Extensible benchmarking of methods that identify and quantify polyadenylation sites from RNA-seq data.

The tremendous rate with which data is generated and analysis methods emerge makes it increasingly difficult to keep track of their domain of applicability, assumptions, limitations, and consequently, of the efficacy and precision with which they solve specific tasks. Therefore, there is an increasing need for benchmarks, and for the provision of infrastructure for continuous method evaluation. APAeval is an international community effort, organized by the RNA Society in 2021, to benchmark tools for the identification and quantification of the usage of alternative polyadenylation (APA) sites from short-read, bulk RNA-sequencing (RNA-seq) data. Here, we reviewed 17 tools and benchmarked eight on their ability to perform APA identification and quantification, using a comprehensive set of RNA-seq experiments comprising real, synthetic, and matched 3'-end sequencing data. To support continuous benchmarking, we have incorporated the results into the OpenEBench online platform, which allows for continuous extension of the set of methods, metrics, and challenges. We envisage that our analyses will assist researchers in selecting the appropriate tools for their studies, while the containers and reproducible workflows could easily be deployed and extended to evaluate new methods or data sets.

Widespread JNK-dependent alternative splicing induces a positive feedback loop through CELF2-mediated regulation of MKK7 during T-cell activation.

Alternative splicing is prevalent among genes encoding signaling molecules; however, the functional consequence of differential isoform expression remains largely unknown. Here we demonstrate that, in response to T-cell activation, the Jun kinase (JNK) kinase MAP kinase kinase 7 (MKK7) is alternatively spliced to favor an isoform that lacks exon 2. This isoform restores a JNK-docking site within MKK7 that is disrupted in the larger isoform. Consistently, we show that skipping of MKK7 exon 2 enhances JNK pathway activity, as indicated by c-Jun phosphorylation and up-regulation of TNF-α. Moreover, this splicing event is itself dependent on JNK signaling. Thus, MKK7 alternative splicing represents a positive feedback loop through which JNK promotes its own signaling. We further show that repression of MKK7 exon 2 is dependent on the presence of flanking sequences and the JNK-induced expression of the RNA-binding protein CELF2, which binds to these regulatory elements. Finally, we found that ∼25% of T-cell receptor-mediated alternative splicing events are dependent on JNK signaling. Strikingly, these JNK-dependent events are also significantly enriched for responsiveness to CELF2. Together, our data demonstrate a widespread role for the JNK-CELF2 axis in controlling splicing during T-cell activation, including a specific role in propagating JNK signaling.

Deep profiling and custom databases improve detection of proteoforms generated by alternative splicing.

Alternative pre-mRNA splicing has long been proposed to contribute greatly to proteome complexity. However, the extent to which mature mRNA isoforms are successfully translated into protein remains controversial. Here, we used high-throughput RNA sequencing and mass spectrometry (MS)-based proteomics to better evaluate the translation of alternatively spliced mRNAs. To increase proteome coverage and improve protein quantitation, we optimized cell fractionation and sample processing steps at both the protein and peptide level. Furthermore, we generated a custom peptide database trained on analysis of RNA-seq data with MAJIQ, an algorithm optimized to detect and quantify differential and unannotated splice junction usage. We matched tandem mass spectra acquired by data-dependent acquisition (DDA) against our custom RNA-seq based database, as well as SWISS-PROT and RefSeq databases to improve identification of splicing-derived proteoforms by 28% compared with use of the SWISS-PROT database alone. Altogether, we identified peptide evidence for 554 alternate proteoforms corresponding to 274 genes. Our increased depth and detection of proteins also allowed us to track changes in the transcriptome and proteome induced by T-cell stimulation, as well as fluctuations in protein subcellular localization. In sum, our data here confirm that use of generic databases in proteomic studies underestimates the number of spliced mRNA isoforms that are translated into protein and provides a workflow that improves isoform detection in large-scale proteomic experiments.

Reciprocal regulation of hnRNP C and CELF2 through translation and transcription tunes splicing activity in T cells.

RNA binding proteins (RBPs) frequently regulate the expression of other RBPs in mammalian cells. Such cross-regulation has been proposed to be important to control networks of coordinated gene expression; however, much remains to be understood about how such networks of cross-regulation are established and what the functional consequence is of coordinated or reciprocal expression of RBPs. Here we demonstrate that the RBPs CELF2 and hnRNP C regulate the expression of each other, such that depletion of one results in reduced expression of the other. Specifically, we show that loss of hnRNP C reduces the transcription of CELF2 mRNA, while loss of CELF2 results in decreased efficiency of hnRNP C translation. We further demonstrate that this reciprocal regulation serves to fine tune the splicing patterns of many downstream target genes. Together, this work reveals new activities of hnRNP C and CELF2, provides insight into a previously unrecognized gene regulatory network, and demonstrates how cross-regulation of RBPs functions to shape the cellular transcriptome.

