Annotation of genomics data using bidirectional hidden Markov models unveils variations in Pol II transcription cycle

A functional connection between the Microprocessor and a variant NEXT complex

In mammalian cells, primary miRNAs are cleaved at their hairpin structures by the Microprocessor complex, whose core is composed of DROSHA and DGCR8. Here, we show that 5' flanking regions, resulting from Microprocessor cleavage, are targeted by the RNA exosome in mouse embryonic stem cells (mESCs). This is facilitated by a physical link between DGCR8 and the nuclear exosome targeting (NEXT) component ZCCHC8. Surprisingly, however, both biochemical and mutagenesis studies demonstrate that a variant NEXT complex, containing the RNA helicase MTR4 but devoid of the RNA-binding protein RBM7, is the active entity. This Microprocessor-NEXT variant also targets stem-loop-containing RNAs expressed from other genomic regions, such as enhancers. By contrast, Microprocessor does not contribute to the turnover of less structured NEXT substrates. Our results therefore demonstrate that MTR4-ZCCHC8 can link to either RBM7 or DGCR8/DROSHA to target different RNA substrates depending on their structural context.

Nuclear RNA catabolism controls endogenous retroviruses, gene expression asymmetry, and dedifferentiation

Endogenous retroviruses (ERVs) are remnants of ancient parasitic infections and comprise sizable portions of most genomes. Although epigenetic mechanisms silence most ERVs by generating a repressive environment that prevents their expression (heterochromatin), little is known about mechanisms silencing ERVs residing in open regions of the genome (euchromatin). This is particularly important during embryonic development, where induction and repression of distinct classes of ERVs occur in short temporal windows. Here, we demonstrate that transcription-associated RNA degradation by the nuclear RNA exosome and Integrator is a regulatory mechanism that controls the productive transcription of most genes and many ERVs involved in preimplantation development. Disrupting nuclear RNA catabolism promotes dedifferentiation to a totipotent-like state characterized by defects in RNAPII elongation and decreased expression of long genes (gene-length asymmetry). Our results indicate that RNA catabolism is a core regulatory module of gene networks that safeguards RNAPII activity, ERV expression, cell identity, and developmental potency.

Automatic block-wise genotype-phenotype association detection based on hidden Markov model

BACKGROUND

For detecting genotype-phenotype association from case-control single nucleotide polymorphism (SNP) data, one class of methods relies on testing each genomic variant site individually. However, this approach ignores the tendency for associated variant sites to be spatially clustered instead of uniformly distributed along the genome. Therefore, a more recent class of methods looks for blocks of influential variant sites. Unfortunately, existing such methods either assume prior knowledge of the blocks, or rely on ad hoc moving windows. A principled method is needed to automatically detect genomic variant blocks which are associated with the phenotype.

RESULTS

In this paper, we introduce an automatic block-wise Genome-Wide Association Study (GWAS) method based on Hidden Markov model. Using case-control SNP data as input, our method detects the number of blocks associated with the phenotype and the locations of the blocks. Correspondingly, the minor allele of each variate site will be classified as having negative influence, no influence or positive influence on the phenotype. We evaluated our method using both datasets simulated from our model and datasets from a block model different from ours, and compared the performance with other methods. These included both simple methods based on the Fisher's exact test, applied site-by-site, as well as more complex methods built into the recent Zoom-Focus Algorithm. Across all simulations, our method consistently outperformed the comparisons.

CONCLUSIONS

With its demonstrated better performance, we expect our algorithm for detecting influential variant sites may help find more accurate signals across a wide range of case-control GWAS.

Integrator is a genome-wide attenuator of non-productive transcription

Termination of RNA polymerase II (RNAPII) transcription in metazoans relies largely on the cleavage and polyadenylation (CPA) and integrator (INT) complexes originally found to act at the ends of protein-coding and small nuclear RNA (snRNA) genes, respectively. Here, we monitor CPA- and INT-dependent termination activities genome-wide, including at thousands of previously unannotated transcription units (TUs), producing unstable RNA. We verify the global activity of CPA occurring at pA sites indiscriminately of their positioning relative to the TU promoter. We also identify a global activity of INT, which is largely sequence-independent and restricted to a ~3-kb promoter-proximal region. Our analyses suggest two functions of genome-wide INT activity: it dampens transcriptional output from weak promoters, and it provides quality control of RNAPII complexes that are unfavorably configured for transcriptional elongation. We suggest that the function of INT in stable snRNA production is an exception from its general cellular role, the attenuation of non-productive transcription.

