A deep learning approach to identify gene targets of a therapeutic for human splicing disorders

The impact and future of artificial intelligence in medical genetics and molecular medicine: an ongoing revolution

Artificial intelligence (AI) platforms have emerged as pivotal tools in genetics and molecular medicine, as in many other fields. The growth in patient data, identification of new diseases and phenotypes, discovery of new intracellular pathways, availability of greater sets of omics data, and the need to continuously analyse them have led to the development of new AI platforms. AI continues to weave its way into the fabric of genetics with the potential to unlock new discoveries and enhance patient care. This technology is setting the stage for breakthroughs across various domains, including dysmorphology, rare hereditary diseases, cancers, clinical microbiomics, the investigation of zoonotic diseases, omics studies in all medical disciplines. AI's role in facilitating a deeper understanding of these areas heralds a new era of personalised medicine, where treatments and diagnoses are tailored to the individual's molecular features, offering a more precise approach to combating genetic or acquired disorders. The significance of these AI platforms is growing as they assist healthcare professionals in the diagnostic and treatment processes, marking a pivotal shift towards more informed, efficient, and effective medical practice. In this review, we will explore the range of AI tools available and show how they have become vital in various sectors of genomic research supporting clinical decisions.

Age-dependent regulation of ELP1 exon 20 splicing in Familial Dysautonomia by RNA Polymerase II kinetics and chromatin structure

Familial Dysautonomia (FD) is a rare disease caused by ELP1 exon 20 skipping. Here we clarify the role of RNA Polymerase II (RNAPII) and chromatin on this splicing event. A slow RNAPII mutant and chromatin-modifying chemicals that reduce the rate of RNAPII elongation induce exon skipping whereas chemicals that create a more relaxed chromatin exon inclusion. In the brain of a mouse transgenic for the human FD-ELP1 we observed on this gene an age-dependent decrease in the RNAPII density profile that was most pronounced on the alternative exon, a robust increase in the repressive marks H3K27me3 and H3K9me3 and a decrease of H3K27Ac, together with a progressive reduction in ELP1 exon 20 inclusion level. In HEK 293T cells, selective drug-induced demethylation of H3K27 increased RNAPII elongation on ELP1 and SMN2, promoted the inclusion of the corresponding alternative exons, and, by RNA-sequencing analysis, induced changes in several alternative splicing events. These data suggest a co-transcriptional model of splicing regulation in which age-dependent changes in H3K27me3/Ac modify the rate of RNAPII elongation and affect processing of ELP1 alternative exon 20.

Big data and deep learning for RNA biology

The exponential growth of big data in RNA biology (RB) has led to the development of deep learning (DL) models that have driven crucial discoveries. As constantly evidenced by DL studies in other fields, the successful implementation of DL in RB depends heavily on the effective utilization of large-scale datasets from public databases. In achieving this goal, data encoding methods, learning algorithms, and techniques that align well with biological domain knowledge have played pivotal roles. In this review, we provide guiding principles for applying these DL concepts to various problems in RB by demonstrating successful examples and associated methodologies. We also discuss the remaining challenges in developing DL models for RB and suggest strategies to overcome these challenges. Overall, this review aims to illuminate the compelling potential of DL for RB and ways to apply this powerful technology to investigate the intriguing biology of RNA more effectively.

Splice modulators target PMS1 to reduce somatic expansion of the Huntington's disease-associated CAG repeat

Huntington's disease (HD) is a dominant neurological disorder caused by an expanded HTT exon 1 CAG repeat that lengthens huntingtin's polyglutamine tract. Lowering mutant huntingtin has been proposed for treating HD, but genetic modifiers implicate somatic CAG repeat expansion as the driver of onset. We find that branaplam and risdiplam, small molecule splice modulators that lower huntingtin by promoting HTT pseudoexon inclusion, also decrease expansion of an unstable HTT exon 1 CAG repeat in an engineered cell model. Targeted CRISPR-Cas9 editing shows this effect is not due to huntingtin lowering, pointing instead to pseudoexon inclusion in PMS1. Homozygous but not heterozygous inactivation of PMS1 also reduces CAG repeat expansion, supporting PMS1 as a genetic modifier of HD and a potential target for therapeutic intervention. Although splice modulation provides one strategy, genome-wide transcriptomics also emphasize consideration of cell-type specific effects and polymorphic variation at both target and off-target sites.

