Discovery of Allele-Specific Protein-RNA Interactions in Human Transcriptomes

AStruct: detection of allele-specific RNA secondary structure in structuromic probing data

BACKGROUND

Uncovering functional genetic variants from an allele-specific perspective is of paramount importance in advancing our understanding of gene regulation and genetic diseases. Recently, various allele-specific events, such as allele-specific gene expression, allele-specific methylation, and allele-specific binding, have been explored on a genome-wide scale due to the development of high-throughput sequencing methods. RNA secondary structure, which plays a crucial role in multiple RNA-associated processes like RNA modification, translation and splicing, has emerged as an essential focus of relevant research. However, tools to identify genetic variants associated with allele-specific RNA secondary structures are still lacking.

RESULTS

Here, we develop a computational tool called 'AStruct' that enables us to detect allele-specific RNA secondary structure (ASRS) from RT-stop based structuromic probing data. AStruct shows robust performance in both simulated datasets and public icSHAPE datasets. We reveal that single nucleotide polymorphisms (SNPs) with higher AStruct scores are enriched in coding regions and tend to be functional. These SNPs are highly conservative, have the potential to disrupt sites involved in m6A modification or protein binding, and are frequently associated with disease.

CONCLUSIONS

AStruct is a tool dedicated to invoke allele-specific RNA secondary structure events at heterozygous SNPs in RT-stop based structuromic probing data. It utilizes allelic variants, base pairing and RT-stop information under different cell conditions to detect dynamic and functional ASRS. Compared to sequence-based tools, AStruct considers dynamic cell conditions and outperforms in detecting functional variants. AStruct is implemented in JAVA and is freely accessible at: https://github.com/canceromics/AStruct .

Nuclear and cytoplasmic specific RNA binding proteome enrichment and its changes upon ferroptosis induction

The key role of RNA-binding proteins (RBPs) in posttranscriptional regulation of gene expression is intimately tied to their subcellular localization. Here, we show a subcellular-specific RNA labeling method for efficient enrichment and deep profiling of nuclear and cytoplasmic RBPs. A total of 1221 nuclear RBPs and 1333 cytoplasmic RBPs were enriched and identified using nuclear/cytoplasm targeting enrichment probes, representing an increase of 54.4% and 85.7% compared with previous reports. The probes were further applied in the omics-level investigation of subcellular-specific RBP-RNA interactions upon ferroptosis induction. Interestingly, large-scale RBPs display enhanced interaction with RNAs in nucleus but reduced association with RNAs in cytoplasm during ferroptosis process. Furthermore, we discovered dozens of nucleoplasmic translocation candidate RBPs upon ferroptosis induction and validated representative ones by immunofluorescence imaging. The enrichment of Tricarboxylic acid cycle in the translocation candidate RBPs may provide insights for investigating their possible roles in ferroptosis induced metabolism dysregulation.

Effect of single nucleotide polymorphisms on the structure of long noncoding RNAs and their interaction with RNA binding proteins

Long non-coding RNAs (lncRNA) are emerging as a new class of regulatory RNAs with remarkable potential to be utilized as therapeutic targets against many human diseases. Several genome-wide association studies (GWAS) have catalogued Single Nucleotide Polymorphisms (SNPs) present in the noncoding regions of the genome from where lncRNAs originate. In this study, we have selected 67 lncRNAs with GWAS-tagged SNPs and have also investigated their role in affecting the local secondary structures. Majority of the SNPs lead to changes in the secondary structure of lncRNAs to a different extent by altering the base pairing patterns. These structural changes in lncRNA are also manifested in form of alteration in the binding site for RNA binding proteins (RBPs) along with affecting their binding efficacies. Ultimately, these structural modifications may influence the transcriptional and post-transcriptional pathways of these RNAs, leading to the causation of diseases. Hence, it is important to understand the possible underlying mechanism of RBPs in association with GWAS-tagged SNPs in human diseases.

Allele-specific RNA N 6-methyladenosine modifications reveal functional genetic variants in human tissues

An intricate network of cis- and trans-elements acts on RNA N 6-methyladenosine (m6A), which in turn may affect gene expression and, ultimately, human health. A complete understanding of this network requires new approaches to accurately measure the subtle m6A differences arising from genetic variants, many of which have been associated with common diseases. To address this gap, we developed a method to accurately and sensitively detect transcriptome-wide allele-specific m6A (ASm6A) from MeRIP-seq data and applied it to uncover 12,056 high-confidence ASm6A modifications from 25 human tissues. We also identified 1184 putative functional variants for ASm6A regulation, a subset of which we experimentally validated. Importantly, we found that many of these ASm6A-associated genetic variants were enriched for common disease-associated and complex trait-associated risk loci, and verified that two disease risk variants can change m6A modification status. Together, this work provides a tool to detangle the dynamic network of RNA modifications at the allelic level and highlights the interplay of m6A and genetics in human health and disease.

