequence specificities and integrates RNA-seq or PRO-seq data to restrict analyses to motifs recognized by TFs expressed in the cell type of interest. Our package allows common analyses to be performed rapidly in an integrated environment. AVAILABILITY AND IMPLEMENTATION: rtfbsdb available at (https://github.com/Danko-Lab/rtfbs_db). CONTACT: [email protected] SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

BACKGROUND: Global run-on coupled with deep sequencing (GRO-seq) provides extensive information on the location and function of coding and non-coding transcripts, including primary microRNAs (miRNAs), long non-coding RNAs (lncRNAs), and enhancer RNAs (eRNAs), as well as yet undiscovered classes of transcripts. However, few computational tools tailored toward this new type of sequencing data are available, limiting the applicability of GRO-seq data for identifying novel transcription units. RESULTS: Here, we present groHMM, a computational tool in R, which defines the boundaries of transcription units de novo using a two state hidden-Markov model (HMM). A systematic comparison of the performance between groHMM and two existing peak-calling methods tuned to identify broad regions (SICER and HOMER) favorably supports our approach on existing GRO-seq data from MCF-7 breast cancer cells. To demonstrate the broader utility of our approach, we have used groHMM to annotate a diverse array of transcription units (i.e., primary transcripts) from four GRO-seq data sets derived from cells representing a variety of different human tissue types, including non-transformed cells (cardiomyocytes and lung fibroblasts) and transformed cells (LNCaP and MCF-7 cancer cells), as well as non-mammalian cells (from flies and worms). As an example of the utility of groHMM and its application to questions about the transcriptome, we show how groHMM can be used to analyze cell type-specific enhancers as defined by newly annotated enhancer transcripts. CONCLUSIONS: Our results show that groHMM can reveal new insights into cell type-specific transcription by identifying novel transcription units, and serve as a complete and useful tool for evaluating functional genomic elements in cells.

BACKGROUND: Cell specific gene expression is largely regulated by different combinations of transcription factors that bind cis-elements in the upstream promoter sequence. However, experimental detection of cis-elements is difficult, expensive, and time-consuming. This provides a motivation for developing bioinformatic methods to identify cis-elements that could prioritize future experimental studies. Here, we use motif discovery algorithms to predict transcription factor binding sites involved in regulating the differences between murine rod and cone photoreceptor populations. RESULTS: To identify highly conserved motifs enriched in promoters that drive expression in either rod or cone photoreceptors, we assembled a set of murine rod-specific, cone-specific, and non-photoreceptor background promoter sequences. These sets were used as input to a newly devised motif discovery algorithm called Iterative Alignment/Modular Motif Selection (IAMMS). Using IAMMS, we predicted 34 motifs that may contribute to rod-specific (19 motifs) or cone-specific (15 motifs) expression patterns. Of these, 16 rod- and 12 cone-specific motifs were found in clusters near the transcription start site. New findings include the observation that cone promoters tend to contain TATA boxes, while rod promoters tend to be TATA-less (exempting Rho and Cnga1). Additionally, we identify putative sites for IL-6 effectors (in rods) and RXR family members (in cones) that can explain experimental data showing changes to cell-fate by activating these signaling pathways during rod/cone development. Two of the predicted motifs (NRE and ROP2) have been confirmed experimentally to be involved in cell-specific expression patterns. We provide a full database of predictions as additional data that may contain further valuable information. IAMMS predictions are compared with existing motif discovery algorithms, DME and BioProspector. We find that over 60% of IAMMS predictions are confirmed by at least one other motif discovery algorithm. CONCLUSION: We predict novel, putative cis-elements enriched in the promoter of rod-specific or cone-specific genes. These are candidate binding sites for transcription factors involved in maintaining functional differences between rod and cone photoreceptor populations.