Studying alternative splicing regulatory networks through partial correlation analysis

Exon inclusion signatures enable accurate estimation of splicing factor activity

Splicing factors control exon inclusion in messenger RNA, shaping transcriptome and proteome diversity. Their catalytic activity is regulated by multiple layers, making single-omic measurements on their own fall short in identifying which splicing factors underlie a phenotype. Here, we propose splicing factor activity can be estimated by interpreting changes in exon inclusion. We benchmark methods to construct splicing factor→exon networks and calculate activity. Combining RNA-seq perturbation-based networks with VIPER (virtual inference of protein activity by enriched regulon analysis) accurately captures splicing factor activation modulated by different regulatory layers. This approach consolidates splicing factor regulation into a single score derived solely from exon inclusion signatures, allowing functional interpretation of heterogeneous conditions. As a proof of concept, we identify recurrent cancer splicing programs, revealing oncogenic- and tumor suppressor-like splicing factors missed by conventional methods. These programs correlate with patient survival and key cancer hallmarks: initiation, proliferation, and immune evasion. Altogether, we show splicing factor activity can be accurately estimated from exon inclusion changes, enabling comprehensive analyses of splicing regulation with minimal data requirements.

GLIMS: A two-stage gradual-learning method for cancer genes prediction using multi-omics data and co-splicing network

Identifying cancer genes is vital for cancer diagnosis and treatment. However, because of the complexity of cancer occurrence and limited cancer genes knowledge, it is hard to identify cancer genes accurately using only a few omics data, and the overall performance of existing methods is being called for further improvement. Here, we introduce a two-stage gradual-learning strategy GLIMS to predict cancer genes using integrative features from multi-omics data. Firstly, it uses a semi-supervised hierarchical graph neural network to predict the initial candidate cancer genes by integrating multi-omics data and protein-protein interaction (PPI) network. Then, it uses an unsupervised approach to further optimize the initial prediction by integrating the co-splicing network in post-transcriptional regulation, which plays an important role in cancer development. Systematic experiments on multi-omics cancer data demonstrated that GLIMS outperforms the state-of-the-art methods for the identification of cancer genes and it could be a useful tool to help advance cancer analysis.

DUX4 is a common driver of immune evasion and immunotherapy failure in metastatic cancers

Cancer immune evasion contributes to checkpoint immunotherapy failure in many patients with metastatic cancers. The embryonic transcription factor DUX4 was recently characterized as a suppressor of interferon-γ signaling and antigen presentation that is aberrantly expressed in a small subset of primary tumors. Here, we report that DUX4 expression is a common feature of metastatic tumors, with ~10-50% of advanced bladder, breast, kidney, prostate, and skin cancers expressing DUX4. DUX4 expression is significantly associated with immune cell exclusion and decreased objective response to PD-L1 blockade in a large cohort of urothelial carcinoma patients. DUX4 expression is a significant predictor of survival even after accounting for tumor mutational burden and other molecular and clinical features in this cohort, with DUX4 expression associated with a median reduction in survival of over one year. Our data motivate future attempts to develop DUX4 as a biomarker and therapeutic target for checkpoint immunotherapy resistance.

Alternative splicing programming of axon formation

Alternative pre-mRNA splicing generates multiple mRNA isoforms of different structures and functions from a single gene. While the prevalence of alternative splicing control is widely recognized, and the underlying regulatory mechanisms have long been studied, the physiological relevance and biological necessity for alternative splicing are only slowly being revealed. Significant inroads have been made in the brain, where alternative splicing regulation is particularly pervasive and conserved. Various aspects of brain development and function (from neurogenesis, neuronal migration, synaptogenesis, to the homeostasis of neuronal activity) involve alternative splicing regulation. Recent studies have begun to interrogate the possible role of alternative splicing in axon formation, a neuron-exclusive morphological and functional characteristic. We discuss how alternative splicing plays an instructive role in each step of axon formation. Converging genetic, molecular, and cellular evidence from studies of multiple alternative splicing regulators in different systems shows that a biological process as complicated and unique as axon formation requires highly coordinated and specific alternative splicing events. This article is categorized under: RNA Processing > Splicing Regulation/Alternative Splicing RNA in Disease and Development > RNA in Development.

Neuronal impact of patient-specific aberrant NRXN1α splicing

NRXN1 undergoes extensive alternative splicing, and non-recurrent heterozygous deletions in NRXN1 are strongly associated with neuropsychiatric disorders. We establish that human induced pluripotent stem cell (hiPSC)-derived neurons represent well the diversity of NRXN1α alternative splicing observed in the human brain, cataloguing 123 high-confidence in-frame human NRXN1α isoforms. Patient-derived NRXN1^+/− hiPSC-neurons show greater than two-fold reduction of half of the wild-type NRXN1α isoforms and express dozens of novel isoforms expressed from the mutant allele. Reduced neuronal activity in patient-derived NRXN1^+/− hiPSC-neurons is ameliorated by overexpression of individual control isoforms in a genotype-dependent manner, whereas individual mutant isoforms decrease neuronal activity levels in control hiPSC-neurons. In a genotype-dependent manner, the phenotypic impact of patient-specific NRXN1^+/− mutations can occur through a reduction in wild-type NRXN1α isoform levels as well as the presence of mutant NRXN1α isoforms.

Developmental Xist induction is mediated by enhanced splicing

X-inactive-specific transcript (Xist) is a long noncoding RNA (lncRNA) essential for inactivating one of the two X chromosomes in mammalian females. Random X chromosome inactivation is mediated by Xist RNA expressed from the inactive X chromosome. We found that Xist RNA is unspliced in naïve embryonic stem (ES) cells. Upon differentiation, Xist splicing becomes efficient across all exons independent of transcription, suggesting interdependent or coordinated removal of Xist introns. In female cells with mutated polypyrimidine tract binding protein 1 (Ptbp1), differentiation fails to substantially upregulate mature Xist RNA because of a defect in Xist splicing. We further found both Xist129 and XistCAS RNA are unspliced in Mus musculus 129SvJ/Mus castaneous (CAS) hybrid female ES cells. Upon differentiation, Xist129 exhibits a higher splicing efficiency than XistCAS, likely contributing to preferential inhibition of the X129 chromosome. Single cell analysis shows that the allelic choice of Xist splicing is linked to the inactive X chromosome. We conclude post-transcriptional control of Xist RNA splicing is an essential regulatory step of Xist induction. Our studies shed light on the developmental roles of splicing for nuclear-retained Xist lncRNA and suggest inefficient Xist splicing is an additional fail-safe mechanism to prevent Xist activity in ES cells.