Specificity, synergy, and mechanisms of splice-modifying drugs

Drugs that target pre-mRNA splicing hold great therapeutic potential, but the quantitative understanding of how these drugs work is limited. Here we introduce mechanistically interpretable quantitative models for the sequence-specific and concentration-dependent behavior of splice-modifying drugs. Using massively parallel splicing assays, RNA-seq experiments, and precision dose-response curves, we obtain quantitative models for two small-molecule drugs, risdiplam and branaplam, developed for treating spinal muscular atrophy. The results quantitatively characterize the specificities of risdiplam and branaplam for 5' splice site sequences, suggest that branaplam recognizes 5' splice sites via two distinct interaction modes, and contradict the prevailing two-site hypothesis for risdiplam activity at SMN2 exon 7. The results also show that anomalous single-drug cooperativity, as well as multi-drug synergy, are widespread among small-molecule drugs and antisense-oligonucleotide drugs that promote exon inclusion. Our quantitative models thus clarify the mechanisms of existing treatments and provide a basis for the rational development of new therapies.

A drug-free cardiovascular stent functionalized with tailored collagen supports in-situ healing of vascular tissues

Drug-eluting stent implantation suppresses the excessive proliferation of smooth muscle cells to reduce in-stent restenosis. However, the efficacy of drug-eluting stents remains limited due to delayed reendothelialization, impaired intimal remodeling, and potentially increased late restenosis. Here, we show that a drug-free coating formulation functionalized with tailored recombinant humanized type III collagen exerts one-produces-multi effects in response to injured tissue following stent implantation. We demonstrate that the one-produces-multi coating possesses anticoagulation, anti-inflammatory, and intimal hyperplasia suppression properties. We perform transcriptome analysis to indicate that the drug-free coating favors the endothelialization process and induces the conversion of smooth muscle cells to a contractile phenotype. We find that compared to drug-eluting stents, our drug-free stent reduces in-stent restenosis in rabbit and porcine models and improves vascular neointimal healing in a rabbit model. Collectively, the one-produces-multi drug-free system represents a promising strategy for the next-generation of stents.

Deep Learning for Genomics: From Early Neural Nets to Modern Large Language Models

The data explosion driven by advancements in genomic research, such as high-throughput sequencing techniques, is constantly challenging conventional methods used in genomics. In parallel with the urgent demand for robust algorithms, deep learning has succeeded in various fields such as vision, speech, and text processing. Yet genomics entails unique challenges to deep learning, since we expect a superhuman intelligence that explores beyond our knowledge to interpret the genome from deep learning. A powerful deep learning model should rely on the insightful utilization of task-specific knowledge. In this paper, we briefly discuss the strengths of different deep learning models from a genomic perspective so as to fit each particular task with proper deep learning-based architecture, and we remark on practical considerations of developing deep learning architectures for genomics. We also provide a concise review of deep learning applications in various aspects of genomic research and point out current challenges and potential research directions for future genomics applications. We believe the collaborative use of ever-growing diverse data and the fast iteration of deep learning models will continue to contribute to the future of genomics.

Predicting RNA structures and functions by artificial intelligence

RNA functions by interacting with its intended targets structurally. However, due to the dynamic nature of RNA molecules, RNA structures are difficult to determine experimentally or predict computationally. Artificial intelligence (AI) has revolutionized many biomedical fields and has been progressively utilized to deduce RNA structures, target binding, and associated functionality. Integrating structural and target binding information could also help improve the robustness of AI-based RNA function prediction and RNA design. Given the rapid development of deep learning (DL) algorithms, AI will provide an unprecedented opportunity to elucidate the sequence-structure-function relation of RNAs.