Ancient antagonism between CELF and RBFOX families tunes mRNA splicing outcomes.

Over 95% of human multi-exon genes undergo alternative splicing, a process important in normal development and often dysregulated in disease. We sought to analyze the global splicing regulatory network of CELF2 in human T cells, a well-studied splicing regulator critical to T cell development and function. By integrating high-throughput sequencing data for binding and splicing quantification with sequence features and probabilistic splicing code models, we find evidence of splicing antagonism between CELF2 and the RBFOX family of splicing factors. We validate this functional antagonism through knockdown and overexpression experiments in human cells and find CELF2 represses RBFOX2 mRNA and protein levels. Because both families of proteins have been implicated in the development and maintenance of neuronal, muscle, and heart tissues, we analyzed publicly available data in these systems. Our analysis suggests global, antagonistic coregulation of splicing by the CELF and RBFOX proteins in mouse muscle and heart in several physiologically relevant targets, including proteins involved in calcium signaling and members of the MEF2 family of transcription factors. Importantly, a number of these coregulated events are aberrantly spliced in mouse models and human patients with diseases that affect these tissues, including heart failure, diabetes, or myotonic dystrophy. Finally, analysis of exons regulated by ancient CELF family homologs in chicken, Drosophila, and Caenorhabditis elegans suggests this antagonism is conserved throughout evolution.

Aberrant splicing in B-cell acute lymphoblastic leukemia.

Aberrant splicing is a hallmark of leukemias with mutations in splicing factor (SF)-encoding genes. Here we investigated its prevalence in pediatric B-cell acute lymphoblastic leukemias (B-ALL), where SFs are not mutated. By comparing these samples to normal pro-B cells, we found thousands of aberrant local splice variations (LSVs) per sample, with 279 LSVs in 241 genes present in every comparison. These genes were enriched in RNA processing pathways and encoded ∼100 SFs, e.g. hnRNPA1. HNRNPA1 3'UTR was most pervasively mis-spliced, yielding the transcript subject to nonsense-mediated decay. To mimic this event, we knocked it down in B-lymphoblastoid cells and identified 213 hnRNPA1-regulated exon usage events comprising the hnRNPA1 splicing signature in pediatric leukemia. Some of its elements were LSVs in DICER1 and NT5C2, known cancer drivers. We searched for LSVs in other leukemia and lymphoma drivers and discovered 81 LSVs in 41 additional genes. Seventy-seven LSVs out of 81 were confirmed using two large independent B-ALL RNA-seq datasets, and the twenty most common B-ALL drivers, including NT5C2, showed higher prevalence of aberrant splicing than of somatic mutations. Thus, post-transcriptional deregulation of SF can drive widespread changes in B-ALL splicing and likely contributes to disease pathogenesis.

MAJIQ-SPEL: web-tool to interrogate classical and complex splicing variations from RNA-Seq data.

SUMMARY: Analysis of RNA sequencing (RNA-Seq) data have highlighted the fact that most genes undergo alternative splicing (AS) and that these patterns are tightly regulated. Many of these events are complex, resulting in numerous possible isoforms that quickly become difficult to visualize, interpret and experimentally validate. To address these challenges we developed MAJIQ-SPEL, a web-tool that takes as input local splicing variations (LSVs) quantified from RNA-Seq data and provides users with visualization and quantification of gene isoforms associated with those. Importantly, MAJIQ-SPEL is able to handle both classical (binary) and complex, non-binary, splicing variations. Using a matching primer design algorithm it also suggests to users possible primers for experimental validation by RT-PCR and displays those, along with the matching protein domains affected by the LSV, on UCSC Genome Browser for further downstream analysis. AVAILABILITY AND IMPLEMENTATION: Program and code will be available at http://majiq.biociphers.org/majiq-spel. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Integrative deep models for alternative splicing.