Axonogenesis Is Coordinated by Neuron-Specific Alternative Splicing Programming and Splicing Regulator PTBP2

How a neuron acquires an axon is a fundamental question. Piecemeal identification of many axonogenesis-related genes has been done, but coordinated regulation is unknown. Through unbiased transcriptome profiling of immature primary cortical neurons during early axon formation, we discovered an association between axonogenesis and neuron-specific alternative splicing. Known axonogenesis genes exhibit little expression alternation but widespread splicing changes. Axonogenesis-associated splicing is governed by RNA binding protein PTBP2, which is enriched in neurons and peaks around axonogenesis in the brain. Cortical depletion of PTBP2 prematurely induces axonogenesis-associated splicing, causes imbalanced expression of axonogenesis-associated isoforms, and specifically affects axon formation in vitro and in vivo. PTBP2-controlled axonogenesis-associated Shtn1 splicing determines SHTN1's capacity to regulate actin interaction, polymerization, and axon growth. Precocious Shtn1 isoform switch contributes to disorganized axon formation of Ptbp2^-/- neurons. We conclude that PTBP2-orchestrated alternative splicing programming is required for robust generation of a single axon in mammals.

Alternative processing of its precursor is related to miR319 decreasing in melon plants exposed to cold

miRNAs are fundamental endogenous regulators of gene expression in higher organisms. miRNAs modulate multiple biological processes in plants. Consequently, miRNA accumulation is strictly controlled through miRNA precursor accumulation and processing. Members of the miRNA319 family are ancient ribo-regulators that are essential for plant development and stress responses and exhibit an unusual biogenesis that is characterized by multiple processing of their precursors. The significance of the high conservation of these non-canonical biogenesis pathways remains unknown. Here, we analyze data obtained by massive sRNA sequencing and 5′ - RACE to explore the accumulation and infer the processing of members of the miR319 family in melon plants exposed to adverse environmental conditions. Sequence data showed that miR319c was down regulated in response to low temperature. However, the level of its precursor was increased by cold, indicating that miR319c accumulation is not related to the stem loop levels. Furthermore, we found that a decrease in miR319c was inversely correlated with the stable accumulation of an alternative miRNA (#miR319c) derived from multiple processing of the miR319c precursor. Interestingly, the alternative accumulation of miR319c and #miR319c was associated with an additional and non-canonical partial cleavage of the miR319c precursor during its loop-to-base-processing. Analysis of the transcriptional activity showed that miR319c negatively regulated the accumulation of HY5 via TCP2 in melon plants exposed to cold, supporting its involvement in the low temperature signaling pathway associated with anthocyanin biosynthesis. Our results provide new insights regarding the versatility of plant miRNA processing and the mechanisms regulating them as well as the hypothetical mechanism for the response to cold-induced stress in melon, which is based on the alternative regulation of miRNA biogenesis.

Covariate-adjusted heatmaps for visualizing biological data via correlation decomposition

Motivation

Heatmap is a popular visualization technique in biology and related fields. In this study, we extend heatmaps within the framework of matrix visualization (MV) by incorporating a covariate adjustment process through the estimation of conditional correlations. MV can explore the embedded information structure of high-dimensional large-scale datasets effectively without dimension reduction. The benefit of the proposed covariate-adjusted heatmap is in the exploration of conditional association structures among the subjects or variables that cannot be done with conventional MV.

Results

For adjustment of a discrete covariate, the conditional correlation is estimated by the within and between analysis. This procedure decomposes a correlation matrix into the within- and between-component matrices. The contribution of the covariate effects can then be assessed through the relative structure of the between-component to the original correlation matrix while the within-component acts as a residual. When a covariate is of continuous nature, the conditional correlation is equivalent to the partial correlation under the assumption of a joint normal distribution. A test is then employed to identify the variable pairs which possess the most significant differences at varying levels of correlation before and after a covariate adjustment. In addition, a z-score significance map is constructed to visualize these results. A simulation and three biological datasets are employed to illustrate the power and versatility of our proposed method.

Availability and implementation

GAP is available to readers and is free to non-commercial applications. The installation instructions, the user's manual, and the detailed tutorials can be found at http://gap.stat.sinica.edu.tw/Software/GAP.

Supplementary information

Supplementary Data are available at Bioinformatics online.

l-Arabinose triggers its own uptake via induction of the arabinose-specific Gal2p transporter in an industrial Saccharomyces cerevisiae strain

Bioethanol production processes with Saccharomyces cerevisiae using lignocellulosic biomass as feedstock are challenged by the simultaneous utilization of pentose and hexose sugars from biomass hydrolysates. The pentose uptake into the cell represents a crucial role for the efficiency of the process. The focus of the here presented study was to understand the uptake and conversion of the pentose l-arabinose in S. cerevisiae and reveal its regulation by d-glucose and d-galactose. Gal2p-the most prominent transporter enabling l-arabinose uptake in S. cerevisiae wild-type strains-has an affinity for the transport of l-arabinose, d-glucose, and d-galactose. d-Galactose was reported for being mandatory for inducing GAL2 expression. GAL2 expression is also known to be regulated by d-glucose-mediated carbon catabolite repression, as well as catabolite inactivation. The results of the present study demonstrate that l-arabinose can be used as sole carbon and energy source by the recombinant industrial strain S. cerevisiae DS61180. RT-qPCR and RNA-Seq experiments confirmed that l-arabinose can trigger its own uptake via the induction of GAL2 expression. Expression levels of GAL2 during growth on l-arabinose reached up to 21% of those obtained with d-galactose as sole carbon and energy source. l-Arabinose-induced GAL2 expression was also subject to catabolite repression by d-glucose. Kinetic investigations of substrate uptake, biomass, and product formation during growth on a mixture of d-glucose/l-arabinose revealed impairment of growth and ethanol production from l-arabinose upon d-glucose depletion. The presence of d-glucose is thus preventing the fermentation of l-arabinose in S. cerevisiae DS61180. Comparative transcriptome studies including the wild-type and a precursor strain delivered hints for an increased demand in ATP production and cofactor regeneration during growth of S. cerevisiae DS61180 on l-arabinose. Our results thus emphasize that cofactor and energy metabolism demand attention if the combined conversion of hexose and pentose sugars is intended, for example in biorefineries using lignocellulosics.