Revealing metabolic mechanisms of interaction in the anaerobic digestion microbiome by flux balance analysis

Anaerobic digestion is a key biological process for renewable energy, yet the mechanistic knowledge on its hidden microbial dynamics is still limited. The present work charted the interaction network in the anaerobic digestion microbiome via the full characterization of pairwise interactions and the associated metabolite exchanges. To this goal, a novel collection of 836 genome-scale metabolic models was built to represent the functional capabilities of bacteria and archaea species derived from genome-centric metagenomics. Dominant microbes were shown to prefer mutualistic, parasitic and commensalistic interactions over neutralism, amensalism and competition, and are more likely to behave as metabolite importers and profiteers of the coexistence. Additionally, external hydrogen injection positively influences microbiome dynamics by promoting commensalism over amensalism. Finally, exchanges of glucogenic amino acids were shown to overcome auxotrophies caused by an incomplete tricarboxylic acid cycle. Our novel strategy predicted the most favourable growth conditions for the microbes, overall suggesting strategies to increasing the biogas production efficiency. In principle, this approach could also be applied to microbial populations of biomedical importance, such as the gut microbiome, to allow a broad inspection of the microbial interplays.

EpiAlign: an alignment-based bioinformatic tool for comparing chromatin state sequences

The availability of genome-wide epigenomic datasets enables in-depth studies of epigenetic modifications and their relationships with chromatin structures and gene expression. Various alignment tools have been developed to align nucleotide or protein sequences in order to identify structurally similar regions. However, there are currently no alignment methods specifically designed for comparing multi-track epigenomic signals and detecting common patterns that may explain functional or evolutionary similarities. We propose a new local alignment algorithm, EpiAlign, designed to compare chromatin state sequences learned from multi-track epigenomic signals and to identify locally aligned chromatin regions. EpiAlign is a dynamic programming algorithm that novelly incorporates varying lengths and frequencies of chromatin states. We demonstrate the efficacy of EpiAlign through extensive simulations and studies on the real data from the NIH Roadmap Epigenomics project. EpiAlign is able to extract recurrent chromatin state patterns along a single epigenome, and many of these patterns carry cell-type-specific characteristics. EpiAlign can also detect common chromatin state patterns across multiple epigenomes, and it will serve as a useful tool to group and distinguish epigenomic samples based on genome-wide or local chromatin state patterns.

Reveal cell type-specific regulatory elements and their characterized histone code classes via a hidden Markov model

BACKGROUND

With the maturity of next generation sequencing technology, a huge amount of epigenomic data have been generated by several large consortia in the last decade. These plenty resources leave us the opportunity about sufficiently utilizing those data to explore biological problems.

RESULTS

Here we developed an integrative and comparative method, CsreHMM, which is based on a hidden Markov model, to systematically reveal cell type-specific regulatory elements (CSREs) along the whole genome, and simultaneously recognize the histone codes (mark combinations) charactering them. This method also reveals the subclasses of CSREs and explicitly label those shared by a few cell types. We applied this method to a data set of 9 cell types and 9 chromatin marks to demonstrate its effectiveness and found that the revealed CSREs relates to different kinds of functional regulatory regions significantly. Their proximal genes have consistent expression and are likely to participate in cell type-specific biological functions.

CONCLUSIONS

These results suggest CsreHMM has the potential to help understand cell identity and the diverse mechanisms of gene regulation.

Chromatin-state discovery and genome annotation with ChromHMM

Noncoding DNA regions have central roles in human biology, evolution, and disease. ChromHMM helps to annotate the noncoding genome using epigenomic information across one or multiple cell types. It combines multiple genome-wide epigenomic maps, and uses combinatorial and spatial mark patterns to infer a complete annotation for each cell type. ChromHMM learns chromatin-state signatures using a multivariate hidden Markov model (HMM) that explicitly models the combinatorial presence or absence of each mark. ChromHMM uses these signatures to generate a genome-wide annotation for each cell type by calculating the most probable state for each genomic segment. ChromHMM provides an automated enrichment analysis of the resulting annotations to facilitate the functional interpretations of each chromatin state. ChromHMM is distinguished by its modeling emphasis on combinations of marks, its tight integration with downstream functional enrichment analyses, its speed, and its ease of use. Chromatin states are learned, annotations are produced, and enrichments are computed within 1 d.

Similarity-Based Segmentation of Multi-Dimensional Signals

The segmentation of time series and genomic data is a common problem in computational biology. With increasingly complex measurement procedures individual data points are often not just numbers or simple vectors in which all components are of the same kind. Analysis methods that capitalize on slopes in a single real-valued data track or that make explicit use of the vectorial nature of the data are not applicable in such scenaria. We develop here a framework for segmentation in arbitrary data domains that only requires a minimal notion of similarity. Using unsupervised clustering of (a sample of) the input yields an approximate segmentation algorithm that is efficient enough for genome-wide applications. As a showcase application we segment a time-series of transcriptome sequencing data from budding yeast, in high temporal resolution over ca. 2.5 cycles of the short-period respiratory oscillation. The algorithm is used with a similarity measure focussing on periodic expression profiles across the metabolic cycle rather than coverage per time point.