MOTIVATION: Advancements in sequencing technologies have highlighted the role of alternative splicing (AS) in increasing transcriptome complexity. This role of AS, combined with the relation of aberrant splicing to malignant states, motivated two streams of research, experimental and computational. The first involves a myriad of techniques such as RNA-Seq and CLIP-Seq to identify splicing regulators and their putative targets. The second involves probabilistic models, also known as splicing codes, which infer regulatory mechanisms and predict splicing outcome directly from genomic sequence. To date, these models have utilized only expression data. In this work, we address two related challenges: Can we improve on previous models for AS outcome prediction and can we integrate additional sources of data to improve predictions for AS regulatory factors. RESULTS: We perform a detailed comparison of two previous modeling approaches, Bayesian and Deep Neural networks, dissecting the confounding effects of datasets and target functions. We then develop a new target function for AS prediction in exon skipping events and show it significantly improves model accuracy. Next, we develop a modeling framework that leverages transfer learning to incorporate CLIP-Seq, knockdown and over expression experiments, which are inherently noisy and suffer from missing values. Using several datasets involving key splice factors in mouse brain, muscle and heart we demonstrate both the prediction improvements and biological insights offered by our new models. Overall, the framework we propose offers a scalable integrative solution to improve splicing code modeling as vast amounts of relevant genomic data become available. AVAILABILITY AND IMPLEMENTATION: Code and data available at: majiq.biociphers.org/jha_et_al_2017/. CONTACT: yosephb@upenn.edu. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Induced transcription and stability of CELF2 mRNA drives widespread alternative splicing during T-cell signaling.

Studies in several cell types have highlighted dramatic and diverse changes in mRNA processing that occur upon cellular stimulation. However, the mechanisms and pathways that lead to regulated changes in mRNA processing remain poorly understood. Here we demonstrate that expression of the splicing factor CELF2 (CUGBP, Elav-like family member 2) is regulated in response to T-cell signaling through combined increases in transcription and mRNA stability. Transcriptional induction occurs within 6 h of stimulation and is dependent on activation of NF-κB. Subsequently, there is an increase in the stability of the CELF2 mRNA that correlates with a change in CELF2 3'UTR length and contributes to the total signal-induced enhancement of CELF2 expression. Importantly, we uncover dozens of splicing events in cultured T cells whose changes upon stimulation are dependent on CELF2 expression, and provide evidence that CELF2 controls a similar proportion of splicing events during human thymic T-cell development. Taken together, these findings expand the physiologic impact of CELF2 beyond that previously documented in developing neuronal and muscle cells to T-cell development and function, identify unappreciated instances of alternative splicing in the human thymus, and uncover novel mechanisms for CELF2 regulation that may broadly impact CELF2 expression across diverse cell types.

Position-dependent activity of CELF2 in the regulation of splicing and implications for signal-responsive regulation in T cells.

CELF2 is an RNA binding protein that has been implicated in developmental and signal-dependent splicing in the heart, brain and T cells. In the heart, CELF2 expression decreases during development, while in T cells CELF2 expression increases both during development and in response to antigen-induced signaling events. Although hundreds of CELF2-responsive splicing events have been identified in both heart and T cells, the way in which CELF2 functions has not been broadly investigated. Here we use CLIP-Seq to identified physical targets of CELF2 in a cultured human T cell line. By comparing the results with known functional targets of CELF2 splicing regulation from the same cell line we demonstrate a generalizable position-dependence of CELF2 activity that is consistent with previous mechanistic studies of individual CELF2 target genes in heart and brain. Strikingly, this general position-dependence is sufficient to explain the bi-directional activity of CELF2 on 2 T cell targets recently reported. Therefore, we propose that the location of CELF2 binding around an exon is a primary predictor of CELF2 function in a broad range of cellular contexts.

In silico to in vivo splicing analysis using splicing code models.

With the growing appreciation of RNA splicing's role in gene regulation, development, and disease, researchers from diverse fields find themselves investigating exons of interest. Commonly, researchers are interested in knowing if an exon is alternatively spliced, if it is differentially included in specific tissues or in developmental stages, and what regulatory elements control its inclusion. An important step towards the ability to perform such analysis in silico was made with the development of computational splicing code models. Aimed as a practical how-to guide, we demonstrate how researchers can now use these code models to analyze a gene of interest, focusing on Bin1 as a case study. Bridging integrator 1 (BIN1) is a nucleocytoplasmic adaptor protein known to be functionally regulated through alternative splicing in a tissue-specific manner. Specific Bin1 isoforms have been associated with muscular diseases and cancers, making the study of its splicing regulation of wide interest. Using AVISPA, a recently released web tool based on splicing code models, we show that many Bin1 tissue-dependent isoforms are correctly predicted, along with many of its known regulators. We review the best practices and constraints of using the tool, demonstrate how AVISPA is used to generate high confidence novel regulatory hypotheses, and experimentally validate predicted regulators of Bin1 alternative splicing.