A Cell-Based High-Throughput Method for Identifying Modulators of Alternative Splicing

Alternative splicing, a key regulatory process of gene expression, is controlled by trans-acting factors that recognize cis-elements in premature RNA transcripts to affect spliceosome assembly and splice site choices. Extracellular stimuli and signaling cascades can converge on RNA binding splicing regulators to affect alternative splicing. Defects in splicing regulation have been associated with various human diseases, and modification of disease-causing splicing events presents great therapeutic promise. Determining splicing regulators and/or upstream modulators has been difficult with low throughput, low sensitivity, and low specificity. IRAS (identifying regulators of alternative splicing) is a novel cell-based high-throughput screening strategy designed specifically to address these challenges and has achieved high throughput, high sensitivity, and high specificity. Here, we describe the IRAS method in detail with a pair of dual-fluorescence minigene reporters that produces GFP and RFP fluorescent signals to assay the two spliced isoforms exclusively. These two complementary mini-gene reporters alter GFP/RFP output ratios in the opposite direction in response to only a true splicing change. False positives from a signal screen do not stimulate opposite changes in GFP/RFP ratios. The reporter pair in conjunction with robotic liquid handlers and arrayed libraries allows IRAS to screen for both positive and negative splicing regulators with high sensitivity and specificity in a high-throughput manner.

High-Throughput Screening Method to Identify Alternative Splicing Regulators

Misregulation of alternative pre-mRNA splicing contributes to various diseases. Understanding how alternative splicing is regulated paves the way to modulating or correcting molecular pathogenesis of the diseases. Alternative splicing is typically regulated by trans RNA binding proteins and their upstream modulators. Identification of these splicing regulators has been difficult and traditionally done piecemeal. High-throughput screening strategies to find multiple regulators of exon splicing have great potential to accelerate the discovery process, but typically confront low sensitivity and specificity of splicing assays. Here we describe a high-throughput screening method using dual-fluorescence minigene reporters to allow for sensitive detection of exon splicing changes. To enhance specificity we introduce two complementary dual-fluorescence minigenes that each express both GFP and RFP in response to exon inclusion and exclusion but oppositely. The method significantly eliminates false positives and allows for sensitive identification of true regulators of splicing. The method described here is designed to screen cDNA libraries, but can be applied to isolate splicing regulators from shRNA libraries or chemical libraries.

Beyond Transcription: Roles of Transcription Factors in Pre-mRNA Splicing

Whereas individual steps of protein-coding gene expression in eukaryotes can be studied in isolation in vitro, it has become clear that these steps are intimately connected within cells. Connections not only ensure quality control but also fine-tune the gene expression process, which must adapt to environmental changes while remaining robust. In this review, we systematically present proven and potential mechanisms by which sequence-specific DNA-binding transcription factors can alter gene expression beyond transcription initiation and regulate pre-mRNA splicing, and thereby mRNA isoform production, by (i) influencing transcription elongation rates, (ii) binding to pre-mRNA to recruit splicing factors, and/or (iii) blocking the association of splicing factors with pre-mRNA. We propose various mechanistic models throughout the review, in some cases without explicit supportive evidence, in hopes of providing fertile ground for future studies.

Decoding critical long non-coding RNA in ovarian cancer epithelial-to-mesenchymal transition

Long non-coding RNA (lncRNA) are emerging as contributors to malignancies. Little is understood about the contribution of lncRNA to epithelial-to-mesenchymal transition (EMT), which correlates with metastasis. Ovarian cancer is usually diagnosed after metastasis. Here we report an integrated analysis of >700 ovarian cancer molecular profiles, including genomic data sets, from four patient cohorts identifying lncRNA DNM3OS, MEG3, and MIAT overexpression and their reproducible gene regulation in ovarian cancer EMT. Genome-wide mapping shows 73% of MEG3-regulated EMT-linked pathway genes contain MEG3 binding sites. DNM3OS overexpression, but not MEG3 or MIAT, significantly correlates to worse overall patient survival. DNM3OS knockdown results in altered EMT-linked genes/pathways, mesenchymal-to-epithelial transition, and reduced cell migration and invasion. Proteotranscriptomic characterization further supports the DNM3OS and ovarian cancer EMT connection. TWIST1 overexpression and DNM3OS amplification provides an explanation for increased DNM3OS levels. Therefore, our results elucidate lncRNA that regulate EMT and demonstrate DNM3OS specifically contributes to EMT in ovarian cancer.

Transcriptomic study of Salmonella enterica subspecies enterica serovar Typhi biofilm

Background

Typhoid fever is an acute systemic infection of humans caused by Salmonella enterica subspecies enterica serovar Typhi (S. Typhi). In chronic carriers, the bacteria survive the harsh environment of the gallbladder by producing biofilm. The phenotype of S. Typhi biofilm cells is significantly different from the free-swimming planktonic cells, and studies have shown that they are associated with antibiotic resistance, immune system evasion, and bacterial persistence. However, the mechanism of this transition and the events leading to biofilm formation are unknown. High throughput sequencing was performed to identify the genes involved in biofilm formation and to postulate the mechanism of action.

Results

Planktonic S. Typhi cells were cultured using standard nutrient broth whereas biofilm cells were cultured in a stressful environment using high shearing-force and bile to mimic the gallbladder. Sequencing libraries were prepared from S. Typhi planktonic cells and mature biofilm cells using the Illumina HiSeq 2500 platform, and the transcriptome data obtained were processed using Cufflinks bioinformatics suite of programs to investigate differential gene expression between the two phenotypes. A total of 35 up-regulated and 29 down-regulated genes were identified. The identities of the differentially expressed genes were confirmed using NCBI BLAST and their functions were analyzed. The results showed that the genes associated with metabolic processes and biofilm regulations were down-regulated while those associated with the membrane matrix and antibiotic resistance were highly up-regulated.