Diagnostic and therapeutic potential of RNASET2 in Crohn's disease: Disease-risk polymorphism modulates allelic-imbalance in expression and circulating protein levels and recombinant-RNASET2 attenuates pro-inflammatory cytokine secretion

Ribonuclease T2 gene (RNASET2) variants are associated in genome wide association studies (GWAS) with risk for several autoimmune diseases, including Crohn's disease (CD). In T cells, a functional and biological relationship exists between TNFSF15-mediated enhancement of IFN-γ production, mucosal inflammation and RNASET2. Disease risk variants are associated with decreased mRNA expression and clinical characteristics of severe CD; however, functional classifications of variants and underlying molecular mechanisms contributing to pathogenesis remain largely unknown. In this study we demonstrate that allelic imbalance of RNASET2 disease risk variant rs2149092 is associated with transcriptional and post-transcriptional mechanisms regulating transcription factor binding, promoter-transactivation and allele-specific expression. RNASET2 mRNA expression decreases in response to multiple modes of T cell activation and recovers following elimination of activator. In CD patients with severe disease necessitating surgical intervention, preoperative circulating RNASET2 protein levels were decreased compared to non-IBD subjects and rebounded post-operatively following removal of the inflamed region, with levels associated with allelic carriage. Furthermore, overexpression or treatment with recombinant RNASET2 significantly reduced IFN-γ secretion. These findings reveal that RNASET2 cis- and trans-acting variation contributed regulatory complexity and determined expression and provide a basis for linking genetic variation with CD pathobiology. These data may ultimately identify RNASET2 as an effective therapeutic target in a subset of CD patients with severe disease.

Inferring RNA-binding protein target preferences using adversarial domain adaptation

Precise identification of target sites of RNA-binding proteins (RBP) is important to understand their biochemical and cellular functions. A large amount of experimental data is generated by in vivo and in vitro approaches. The binding preferences determined from these platforms share similar patterns but there are discernable differences between these datasets. Computational methods trained on one dataset do not always work well on another dataset. To address this problem which resembles the classic "domain shift" in deep learning, we adopted the adversarial domain adaptation (ADDA) technique and developed a framework (RBP-ADDA) that can extract RBP binding preferences from an integration of in vivo and vitro datasets. Compared with conventional methods, ADDA has the advantage of working with two input datasets, as it trains the initial neural network for each dataset individually, projects the two datasets onto a feature space, and uses an adversarial framework to derive an optimal network that achieves an optimal discriminative predictive power. In the first step, for each RBP, we include only the in vitro data to pre-train a source network and a task predictor. Next, for the same RBP, we initiate the target network by using the source network and use adversarial domain adaptation to update the target network using both in vitro and in vivo data. These two steps help leverage the in vitro data to improve the prediction on in vivo data, which is typically challenging with a lower signal-to-noise ratio. Finally, to further take the advantage of the fused source and target data, we fine-tune the task predictor using both data. We showed that RBP-ADDA achieved better performance in modeling in vivo RBP binding data than other existing methods as judged by Pearson correlations. It also improved predictive performance on in vitro datasets. We further applied augmentation operations on RBPs with less in vivo data to expand the input data and showed that it can improve prediction performances. Lastly, we explored the predictive interpretability of RBP-ADDA, where we quantified the contribution of the input features by Integrated Gradients and identified nucleotide positions that are important for RBP recognition.

Perspectives on Allele-Specific Expression

Diploidy has profound implications for population genetics and susceptibility to genetic diseases. Although two copies are present for most genes in the human genome, they are not necessarily both active or active at the same level in a given individual. Genomic imprinting, resulting in exclusive or biased expression in favor of the allele of paternal or maternal origin, is now believed to affect hundreds of human genes. A far greater number of genes display unequal expression of gene copies due to cis-acting genetic variants that perturb gene expression. The availability of data generated by RNA sequencing applied to large numbers of individuals and tissue types has generated unprecedented opportunities to assess the contribution of genetic variation to allelic imbalance in gene expression. Here we review the insights gained through the analysis of these data about the extent of the genetic contribution to allelic expression imbalance, the tools and statistical models for gene expression imbalance, and what the results obtained reveal about the contribution of genetic variants that alter gene expression to complex human diseases and phenotypes.