Alternative 3'UTR expression induced by T cell activation is regulated in a temporal and signal dependent manner.

The length of 3' untranslated regions (3'UTR) is highly regulated during many transitions in cell state, including T cell activation, through the process of alternative polyadenylation (APA). However, the regulatory mechanisms and functional consequences of APA remain largely unexplored. Here we present a detailed analysis of the temporal and condition-specific regulation of APA following activation of primary human CD4+ T cells. We find that global APA changes are regulated temporally and CD28 costimulatory signals enhance a subset of these changes. Most APA changes upon T cell activation involve 3'UTR shortening, although a set of genes enriched for function in the mTOR pathway exhibit 3'UTR lengthening. While upregulation of the core polyadenylation machinery likely induces 3'UTR shortening following prolonged T cell stimulation; a significant program of APA changes occur prior to cellular proliferation or upregulation of the APA machinery. Motif analysis suggests that at least a subset of these early changes in APA are driven by upregulation of RBM3, an RNA-binding protein which competes with the APA machinery for binding. Together this work expands our understanding of the impact and mechanisms of APA in response to T cell activation and suggests new mechanisms by which APA may be regulated.

A Deep Dive into Statistical Modeling of RNA Splicing QTLs Reveals New Variants that Explain Neurodegenerative Disease.

Genome-wide association studies (GWAS) have identified thousands of putative disease causing variants with unknown regulatory effects. Efforts to connect these variants with splicing quantitative trait loci (sQTLs) have provided functional insights, yet sQTLs reported by existing methods cannot explain many GWAS signals. We show current sQTL modeling approaches can be improved by considering alternative splicing representation, model calibration, and covariate integration. We then introduce MAJIQTL, a new pipeline for sQTL discovery. MAJIQTL includes two new statistical methods: a weighted multiple testing approach for sGene discovery and a model for sQTL effect size inference to improve variant prioritization. By applying MAJIQTL to GTEx, we find significantly more sGenes harboring sQTLs with functional significance. Notably, our analysis implicates the novel variant rs582283 in Alzheimer's disease. Using antisense oligonucleotides, we validate this variant's effect by blocking the implicated YBX3 binding site, leading to exon skipping in the gene MS4A3.

An anciently diverged family of RNA binding proteins maintain correct splicing of a class of ultra-long exons through cryptic splice site repression.

Previously, we showed that the germ cell-specific nuclear protein RBMXL2 represses cryptic splicing patterns during meiosis and is required for male fertility (Ehrmann et al., 2019). Here, we show that in somatic cells the similar yet ubiquitously expressed RBMX protein has similar functions. RBMX regulates a distinct class of exons that exceed the median human exon size. RBMX protein-RNA interactions are enriched within ultra-long exons, particularly within genes involved in genome stability, and repress the selection of cryptic splice sites that would compromise gene function. The RBMX gene is silenced during male meiosis due to sex chromosome inactivation. To test whether RBMXL2 might replace the function of RBMX during meiosis we induced expression of RBMXL2 and the more distantly related RBMY protein in somatic cells, finding each could rescue aberrant patterns of RNA processing caused by RBMX depletion. The C-terminal disordered domain of RBMXL2 is sufficient to rescue proper splicing control after RBMX depletion. Our data indicate that RBMX and RBMXL2 have parallel roles in somatic tissues and the germline that must have been conserved for at least 200 million years of mammalian evolution. We propose RBMX family proteins are particularly important for the splicing inclusion of some ultra-long exons with increased intrinsic susceptibility to cryptic splice site selection.

Alternative splicing of HDAC7 regulates its interaction with 14-3-3 proteins to alter histone marks and target gene expression.

Chromatin regulation and alternative splicing are both critical mechanisms guiding gene expression. Studies have demonstrated that histone modifications can influence alternative splicing decisions, but less is known about how alternative splicing may impact chromatin. Here, we demonstrate that several genes encoding histone-modifying enzymes are alternatively spliced downstream of T cell signaling pathways, including HDAC7, a gene previously implicated in controlling gene expression and differentiation in T cells. Using CRISPR-Cas9 gene editing and cDNA expression, we show that differential inclusion of HDAC7 exon 9 controls the interaction of HDAC7 with protein chaperones, resulting in changes to histone modifications and gene expression. Notably, the long isoform, which is induced by the RNA-binding protein CELF2, promotes expression of several critical T cell surface proteins including CD3, CD28, and CD69. Thus, we demonstrate that alternative splicing of HDAC7 has a global impact on histone modification and gene expression that contributes to T cell development.