Conclusions

It is proposed that the biofilm phenotype of S. Typhi allows the bacteria to increase production of the membrane matrix in order to serve as a physical shield and to adhere to surfaces, and enter an energy conservation state in response to the stressful environment. Conversely, the planktonic phenotype allows the bacteria to produce flagella and increase metabolic activity to enable the bacteria to migrate and form new colonies of infection. This data provide a basis for further studies to uncover the mechanism of biofilm formation in S. Typhi and to discover novel genes or pathways associated with the development of the typhoid carrier state.

Electronic supplementary material

The online version of this article (10.1186/s12864-017-4212-6) contains supplementary material, which is available to authorized users.

IRAS: High-Throughput Identification of Novel Alternative Splicing Regulators

Alternative splicing is a fundamental regulatory process of gene expression. Defects in alternative splicing can lead to various diseases, and modification of disease-causing splicing events presents great therapeutic promise. Splicing outcome is commonly affected by extracellular stimuli and signaling cascades that converge on RNA-binding splicing regulators. These trans-acting factors recognize cis-elements in pre-mRNA transcripts to affect spliceosome assembly and splice site choices. Identification of these splicing regulators and/or upstream modulators has been difficult and traditionally done by piecemeal. High-throughput screening strategies to find multiple regulators of exon splicing have great potential to accelerate the discovery process, but typically confront low sensitivity and low specificity of screening assays. Here we describe a unique screening strategy, IRAS (identifying regulators of alternative splicing), using a pair of dual-output minigene reporters to allow for sensitive detection of exon splicing changes. Each dual-output reporter produces green fluorescent protein (GFP) and red fluorescent protein (RFP) fluorescent signals to assay the two spliced isoforms exclusively. The two complementary minigene reporters alter GFP/RFP output ratios in the opposite direction in response to splicing change. Applying IRAS in cell-based high-throughput screens allows sensitive and specific identification of splicing regulators and modulators for any alternative exons of interest. In comparison to previous high-throughput screening methods, IRAS substantially enhances the specificity of the screening assay. This strategy significantly eliminates false positives without sacrificing sensitive identification of true regulators of splicing.

Discovery of new methylation markers to improve screening for cervical intraepithelial neoplasia grade 2/3

Background

Assessment of DNA promoter methylation markers in cervical scrapings for the detection of cervical intraepithelial neoplasia (CIN) and cervical cancer is feasible, but finding methylation markers with both high sensitivity as well as high specificity remains a challenge. In this study, we aimed to identify new methylation markers for the detection of high-grade CIN (CIN2/3 or worse, CIN2+) by using innovative genome-wide methylation analysis (MethylCap-seq). We focused on diagnostic performance of methylation markers with high sensitivity and high specificity considering any methylation level as positive.

Results

MethylCap-seq of normal cervices and CIN2/3 revealed 176 differentially methylated regions (DMRs) comprising 164 genes. After verification and validation of the 15 best discriminating genes with methylation-specific PCR (MSP), 9 genes showed significant differential methylation in an independent cohort of normal cervices versus CIN2/3 lesions (p < 0.05). For further diagnostic evaluation, these 9 markers were tested with quantitative MSP (QMSP) in cervical scrapings from 2 cohorts: (1) cervical carcinoma versus healthy controls and (2) patients referred from population-based screening with an abnormal Pap smear in whom also HPV status was determined. Methylation levels of 8/9 genes were significantly higher in carcinoma compared to normal scrapings. For all 8 genes, methylation levels increased with the severity of the underlying histological lesion in scrapings from patients referred with an abnormal Pap smear. In addition, the diagnostic performance was investigated, using these 8 new genes and 4 genes (previously identified by our group: C13ORF18, JAM3, EPB41L3, and TERT). In a triage setting (after a positive Pap smear), sensitivity for CIN2+ of the best combination of genes (C13ORF18/JAM3/ANKRD18CP) (74 %) was comparable to hrHPV testing (79 %), while specificity was significantly higher (76 % versus 42 %, p ≤ 0.05). In addition, in hrHPV-positive scrapings, sensitivity and specificity for CIN2+ of this best-performing combination was comparable to the population referred with abnormal Pap smear.

Conclusions

We identified new CIN2/3-specific methylation markers using genome-wide DNA methylation analysis. The diagnostic performance of our new methylation panel shows higher specificity, which should result in prevention of unnecessary colposcopies for women referred with abnormal cytology. In addition, these newly found markers might be applied as a triage test in hrHPV-positive women from population-based screening. The next step before implementation in primary screening programs will be validation in population-based cohorts.

Electronic supplementary material

The online version of this article (doi:10.1186/s13148-016-0196-3) contains supplementary material, which is available to authorized users.

Concordant dysregulation of miR-5p and miR-3p arms of the same precursor microRNA may be a mechanism in inducing cell proliferation and tumorigenesis: a lung cancer study

A precursor microRNA (miRNA) has two arms: miR-5p and miR-3p (miR-5p/-3p). Depending on the tissue or cell types, both arms can become functional. However, little is known about their coregulatory mechanisms during the tumorigenic process. Here, by using the large-scale miRNA expression profiles of five cancer types, we revealed that several of miR-5p/-3p arms were concordantly dysregulated in each cancer. To explore possible coregulatory mechanisms of concordantly dysregulated miR-5p/-3p pairs, we developed a robust computational framework and applied it to lung cancer data. The framework deciphers miR-5p/-3p coregulated protein interaction networks critical to lung cancer development. As a novel part in the method, we uniquely applied the second-order partial correlation to minimize false-positive regulations. Using 279 matched miRNA and mRNA expression profiles extracted from tumor and normal lung tissue samples, we identified 17 aberrantly expressed miR-5p/-3p pairs that potentially modulate the gene expression of 35 protein complexes. Functional analyses revealed that these complexes are associated with cancer-related biological processes, suggesting the oncogenic potential of the reported miR-5p/-3p pairs. Specifically, we revealed that the reduced expression of miR-145-5p/-3p pair potentially contributes to elevated expression of genes in the "FOXM1 transcription factor network" pathway, which may consequently lead to uncontrolled cell proliferation. Subsequently, the regulation of miR-145-5p/-3p in the FOXM1signaling pathway was validated by a cohort of 104 matched miRNA and protein (reverse-phase protein array) expression profiles in lung cancer. In summary, our computational framework provides a novel tool to study miR-5p/-3p coregulatory mechanisms in cancer and other diseases.