Mutations in cis that affect mRNA synthesis, processing and translation

Genetic mutations that cause hereditary diseases usually affect the composition of the transcribed mRNA and its encoded protein, leading to instability of the mRNA and/or the protein. Sometimes, however, such mutations affect the synthesis, the processing or the translation of the mRNA, with similar disastrous effects. We here present an overview of mRNA synthesis, its posttranscriptional modification and its translation into protein. We then indicate which elements in these processes are known to be affected by pathogenic mutations, but we restrict our review to mutations in cis, in the DNA of the gene that encodes the affected protein. These mutations can be in enhancer or promoter regions of the gene, which act as binding sites for transcription factors involved in pre-mRNA synthesis. We also describe mutations in polyadenylation sequences and in splice site regions, exonic and intronic, involved in intron removal. Finally, we include mutations in the Kozak sequence in mRNA, which is involved in protein synthesis. We provide examples of genetic diseases caused by mutations in these DNA regions and refer to databases to help identify these regions. The over-all knowledge of mRNA synthesis, processing and translation is essential for improvement of the diagnosis of patients with genetic diseases.

Predicting dynamic cellular protein-RNA interactions by deep learning using in vivo RNA structures

Interactions with RNA-binding proteins (RBPs) are integral to RNA function and cellular regulation, and dynamically reflect specific cellular conditions. However, presently available tools for predicting RBP-RNA interactions employ RNA sequence and/or predicted RNA structures, and therefore do not capture their condition-dependent nature. Here, after profiling transcriptome-wide in vivo RNA secondary structures in seven cell types, we developed PrismNet, a deep learning tool that integrates experimental in vivo RNA structure data and RBP binding data for matched cells to accurately predict dynamic RBP binding in various cellular conditions. PrismNet results for 168 RBPs support its utility for both understanding CLIP-seq results and largely extending such interaction data to accurately analyze additional cell types. Further, PrismNet employs an "attention" strategy to computationally identify exact RBP-binding nucleotides, and we discovered enrichment among dynamic RBP-binding sites for structure-changing variants (riboSNitches), which can link genetic diseases with dysregulated RBP bindings. Our rich profiling data and deep learning-based prediction tool provide access to a previously inaccessible layer of cell-type-specific RBP-RNA interactions, with clear utility for understanding and treating human diseases.

Integrative analyses of transcriptome data reveal the mechanisms of post-transcriptional regulation

Post-transcriptional processing of RNAs plays important roles in a variety of physiological and pathological processes. These processes can be precisely controlled by a series of RNA binding proteins and cotranscriptionally regulated by transcription factors as well as histone modifications. With the rapid development of high-throughput sequencing techniques, multiomics data have been broadly used to study the mechanisms underlying the important biological processes. However, how to use these high-throughput sequencing data to elucidate the fundamental regulatory roles of post-transcriptional processes is still of great challenge. This review summarizes the regulatory mechanisms of post-transcriptional processes and the general principles and approaches to dissect these mechanisms by integrating multiomics data as well as public resources.

NCBP3/SNHG6 inhibits GBX2 transcription in a histone modification manner to facilitate the malignant biological behaviour of glioma cells

RNA-binding proteins (RBPs) are significantly dysregulated in glioma. In this study, we demonstrated the upregulation of Nuclear cap-binding subunit 3 (NCBP3) in glioma tissues and cells. Further, knockdown of NCBP3 inhibited the malignant progression of glioma. NCBP3 directly bound to small nucleolar RNA host gene 6 (SNHG6) and stabilized SNHG6 expression. In contrast, the gastrulation brain homeobox 2 (GBX2) transcription factor was downregulated in glioma tissues and cells. SNHG6 inhibited GBX2 transcription by mediating the H3K27me3 modification induced by polycomb repressive complex 2 (PRC2). Moreover, GBX2 decreased the promoter activities and downregulated the expression of the flotillin protein family 1 (FLOT1) oncogene. In conclusion, NCBP3/SNHG6 inhibits GBX2 transcription in a PRC2-dependent manner to facilitate the malignant progression of gliomas.