Retrieving Chromatin Patterns from Deep Sequencing Data Using Correlation Functions

Epigenetic modifications and other chromatin features partition the genome on multiple length scales. They define chromatin domains with distinct biological functions that come in sizes ranging from single modified DNA bases to several megabases in the case of heterochromatic histone modifications. Due to chromatin folding, domains that are well separated along the linear nucleosome chain can form long-range interactions in three-dimensional space. It has now become a routine task to map epigenetic marks and chromatin structure by deep sequencing methods. However, assessing and comparing the properties of chromatin domains and their positional relationships across data sets without a priori assumptions remains challenging. Here, we introduce multiscale correlation evaluation (MCORE), which uses the fluctuation spectrum of mapped sequencing reads to quantify and compare chromatin patterns over a broad range of length scales in a model-independent manner. We applied MCORE to map the chromatin landscape in mouse embryonic stem cells and differentiated neural cells. We integrated sequencing data from chromatin immunoprecipitation, RNA expression, DNA methylation, and chromosome conformation capture experiments into network models that reflect the positional relationships among these features on different genomic scales. Furthermore, we used MCORE to compare our experimental data to models for heterochromatin reorganization during differentiation. The application of correlation functions to deep sequencing data complements current evaluation schemes and will support the development of quantitative descriptions of chromatin networks.

Accurate Promoter and Enhancer Identification in 127 ENCODE and Roadmap Epigenomics Cell Types and Tissues by GenoSTAN

Accurate maps of promoters and enhancers are required for understanding transcriptional regulation. Promoters and enhancers are usually mapped by integration of chromatin assays charting histone modifications, DNA accessibility, and transcription factor binding. However, current algorithms are limited by unrealistic data distribution assumptions. Here we propose GenoSTAN (Genomic STate ANnotation), a hidden Markov model overcoming these limitations. We map promoters and enhancers for 127 cell types and tissues from the ENCODE and Roadmap Epigenomics projects, today’s largest compendium of chromatin assays. Extensive benchmarks demonstrate that GenoSTAN generally identifies promoters and enhancers with significantly higher accuracy than previous methods. Moreover, GenoSTAN-derived promoters and enhancers showed significantly higher enrichment of complex trait-associated genetic variants than current annotations. Altogether, GenoSTAN provides an easy-to-use tool to define promoters and enhancers in any system, and our annotation of human transcriptional cis-regulatory elements constitutes a rich resource for future research in biology and medicine.

Strand-specific, high-resolution mapping of modified RNA polymerase II

Reversible modification of the RNAPII C‐terminal domain links transcription with RNA processing and surveillance activities. To better understand this, we mapped the location of RNAPII carrying the five types of CTD phosphorylation on the RNA transcript, providing strand‐specific, nucleotide‐resolution information, and we used a machine learning‐based approach to define RNAPII states. This revealed enrichment of Ser5P, and depletion of Tyr1P, Ser2P, Thr4P, and Ser7P in the transcription start site (TSS) proximal ~150 nt of most genes, with depletion of all modifications close to the poly(A) site. The TSS region also showed elevated RNAPII relative to regions further 3′, with high recruitment of RNA surveillance and termination factors, and correlated with the previously mapped 3′ ends of short, unstable ncRNA transcripts. A hidden Markov model identified distinct modification states associated with initiating, early elongating and later elongating RNAPII. The initiation state was enriched near the TSS of protein‐coding genes and persisted throughout exon 1 of intron‐containing genes. Notably, unstable ncRNAs apparently failed to transition into the elongation states seen on protein‐coding genes.

Simultaneous characterization of sense and antisense genomic processes by the double-stranded hidden Markov model

Hidden Markov models (HMMs) have been extensively used to dissect the genome into functionally distinct regions using data such as RNA expression or DNA binding measurements. It is a challenge to disentangle processes occurring on complementary strands of the same genomic region. We present the double-stranded HMM (dsHMM), a model for the strand-specific analysis of genomic processes. We applied dsHMM to yeast using strand specific transcription data, nucleosome data, and protein binding data for a set of 11 factors associated with the regulation of transcription.The resulting annotation recovers the mRNA transcription cycle (initiation, elongation, termination) while correctly predicting strand-specificity and directionality of the transcription process. We find that pre-initiation complex formation is an essentially undirected process, giving rise to a large number of bidirectional promoters and to pervasive antisense transcription. Notably, 12% of all transcriptionally active positions showed simultaneous activity on both strands. Furthermore, dsHMM reveals that antisense transcription is specifically suppressed by Nrd1, a yeast termination factor.