Cosplicing network analysis of mammalian brain RNA-Seq data utilizing WGCNA and Mantel correlations

Across species and tissues and especially in the mammalian brain, production of gene isoforms is widespread. While gene expression coordination has been previously described as a scale-free coexpression network, the properties of transcriptome-wide isoform production coordination have been less studied. Here we evaluate the system-level properties of cosplicing in mouse, macaque, and human brain gene expression data using a novel network inference procedure. Genes are represented as vectors/lists of exon counts and distance measures sensitive to exon inclusion rates quantifies differences across samples. For all gene pairs, distance matrices are correlated across samples, resulting in cosplicing or cotranscriptional network matrices. We show that networks including cosplicing information are scale-free and distinct from coexpression. In the networks capturing cosplicing we find a set of novel hubs with unique characteristics distinguishing them from coexpression hubs: heavy representation in neurobiological functional pathways, strong overlap with markers of neurons and neuroglia, long coding lengths, and high number of both exons and annotated transcripts. Further, the cosplicing hubs are enriched in genes associated with autism spectrum disorders. Cosplicing hub homologs across eukaryotes show dramatically increasing intronic lengths but stable coding region lengths. Shared transcription factor binding sites increase coexpression but not cosplicing; the reverse is true for splicing-factor binding sites. Genes with protein-protein interactions have strong coexpression and cosplicing. Additional factors affecting the networks include shared microRNA binding sites, spatial colocalization within the striatum, and sharing a chromosomal folding domain. Cosplicing network patterns remain relatively stable across species.

High throughput RNA sequencing of a hybrid maize and its parents shows different mechanisms responsive to nitrogen limitation

BACKGROUND

Development of crop varieties with high nitrogen use efficiency (NUE) is crucial for minimizing N loss, reducing environmental pollution and decreasing input cost. Maize is one of the most important crops cultivated worldwide and its productivity is closely linked to the amount of fertilizer used. A survey of the transcriptomes of shoot and root tissues of a maize hybrid line and its two parental inbred lines grown under sufficient and limiting N conditions by mRNA-Seq has been conducted to have a better understanding of how different maize genotypes respond to N limitation.

RESULTS

A different set of genes were found to be N-responsive in the three genotypes. Many biological processes important for N metabolism such as the cellular nitrogen compound metabolic process and the cellular amino acid metabolic process were enriched in the N-responsive gene list from the hybrid shoots but not from the parental lines' shoots. Coupled to this, sugar, carbohydrate, monosaccharide, glucose, and sorbitol transport pathways were all up-regulated in the hybrid, but not in the parents under N limitation. Expression patterns also differed between shoots and roots, such as the up-regulation of the cytokinin degradation pathway in the shoots of the hybrid and down-regulation of that pathway in the roots. The change of gene expression under N limitation in the hybrid resembled the parent with the higher NUE trait. The transcript abundances of alleles derived from each parent were estimated using polymorphic sites in mapped reads in the hybrid. While there were allele abundance differences, there was no correlation between these and the expression differences seen between the hybrid and the two parents.

CONCLUSIONS

Gene expression in two parental inbreds and the corresponding hybrid line in response to N limitation was surveyed using the mRNA-Seq technology. The data showed that the three genotypes respond very differently to N-limiting conditions, and the hybrid clearly has a unique expression pattern compared to its parents. Our results expand our current understanding of N responses and will help move us forward towards effective strategies to improve NUE and enhance crop production.

Incorporating significant amino acid pairs and protein domains to predict RNA splicing-related proteins with functional roles

Machinery of pre-mRNA splicing is carried out through the interaction of RNA sequence elements and a variety of RNA splicing-related proteins (SRPs) (e.g. spliceosome and splicing factors). Alternative splicing, which is an important post-transcriptional regulation in eukaryotes, gives rise to multiple mature mRNA isoforms, which encodes proteins with functional diversities. However, the regulation of RNA splicing is not yet fully elucidated, partly because SRPs have not yet been exhaustively identified and the experimental identification is labor-intensive. Therefore, we are motivated to design a new method for identifying SRPs with their functional roles in the regulation of RNA splicing. The experimentally verified SRPs were manually curated from research articles. According to the functional annotation of Splicing Related Gene Database, the collected SRPs were further categorized into four functional groups including small nuclear Ribonucleoprotein, Splicing Factor, Splicing Regulation Factor and Novel Spliceosome Protein. The composition of amino acid pairs indicates that there are remarkable differences among four functional groups of SRPs. Then, support vector machines (SVMs) were utilized to learn the predictive models for identifying SRPs as well as their functional roles. The cross-validation evaluation presents that the SVM models trained with significant amino acid pairs and functional domains could provide a better predictive performance. In addition, the independent testing demonstrates that the proposed method could accurately identify SRPs in mammals/plants as well as effectively distinguish between SRPs and RNA-binding proteins. This investigation provides a practical means to identifying potential SRPs and a perspective for exploring the regulation of RNA splicing.

Coexpression and cosplicing network approaches for the study of mammalian brain transcriptomes

Next-generation sequencing experiments have demonstrated great potential for transcriptome profiling. While transcriptome sequencing greatly increases the level of biological detail, system-level analysis of these high-dimensional datasets is becoming essential. We illustrate gene network approaches to the analysis of transcriptional data, with particular focus on the advantage of RNA-Seq technology compared to microarray platforms. We introduce a novel methodology for constructing cosplicing networks, based on distance measures combined with matrix correlations. We find that the cosplicing network is distinct and complementary to the coexpression network, although it shares the scale-free properties. In the cosplicing network, we find a set of novel hubs that have unique characteristics distinguishing them from coexpression hubs: they are heavily represented in neurobiological functional pathways and have strong overlap with markers of neurons and neuroglia, long-coding lengths, and high number of both exons and annotated transcripts. We also find that gene networks are plastic in the face of genetic and environmental pressures.

Network link prediction by global silencing of indirect correlations

Predictions of physical and functional links between cellular components are often based on correlations between experimental measurements, such as gene expression. However, correlations are affected by both direct and indirect paths, confounding our ability to identify true pairwise interactions. Here we exploit the fundamental properties of dynamical correlations in networks to develop a method to silence indirect effects. The method receives as input the observed correlations between node pairs and uses a matrix transformation to turn the correlation matrix into a highly discriminative silenced matrix, which enhances only the terms associated with direct causal links. Against empirical data for Escherichia coli regulatory interactions, the method enhanced the discriminative power of the correlations by twofold, yielding >50% predictive improvement over traditional correlation measures and 6% over mutual information. Overall this silencing method will help translate the abundant correlation data into insights about a system's interactions, with applications ranging from link prediction to inferring the dynamical mechanisms governing biological networks.