Principles of RNA processing from analysis of enhanced CLIP maps for 150 RNA binding proteins

BACKGROUND

A critical step in uncovering rules of RNA processing is to study the in vivo regulatory networks of RNA binding proteins (RBPs). Crosslinking and immunoprecipitation (CLIP) methods enable mapping RBP targets transcriptome-wide, but methodological differences present challenges to large-scale analysis across datasets. The development of enhanced CLIP (eCLIP) enabled the mapping of targets for 150 RBPs in K562 and HepG2, creating a unique resource of RBP interactomes profiled with a standardized methodology in the same cell types.

RESULTS

Our analysis of 223 eCLIP datasets reveals a range of binding modalities, including highly resolved positioning around splicing signals and mRNA untranslated regions that associate with distinct RBP functions. Quantification of enrichment for repetitive and abundant multicopy elements reveals 70% of RBPs have enrichment for non-mRNA element classes, enables identification of novel ribosomal RNA processing factors and sites, and suggests that association with retrotransposable elements reflects multiple RBP mechanisms of action. Analysis of spliceosomal RBPs indicates that eCLIP resolves AQR association after intronic lariat formation, enabling identification of branch points with single-nucleotide resolution, and provides genome-wide validation for a branch point-based scanning model for 3' splice site recognition. Finally, we show that eCLIP peak co-occurrences across RBPs enable the discovery of novel co-interacting RBPs.

CONCLUSIONS

This work reveals novel insights into RNA biology by integrated analysis of eCLIP profiling of 150 RBPs with distinct functions. Further, our quantification of both mRNA and other element association will enable further research to identify novel roles of RBPs in regulating RNA processing.

Prevalence and architecture of posttranscriptionally impaired synonymous mutations in 8,320 genomes across 22 cancer types

Somatic synonymous mutations are one of the most frequent genetic variants occurring in the coding region of cancer genomes, while their contributions to cancer development remain largely unknown. To assess whether synonymous mutations involved in post-transcriptional regulation contribute to the genetic etiology of cancers, we collected whole exome data from 8,320 patients across 22 cancer types. By employing our developed algorithm, PIVar, we identified a total of 22,948 posttranscriptionally impaired synonymous SNVs (pisSNVs) spanning 2,042 genes. In addition, 35 RNA binding proteins impacted by these identified pisSNVs were significantly enriched. Remarkably, we discovered markedly elevated ratio of somatic pisSNVs across all 22 cancer types, and a high pisSNV ratio was associated with worse patient survival in five cancer types. Intriguing, several well-established cancer genes, including PTEN, RB1 and PIK3CA, appeared to contribute to tumorigenesis at both protein function and posttranscriptional regulation levels, whereas some pisSNV-hosted genes, including UBR4, EP400 and INTS1, exerted their function during carcinogenesis mainly via posttranscriptional mechanisms. Moreover, we predicted three drugs associated with two pisSNVs, and numerous compounds associated with expression signature of pisSNV-hosted genes. Our study reveals the prevalence and clinical relevance of pisSNVs in cancers, and emphasizes the importance of considering posttranscriptional impaired synonymous mutations in cancer biology.

Deep neural networks for interpreting RNA-binding protein target preferences

Deep learning has become a powerful paradigm to analyze the binding sites of regulatory factors including RNA-binding proteins (RBPs), owing to its strength to learn complex features from possibly multiple sources of raw data. However, the interpretability of these models, which is crucial to improve our understanding of RBP binding preferences and functions, has not yet been investigated in significant detail. We have designed a multitask and multimodal deep neural network for characterizing in vivo RBP targets. The model incorporates not only the sequence but also the region type of the binding sites as input, which helps the model to boost the prediction performance. To interpret the model, we quantified the contribution of the input features to the predictive score of each RBP. Learning across multiple RBPs at once, we are able to avoid experimental biases and to identify the RNA sequence motifs and transcript context patterns that are the most important for the predictions of each individual RBP. Our findings are consistent with known motifs and binding behaviors and can provide new insights about the regulatory functions of RBPs.

Gain-of-Function Mutations: An Emerging Advantage for Cancer Biology

Advances in next-generation sequencing have identified thousands of genomic variants that perturb the normal functions of proteins, further contributing to diverse phenotypic consequences in cancer. Elucidating the functional pathways altered by loss-of-function (LOF) or gain-of-function (GOF) mutations will be crucial for prioritizing cancer-causing variants and their resultant therapeutic liabilities. In this review, we highlight the fundamental function of GOF mutations and discuss the potential mechanistic effects in the context of signaling networks. We also summarize advances in experimental and computational resources, which will dramatically help with studies on the functional and phenotypic consequences of mutations. Together, systematic investigations of the function of GOF mutations will provide an important missing piece for cancer biology and precision therapy.