The role of alternative pre-mRNA splicing in cancer progression

Alternative pre-mRNA splicing is a critical mechanism that generates multiple mRNA from a single gene, thereby increasing the diversity of the proteome. Recent research has highlighted the significance of specific splicing isoforms in cellular processes, particularly in regulating cell numbers. In this review, we examine the current understanding of the role of alternative splicing in controlling cancer cell growth and discuss specific splicing factors and isoforms and their molecular mechanisms in cancer progression. These isoforms have been found to intricately control signaling pathways crucial for cell cycle progression, proliferation, and apoptosis. Furthermore, studies have elucidated the characteristics and functional importance of splicing factors that influence cell numbers. Abnormal expression of oncogenic splicing isoforms and splicing factors, as well as disruptions in splicing caused by genetic mutations, have been implicated in the development and progression of tumors. Collectively, these findings provide valuable insights into the complex interplay between alternative splicing and cell proliferation, thereby suggesting the potential of alternative splicing as a therapeutic target for cancer.

PMS1 as a target for splice modulation to prevent somatic CAG repeat expansion in Huntington's disease

Huntington's disease (HD) is a dominantly inherited neurodegenerative disorder whose motor, cognitive, and behavioral manifestations are caused by an expanded, somatically unstable CAG repeat in the first exon of HTT that lengthens a polyglutamine tract in huntingtin. Genome-wide association studies (GWAS) have revealed DNA repair genes that influence the age-at-onset of HD and implicate somatic CAG repeat expansion as the primary driver of disease timing. To prevent the consequent neuronal damage, small molecule splice modulators (e.g., branaplam) that target HTT to reduce the levels of huntingtin are being investigated as potential HD therapeutics. We found that the effectiveness of the splice modulators can be influenced by genetic variants, both at HTT and other genes where they promote pseudoexon inclusion. Surprisingly, in a novel hTERT-immortalized retinal pigment epithelial cell (RPE1) model for assessing CAG repeat instability, these drugs also reduced the rate of HTT CAG expansion. We determined that the splice modulators also affect the expression of the mismatch repair gene PMS1, a known modifier of HD age-at-onset. Genome editing at specific HTT and PMS1 sequences using CRISPR-Cas9 nuclease confirmed that branaplam suppresses CAG expansion by promoting the inclusion of a pseudoexon in PMS1, making splice modulation of PMS1 a potential strategy for delaying HD onset. Comparison with another splice modulator, risdiplam, suggests that other genes affected by these splice modulators also influence CAG instability and might provide additional therapeutic targets.

Harnessing deep learning into hidden mutations of neurological disorders for therapeutic challenges

The relevant study of transcriptome-wide variations and neurological disorders in the evolved field of genomic data science is on the rise. Deep learning has been highlighted utilizing algorithms on massive amounts of data in a human-like manner, and is expected to predict the dependency or druggability of hidden mutations within the genome. Enormous mutational variants in coding and noncoding transcripts have been discovered along the genome by far, despite of the fine-tuned genetic proofreading machinery. These variants could be capable of inducing various pathological conditions, including neurological disorders, which require lifelong care. Several limitations and questions emerge, including the use of conventional processes via limited patient-driven sequence acquisitions and decoding-based inferences as well as how rare variants can be deduced as a population-specific etiology. These puzzles require harnessing of advanced systems for precise disease prediction, drug development and drug applications. In this review, we summarize the pathophysiological discoveries of pathogenic variants in both coding and noncoding transcripts in neurological disorders, and the current advantage of deep learning applications. In addition, we discuss the challenges encountered and how to outperform them with advancing interpretation.