Modeling exon expression using histone modifications

Histones undergo numerous covalent modifications that play important roles in regulating gene expression. Previous investigations have focused on the effects of histone modifications on gene promoters, whereas efforts to unravel their effects on transcribed regions have lagged behind. To elucidate the effects of histone modification on transcribed regions, we constructed a quantitative model, which we suggest can predict the variation of gene expression more faithfully than the model constructed on promoters. Moreover, motivated by the fact that exon spicing is functionally coupled to transcription, we also devised a quantitative model to predict alternative exon expression using histone modifications on exons. This model was found to be general across different exon types and even cell types. Furthermore, an interaction network linking histone modifications to alternative exon expression was constructed using partial correlations. The network indicated that gene expression and specific histone modifications (H3K36me3 and H4K20me1) could directly influence the exon expression, while other modifications could act in an additive way to account for the stability and robustness. In addition, our results suggest that combinations of histone modifications contribute to exon splicing in a redundant and cumulative fashion. To conclude, this study provides a better understanding of the effects of histone modifications on gene transcribed regions.

A broadly applicable high-throughput screening strategy identifies new regulators of Dlg4 (Psd-95) alternative splicing

Most mammalian genes produce multiple mRNA isoforms derived from alternative pre-mRNA splicing, with each alternative exon controlled by a complex network of regulatory factors. The identification of these regulators can be laborious and is usually carried out one factor at a time. We have developed a broadly applicable high-throughput screening method that simultaneously identifies multiple positive and negative regulators of a particular exon. Two minigene reporters were constructed: One produces green fluorescent protein (GFP) from the mRNA including an exon, and red fluorescent protein (RFP) from the mRNA lacking the exon; the other switches these fluorescent products of exon inclusion and exclusion. Combining results from these two reporters eliminates many false positives and greatly enriches for true splicing regulators. After extensive optimization of this method, we performed a gain-of-function screen of 15,779 cDNA clones and identified 40 genes affecting exon 18 of Discs large homolog 4 (Dlg4; also known as post-synaptic density protein 95 [Psd-95]). We confirmed that 28 of the 34 recoverable clones alter reporter splicing in RT-PCR assays. Remarkably, 18 of the identified genes encode splicing factors or RNA binding proteins, including PTBP1, a previously identified regulator of this exon. Loss-of-function experiments examining endogenous Dlg4 transcripts validated the effects of five of eight genes tested in independent cell lines, and two genes were further confirmed to regulate Dlg4 splicing in primary neurons. These results identify multiple new regulators of Dlg4 splicing, and validate an approach to isolating splicing regulators for almost any cassette exon from libraries of cDNAs, shRNAs, or small molecules.

Complex evolutionary relationships among four classes of modular RNA-binding splicing regulators in eukaryotes: the hnRNP, SR, ELAV-like and CELF proteins

Alternative RNA splicing in multicellular organisms is regulated by a large group of proteins of mainly unknown origin. To predict the functions of these proteins, classification of their domains at the sequence and structural level is necessary. We have focused on four groups of splicing regulators, the heterogeneous nuclear ribonucleoprotein (hnRNP), serine-arginine (SR), embryonic lethal, abnormal vision (ELAV)-like, and CUG-BP and ETR-like factor (CELF) proteins, that show increasing diversity among metazoa. Sequence and phylogenetic analyses were used to obtain a broader understanding of their evolutionary relationships. Surprisingly, when we characterised sequence similarities across full-length sequences and conserved domains of ten metazoan species, we found some hnRNPs were more closely related to SR, ELAV-like and CELF proteins than to other hnRNPs. Phylogenetic analyses and the distribution of the RRM domains suggest that these proteins diversified before the last common ancestor of the metazoans studied here through domain acquisition and duplication to create genes of mixed evolutionary origin. We propose that these proteins were derived independently rather than through the expansion of a single protein family. Our results highlight inconsistencies in the current classification system for these regulators, which does not adequately reflect their evolutionary relationships, and suggests that a domain-based classification scheme may have more utility.

Affymetrix GeneChip microarray preprocessing for multivariate analyses

Affymetrix GeneChip microarrays are the most widely used high-throughput technology to measure gene expression, and a wide variety of preprocessing methods have been developed to transform probe intensities reported by a microarray scanner into gene expression estimates. There have been numerous comparisons of these preprocessing methods, focusing on the most common analyses-detection of differential expression and gene or sample clustering. Recently, more complex multivariate analyses, such as gene co-expression, differential co-expression, gene set analysis and network modeling, are becoming more common; however, the same preprocessing methods are typically applied. In this article, we examine the effect of preprocessing methods on some of these multivariate analyses and provide guidance to the user as to which methods are most appropriate.

Integrating many co-splicing networks to reconstruct splicing regulatory modules

BACKGROUND

Alternative splicing is a ubiquitous gene regulatory mechanism that dramatically increases the complexity of the proteome. However, the mechanism for regulating alternative splicing is poorly understood, and study of coordinated splicing regulation has been limited to individual cases. To study genome-wide splicing regulation, we integrate many human RNA-seq datasets to identify splicing module, which we define as a set of cassette exons co-regulated by the same splicing factors.

RESULTS

We have designed a tensor-based approach to identify co-splicing clusters that appear frequently across multiple conditions, thus very likely to represent splicing modules - a unit in the splicing regulatory network. In particular, we model each RNA-seq dataset as a co-splicing network, where the nodes represent exons and the edges are weighted by the correlations between exon inclusion rate profiles. We apply our tensor-based method to the 38 co-splicing networks derived from human RNA-seq datasets and indentify an atlas of frequent co-splicing clusters. We demonstrate that these identified clusters represent potential splicing modules by validating against four biological knowledge databases. The likelihood that a frequent co-splicing cluster is biologically meaningful increases with its recurrence across multiple datasets, highlighting the importance of the integrative approach.