RNA splicing analysis using heterogeneous and large RNA-seq datasets.

The ubiquity of RNA-seq has led to many methods that use RNA-seq data to analyze variations in RNA splicing. However, available methods are not well suited for handling heterogeneous and large datasets. Such datasets scale to thousands of samples across dozens of experimental conditions, exhibit increased variability compared to biological replicates, and involve thousands of unannotated splice variants resulting in increased transcriptome complexity. We describe here a suite of algorithms and tools implemented in the MAJIQ v2 package to address challenges in detection, quantification, and visualization of splicing variations from such datasets. Using both large scale synthetic data and GTEx v8 as benchmark datasets, we assess the advantages of MAJIQ v2 compared to existing methods. We then apply MAJIQ v2 package to analyze differential splicing across 2,335 samples from 13 brain subregions, demonstrating its ability to offer insights into brain subregion-specific splicing regulation.

Structural basis for high-order complex of SARNP and DDX39B to facilitate mRNP assembly.

mRNA in eukaryotic cells is packaged into highly compacted ribonucleoprotein particles (mRNPs) in the nucleus and exported to the cytoplasm for translation. mRNP packaging and export require the evolutionarily conserved transcription-export (TREX) complex. TREX facilitates loading of various RNA-binding proteins on mRNA through the action of its DDX39B subunit. SARNP (Tho1 [transcriptional defect of Hpr1 by overexpression 1] in yeast) is shown to interact with DDX39B and affect mRNA export. The molecular mechanism of how SARNP recognizes DDX39B and functions in mRNP assembly is unclear. Here, we determine the crystal structure of a Tho1/DDX39B/RNA complex, revealing a multivalent interaction mediated by tandem DDX39B interacting motifs in SARNP/Tho1. The high-order complex of SARNP and DDX39B is evolutionarily conserved, and human SARNP can engage with five DDX39B molecules. RNA sequencing (RNA-seq) from SARNP knockdown cells shows the most affected RNAs in export are GC rich. Our work suggests the role of the high-order SARNP/DDX39B/RNA complex in mRNP assembly and export.

Influenza virus mRNAs encode determinants for nuclear export via the cellular TREX-2 complex.

Nuclear export of influenza A virus (IAV) mRNAs occurs through the nuclear pore complex (NPC). Using the Auxin-Induced Degron (AID) system to rapidly degrade proteins, we show that among the nucleoporins localized at the nucleoplasmic side of the NPC, TPR is the key nucleoporin required for nuclear export of influenza virus mRNAs. TPR recruits the TRanscription and EXport complex (TREX)-2 to the NPC for exporting a subset of cellular mRNAs. By degrading components of the TREX-2 complex (GANP, Germinal-center Associated Nuclear Protein; PCID2, PCI domain containing 2), we show that influenza mRNAs require the TREX-2 complex for nuclear export and replication. Furthermore, we found that cellular mRNAs whose export is dependent on GANP have a small number of exons, a high mean exon length, long 3' UTR, and low GC content. Some of these features are shared by influenza virus mRNAs. Additionally, we identified a 45 nucleotide RNA signal from influenza virus HA mRNA that is sufficient to mediate GANP-dependent mRNA export. Thus, we report a role for the TREX-2 complex in nuclear export of influenza mRNAs and identified RNA determinants associated with the TREX-2-dependent mRNA export.

Extensible benchmarking of methods that identify and quantify polyadenylation sites from RNA-seq data.

The tremendous rate with which data is generated and analysis methods emerge makes it increasingly difficult to keep track of their domain of applicability, assumptions, and limitations and consequently, of the efficacy and precision with which they solve specific tasks. Therefore, there is an increasing need for benchmarks, and for the provision of infrastructure for continuous method evaluation. APAeval is an international community effort, organized by the RNA Society in 2021, to benchmark tools for the identification and quantification of the usage of alternative polyadenylation (APA) sites from short-read, bulk RNA-sequencing (RNA-seq) data. Here, we reviewed 17 tools and benchmarked eight on their ability to perform APA identification and quantification, using a comprehensive set of RNA-seq experiments comprising real, synthetic, and matched 3'-end sequencing data. To support continuous benchmarking, we have incorporated the results into the OpenEBench online platform, which allows for seamless extension of the set of methods, metrics, and challenges. We envisage that our analyses will assist researchers in selecting the appropriate tools for their studies. Furthermore, the containers and reproducible workflows generated in the course of this project can be seamlessly deployed and extended in the future to evaluate new methods or datasets.