SRSF5 Regulates the Expression of BQ323636.1 to Modulate Tamoxifen Resistance in ER-Positive Breast Cancer

About 70% of breast cancer patients are oestrogen receptor-positive (ER +ve). Adjuvant endocrine therapy using tamoxifen (TAM) is an effective approach for preventing local recurrence and metastasis. However, around half of the patients will eventually develop resistance. Overexpression of BQ323636.1 (BQ) is one of the mechanisms that confer TAM resistance. BQ is an alternative splice variant of NCOR2. The inclusion of exon 11 generates mRNA for NCOR2, while the exclusion of exon 11 produces mRNA for BQ. The expression of SRSF5 is low in TAM-resistant breast cancer cells. Modulation of SRSF5 can affect the alternative splicing of NCOR2 to produce BQ. In vitro and in vivo studies confirmed that the knockdown of SRSF5 enhanced BQ expression, and conferred TAM resistance; in contrast, SRSF5 overexpression reduced BQ expression and, thus, reversed TAM resistance. Clinical investigation using a tissue microarray confirmed the inverse correlation of SRSF5 and BQ. Low SRSF5 expression was associated with TAM resistance, local recurrence and metastasis. Survival analyses showed that low SRSF5 expression was associated with poorer prognosis. We showed that SRPK1 can interact with SRSF5 to phosphorylate it. Inhibition of SRPK1 by a small inhibitor, SRPKIN-1, suppressed the phosphorylation of SRSF5. This enhanced the proportion of SRSF5 interacting with exon 11 of NCOR2, reducing the production of BQ mRNA. As expected, SRPKIN-1 reduced TAM resistance. Our study confirms that SRSF5 is essential for BQ expression. Modulating the activity of SRSF5 in ER +ve breast cancer will be a potential approach to combating TAM resistance.

Imaging-based intelligent spectrometer on a plasmonic rainbow chip

Compact, lightweight, and on-chip spectrometers are required to develop portable and handheld sensing and analysis applications. However, the performance of these miniaturized systems is usually much lower than their benchtop laboratory counterparts due to oversimplified optical architectures. Here, we develop a compact plasmonic "rainbow" chip for rapid, accurate dual-functional spectroscopic sensing that can surpass conventional portable spectrometers under selected conditions. The nanostructure consists of one-dimensional or two-dimensional graded metallic gratings. By using a single image obtained by an ordinary camera, this compact system can accurately and precisely determine the spectroscopic and polarimetric information of the illumination spectrum. Assisted by suitably trained deep learning algorithms, we demonstrate the characterization of optical rotatory dispersion of glucose solutions at two-peak and three-peak narrowband illumination across the visible spectrum using just a single image. This system holds the potential for integration with smartphones and lab-on-a-chip systems to develop applications for in situ analysis.

Development of an oral treatment that rescues gait ataxia and retinal degeneration in a phenotypic mouse model of familial dysautonomia

Familial dysautonomia (FD) is a rare neurodegenerative disease caused by a splicing mutation in elongator acetyltransferase complex subunit 1 (ELP1). This mutation leads to the skipping of exon 20 and a tissue-specific reduction of ELP1, mainly in the central and peripheral nervous systems. FD is a complex neurological disorder accompanied by severe gait ataxia and retinal degeneration. There is currently no effective treatment to restore ELP1 production in individuals with FD, and the disease is ultimately fatal. After identifying kinetin as a small molecule able to correct the ELP1 splicing defect, we worked on its optimization to generate novel splicing modulator compounds (SMCs) that can be used in individuals with FD. Here, we optimize the potency, efficacy, and bio-distribution of second-generation kinetin derivatives to develop an oral treatment for FD that can efficiently pass the blood-brain barrier and correct the ELP1 splicing defect in the nervous system. We demonstrate that the novel compound PTC258 efficiently restores correct ELP1 splicing in mouse tissues, including brain, and most importantly, prevents the progressive neuronal degeneration that is characteristic of FD. Postnatal oral administration of PTC258 to the phenotypic mouse model TgFD9;Elp1Δ20/flox increases full-length ELP1 transcript in a dose-dependent manner and leads to a 2-fold increase in functional ELP1 in the brain. Remarkably, PTC258 treatment improves survival, gait ataxia, and retinal degeneration in the phenotypic FD mice. Our findings highlight the great therapeutic potential of this novel class of small molecules as an oral treatment for FD.