CONCLUSIONS

Co-splicing clusters reveal novel functional groups which cannot be identified by co-expression clusters, particularly they can grant new insights into functions associated with post-transcriptional regulation, and the same exons can dynamically participate in different pathways depending on different conditions and different other exons that are co-spliced. We propose that by identifying splicing module, a unit in the splicing regulatory network can serve as an important step to decipher the splicing code.

Incorporating evolutionary information and functional domains for identifying RNA splicing factors in humans

Regulation of pre-mRNA splicing is achieved through the interaction of RNA sequence elements and a variety of RNA-splicing related proteins (splicing factors). The splicing machinery in humans is not yet fully elucidated, partly because splicing factors in humans have not been exhaustively identified. Furthermore, experimental methods for splicing factor identification are time-consuming and lab-intensive. Although many computational methods have been proposed for the identification of RNA-binding proteins, there exists no development that focuses on the identification of RNA-splicing related proteins so far. Therefore, we are motivated to design a method that focuses on the identification of human splicing factors using experimentally verified splicing factors. The investigation of amino acid composition reveals that there are remarkable differences between splicing factors and non-splicing proteins. A support vector machine (SVM) is utilized to construct a predictive model, and the five-fold cross-validation evaluation indicates that the SVM model trained with amino acid composition could provide a promising accuracy (80.22%). Another basic feature, amino acid dipeptide composition, is also examined to yield a similar predictive performance to amino acid composition. In addition, this work presents that the incorporation of evolutionary information and domain information could improve the predictive performance. The constructed models have been demonstrated to effectively classify (73.65% accuracy) an independent data set of human splicing factors. The result of independent testing indicates that in silico identification could be a feasible means of conducting preliminary analyses of splicing factors and significantly reducing the number of potential targets that require further in vivo or in vitro confirmation.

miRTar: an integrated system for identifying miRNA-target interactions in human

Background

MicroRNAs (miRNAs) are small non-coding RNA molecules that are ~22-nt-long sequences capable of suppressing protein synthesis. Previous research has suggested that miRNAs regulate 30% or more of the human protein-coding genes. The aim of this work is to consider various analyzing scenarios in the identification of miRNA-target interactions, as well as to provide an integrated system that will aid in facilitating investigation on the influence of miRNA targets by alternative splicing and the biological function of miRNAs in biological pathways.

Results

This work presents an integrated system, miRTar, which adopts various analyzing scenarios to identify putative miRNA target sites of the gene transcripts and elucidates the biological functions of miRNAs toward their targets in biological pathways. The system has three major features. First, the prediction system is able to consider various analyzing scenarios (1 miRNA:1 gene, 1:N, N:1, N:M, all miRNAs:N genes, and N miRNAs: genes involved in a pathway) to easily identify the regulatory relationships between interesting miRNAs and their targets, in 3'UTR, 5'UTR and coding regions. Second, miRTar can analyze and highlight a group of miRNA-regulated genes that participate in particular KEGG pathways to elucidate the biological roles of miRNAs in biological pathways. Third, miRTar can provide further information for elucidating the miRNA regulation, i.e., miRNA-target interactions, affected by alternative splicing.

Conclusions

In this work, we developed an integrated resource, miRTar, to enable biologists to easily identify the biological functions and regulatory relationships between a group of known/putative miRNAs and protein coding genes. miRTar is now available at http://miRTar.mbc.nctu.edu.tw/.

Revealing epigenetic patterns in gene regulation through integrative analysis of epigenetic interaction network

Chromatin components and DNA methylation play important roles in regulation of gene expression in mammalian genomes. However, the mechanism underlying how they regulate gene transcription, independently or synergistically, remains largely unknown. We constructed an epigenetic interaction network (EIN) of chromatin components, DNA methylation and gene expression by combining partial correlation coefficient with Pearson correlation coefficient. In EIN, we identified nine direct factors for gene expression. They constitute three interaction modules which synergistically affect gene expression. We introduced a new combination strategy to test how these direct factors in each module regulate gene expression synergistically. We found two inter-attracted patterns and one inter-repulsed patterns among the three modules. Furthermore, we identified 22 indirect factors for gene expression which have effect on gene expression via direct factors. DNA methylation, for example, could regulate gene expression through H3K4me3 and Pol II. Our approach has the potential to help in uncovering inherent relationships between epigenetic factors and gene transcription and guiding experiment.

ComB: SNP calling and mapping analysis for color and nucleotide space platforms

The determination of single nucleotide polymorphisms (SNPs) has become faster and more cost effective since the advent of short read data from next generation sequencing platforms such as Roche's 454 Sequencer, Illumina's Solexa platform, and Applied Biosystems SOLiD sequencer. The SOLiD sequencing platform, which is capable of producing more than 6 GB of sequence data in a single run, uses a unique encoding scheme where color reads represent transitions between adjacent nucleotides. The determination of SNPs from color reads usually involves the translation of color alignments to likely nucleotide strings to facilitate the use of tools designed for nucleotide reads. This technique results in the loss of significant information in the color read, producing many incorrect SNP calls, especially if regions exist with dense or adjacent polymorphism. Additionally, color reads align ambiguously and incorrectly more often than nucleotide reads making integrated SNP calling a difficult challenge. We have developed ComB, a SNP calling tool which operates directly in color space, using a Bayesian model to incorporate unique and ambiguous reads to iteratively determine SNP identity. ComB is capable of accurately calling short consecutive nucleotide polymorphisms and densely clustered SNPs; both of which other SNP calling tools fail to identify. ComB, which is capable of using billions of short reads to accurately and efficiently perform whole human genome SNP calling in parallel, is also capable of using sequence data or even integrating sequence and color space data sets. We use real and simulated data to demonstrate that ComB's iterative strategy and recalibration of quality scores allow it to discover more true SNPs while calling fewer false positives than tools which use only color alignments as well as tools which translate color reads to nucleotide strings.