Investigation of early molecular alterations in tauopathy with generative adversarial networks

The recent advances in deep learning-based approaches hold great promise for unravelling biological mechanisms, discovering biomarkers, and predicting gene function. Here, we deployed a deep generative model for simulating the molecular progression of tauopathy and dissecting its early features. We applied generative adversarial networks (GANs) for bulk RNA-seq analysis in a mouse model of tauopathy (TPR50-P301S). The union set of differentially expressed genes from four comparisons (two phenotypes with two time points) was used as input training data. We devised four-way transition curves for a virtual simulation of disease progression, clustered and grouped the curves by patterns, and identified eight distinct pattern groups showing different biological features from Gene Ontology enrichment analyses. Genes that were upregulated in early tauopathy were associated with vasculature development, and these changes preceded immune responses. We confirmed significant disease-associated differences in the public human data for the genes of the different pattern groups. Validation with weighted gene co-expression network analysis suggested that our GAN-based approach can be used to detect distinct patterns of early molecular changes during disease progression, which may be extremely difficult in in vivo experiments. The generative model is a valid systematic approach for exploring the sequential cascades of mechanisms and targeting early molecular events related to dementia.

Targeting strategies for modulating pre-mRNA splicing with small molecules: Recent advances

The concept of using small molecules to therapeutically modulate pre-mRNA splicing was validated with the US Food and Drug Administration (FDA) approval of Evrysdi® (risdiplam) in 2020. Since then, efforts have continued unabated toward the discovery of new splicing-modulating drugs. However, the drug development world has evolved in the 10 years since risdiplam precursors were first identified in high-throughput screening (HTS). Now, new mechanistic insights into RNA-processing pathways and regulatory networks afford increasingly feasible targeted approaches. In this review, organized into classes of biological target, we compile and summarize small molecules discovered, devised, and developed since 2020 to alter pre-mRNA splicing.

Functional Assays Reclassify Suspected Splice-Altering Variants of Uncertain Significance in Mendelian Channelopathies

BACKGROUND

Rare protein-altering variants in SCN5A, KCNQ1, and KCNH2 are major causes of Brugada syndrome and the congenital long QT syndrome. While splice-altering variants lying outside 2-bp canonical splice sites can cause these diseases, their role remains poorly described. We implemented 2 functional assays to assess 12 recently reported putative splice-altering variants of uncertain significance and 1 likely pathogenic variant without functional data observed in Brugada syndrome and long QT syndrome probands.

METHODS

We deployed minigene assays to assess the splicing consequences of 10 variants. Three variants incompatible with the minigene approach were introduced into control induced pluripotent stem cells by CRISPR genome editing. We differentiated cells into induced pluripotent stem cell-derived cardiomyocytes and studied splicing outcomes by reverse transcription-polymerase chain reaction. We used the American College of Medical Genetics and Genomics functional assay criteria (PS3/BS3) to reclassify variants.

RESULTS

We identified aberrant splicing, with presumed disruption of protein sequence, in 8/10 variants studied using the minigene assay and 1/3 studied in induced pluripotent stem cell-derived cardiomyocytes. We reclassified 8 variants of uncertain significance to likely pathogenic, 1 variant of uncertain significance to likely benign, and 1 likely pathogenic variant to pathogenic.

CONCLUSIONS

Functional assays reclassified splice-altering variants outside canonical splice sites in Brugada Syndrome- and long QT syndrome-associated genes.