Dominating clasp of the financial sector revealed by partial correlation analysis of the stock market

What are the dominant stocks which drive the correlations present among stocks traded in a stock market? Can a correlation analysis provide an answer to this question? In the past, correlation based networks have been proposed as a tool to uncover the underlying backbone of the market. Correlation based networks represent the stocks and their relationships, which are then investigated using different network theory methodologies. Here we introduce a new concept to tackle the above question—the partial correlation network. Partial correlation is a measure of how the correlation between two variables, e.g., stock returns, is affected by a third variable. By using it we define a proxy of stock influence, which is then used to construct partial correlation networks. The empirical part of this study is performed on a specific financial system, namely the set of 300 highly capitalized stocks traded at the New York Stock Exchange, in the time period 2001–2003. By constructing the partial correlation network, unlike the case of standard correlation based networks, we find that stocks belonging to the financial sector and, in particular, to the investment services sub-sector, are the most influential stocks affecting the correlation profile of the system. Using a moving window analysis, we find that the strong influence of the financial stocks is conserved across time for the investigated trading period. Our findings shed a new light on the underlying mechanisms and driving forces controlling the correlation profile observed in a financial market.

Identification of two APOBEC3F splice variants displaying HIV-1 antiviral activity and contrasting sensitivity to Vif

Approximately half of all human genes undergo alternative mRNA splicing. This process often yields homologous gene products exhibiting diverse functions. Alternative splicing of APOBEC3G (A3G) and APOBEC3F (A3F), the major host resistance factors targeted by the HIV-1 protein Vif, has not been explored. We investigated the effects of alternative splicing on A3G/A3F gene expression and antiviral activity. Three alternatively spliced A3G mRNAs and two alternatively spliced A3F mRNAs were detected in peripheral blood mononuclear cells in each of 10 uninfected, healthy donors. Expression of these splice variants was altered in different cell subsets and in response to cellular stimulation. Alternatively spliced A3G variants were insensitive to degradation by Vif but displayed no antiviral activity against HIV-1. Conversely, alternative splicing of A3F produced a 37-kDa variant lacking exon 2 (A3FΔ2) that was prominently expressed in macrophages and monocytes and was resistant to Vif-mediated degradation. Alternative splicing also produced a 24-kDa variant of A3F lacking exons 2-4 (A3FΔ2-4) that was highly sensitive to Vif. Both A3FΔ2 and A3FΔ2-4 displayed reduced cytidine deaminase activity and moderate antiviral activity. These alternatively spliced A3F gene products, particularly A3FΔ2, were incorporated into HIV virions, albeit at levels less than wild-type A3F. Thus, alternative splicing of A3F mRNA generates truncated antiviral proteins that differ sharply in their sensitivity to Vif.

Elevated NCOR1 disrupts PPARalpha/gamma signaling in prostate cancer and forms a targetable epigenetic lesion

The loss of anti-proliferative responsiveness in prostate cancer cell lines toward ligands for vitamin D receptor, retinoic acid receptors/retinoid X receptors and peroxisome proliferator activated receptor (PPAR)alpha/gamma may entail underlying epigenetic events, as ligand insensitivity reflects significantly altered messenger RNA expression of corepressors and histone-modifying enzymes. Expression patterns were dependent on phases of the cell cycle and associated with repressed basal gene expression of vitamin D receptor and PPARalpha/gamma target genes, for example CDKN1A [encodes p21((waf1/cip1))]. Elevated nuclear corepressor 1 (NCOR1) and nuclear corepressor 2/silencing mediator of retinoic acid and thyroid hormone receptor protein levels were detected in prostate cancer cell lines compared with non-malignant counterparts. Knockdown of the corepressor NCOR1 significantly elevated basal expression of a cohort of target genes, including CDKN1A. Both chemical [histone deacetylases inhibitor (HDACi)] and NCOR1 knockdown targeting enhanced anti-proliferative sensitivity toward PPARalpha/gamma ligands in prostate cancer cell lines. Pursuing PPARalpha/gamma signaling, microarray approaches were undertaken to identify pathways and genes regulated uniquely by a combination of PPARalpha/gamma activation and HDAC inhibition. Again, HDACi and knockdown approaches demonstrated that elevated NCOR1 expression and activity distorted PPARalpha/gamma gene targets centered on, for example cell cycle control, including CDKN1A and TGFBRAP1. Quantitative real time polymerase chain reaction validation and chromatin immunoprecipitation assays both confirmed that elevated NCOR1 disrupted the ability of PPARalpha/gamma to regulate key target genes (CDKN1A and TGFBRAP1). Interrogation of these relationships in prostate cancer samples using principal component and partial correlation analyses established significant interdependent relationships between NCOR1-PPARalpha/gamma and representative target genes, independently of androgen receptor expression. Therefore, we conclude that elevated NCOR1 distorts the actions of PPARalpha/gamma selectively and generates a potential epigenetic lesion with diagnostic and prognostic significance.

Alternative splicing of transcription factors' genes: beyond the increase of proteome diversity

Functional modification of transcription regulators may lead to developmental changes and phenotypical differences between species. In this work, we study the influence of alternative splicing on transcription factors in human and mouse. Our results show that the impact of alternative splicing on transcription factors is similar in both species, meaning that the ways to increase variability should also be similar. However, when looking at the expression patterns of transcription factors, we observe that they tend to diverge regardless of the role of alternative splicing. Finally, we hypothesise that transcription regulation of alternatively spliced transcription factors could play an important role in the phenotypical differences between species, without discarding other phenomena or functional families.

Large-scale analysis of exonized mammalian-wide interspersed repeats in primate genomes

Transposable elements (TEs) are major sources of new exons in higher eukaryotes. Almost half of the human genome is derived from TEs, and many types of TEs have the potential to exonize. In this work, we conducted a large-scale analysis of human exons derived from mammalian-wide interspersed repeats (MIRs), a class of old TEs which was active prior to the radiation of placental mammals. Using exon array data of 328 MIR-derived exons and RT-PCR analysis of 39 exons in 10 tissues, we identified 15 constitutively spliced MIR exons, and 15 MIR exons with tissue-specific shift in splicing patterns. Analysis of RNAs from multiple species suggests that the splicing events of many strongly included MIR exons have been established before the divergence of primates and rodents, while a small percentage result from recent exonization during primate evolution. Interestingly, exon array data suggest substantially higher splicing activities of MIR exons when compared with exons derived from Alu elements, a class of primate-specific retrotransposons. This appears to be a universal difference between exons derived from young and old TEs, as it is also observed when comparing Alu exons to exons derived from LINE1 and LINE2, two other groups of old TEs. Together, this study significantly expands current knowledge about exonization of TEs. Our data imply that with sufficient evolutionary time, numerous new exons could evolve beyond the evolutionary intermediate state and contribute functional novelties to modern mammalian genomes.