Genetics of Cystic Fibrosis: Clinical Implications

Cystic fibrosis (CF) is a multiorgan disease caused by a wide variety of mutations in the cystic fibrosis transmembrane conductance regulator gene. As treatment has progressed from symptom mitigation to targeting of specific molecular defects, genetics has played an important role in identifying the proper precision therapies for each individual. Novel therapeutic approaches are focused on expanding treatment to a greater number of individuals as well as working toward a cure. This review discusses the role of genetics in our understanding of CF with a particular emphasis on how genetics informs the exciting landscape of current and novel CF therapies.

Rescue of a familial dysautonomia mouse model by AAV9-Exon-specific U1 snRNA

Familial dysautonomia (FD) is a currently untreatable, neurodegenerative disease caused by a splicing mutation (c.2204+6T>C) that causes skipping of exon 20 of the elongator complex protein 1 (ELP1) pre-mRNA. Here, we used adeno-associated virus serotype 9 (AAV9-U1-FD) to deliver an exon-specific U1 (ExSpeU1) small nuclear RNA, designed to cause inclusion of ELP1 exon 20 only in those cells expressing the target pre-mRNA, in a phenotypic mouse model of FD. Postnatal systemic and intracerebral ventricular treatment in these mice increased the inclusion of ELP1 exon 20. This also augmented the production of functional protein in several tissues including brain, dorsal root, and trigeminal ganglia. Crucially, the treatment rescued most of the FD mouse mortality before one month of age (89% vs 52%). There were notable improvements in ataxic gait as well as renal (serum creatinine) and cardiac (ejection fraction) functions. RNA-seq analyses of dorsal root ganglia from treated mice and human cells overexpressing FD-ExSpeU1 revealed only minimal global changes in gene expression and splicing. Overall then, our data prove that AAV9-U1-FD is highly specific and will likely be a safe and effective therapeutic strategy for this debilitating disease.

Long-Read Nanopore Sequencing Identifies Mismatch Repair-Deficient Related Genes with Alternative Splicing in Colorectal Cancer

Background

Alternative splicing (AS) plays a crucial role in regulating the progression of colorectal cancer (CRC), but its distribution remains to be explored. Here, we aim to investigate the genes edited by AS which show differential expression in patients with mismatch repair deficiency (dMMR)/microsatellite instability (MSI).

Materials and Methods

We applied long-read nanopore sequencing to determine the mRNA profiles and screen AS genes using Oxford Nanopore Technologies (ONT) method in ten paired CRC tissues. CRC tissue and plasma samples were used to validate the differential genes with AS using real-time fluorescent quantitative PCR, immunohistochemistry, and enzyme-linked immunosorbent assay.

Results

ONT sequencing identified 404 genes were downregulated, and 348 genes were upregulated in MSI cancer tissues compared with microsatellite stability (MSS) cancer tissues. In total, 6,200 AS events were identified in 2,728 mRNA transcripts. WGCNA revealed dMMR/MSI-correlated gene modules, including INHBA and RPL22L1, which were upregulated; conversely, HMGCS2 was downregulated in MSI cancer. Overexpression of RPL22L1, INHBA, and CAPZA1 was further confirmed in CRC tissues. INHBA was found to be associated with tumor lymphatic metastasis. Importantly, the levels of INHBA in CRC plasma were significantly increased compared with those in noncancer plasma. INHBA showed a higher level in dMMR/MSI CRC than in MSS CRC, indicating that INHBA is a useful biomarker.

Conclusion

Our results showed that ONT-identified genes provide a pool to explore AS-associated markers for dMMR/MSI CRC. We demonstrated INHBA as a promising signature for clinical application in predicting tumor lymphatic metastasis and screening dMMR/MSI candidates.

Selective retinal ganglion cell loss and optic neuropathy in a humanized mouse model of familial dysautonomia

Familial dysautonomia (FD) is an autosomal recessive neurodegenerative disease caused by a splicing mutation in the gene encoding Elongator complex protein 1 (ELP1, also known as IKBKAP). This mutation results in tissue-specific skipping of exon 20 with a corresponding reduction of ELP1 protein, predominantly in the central and peripheral nervous system. Although FD patients have a complex neurological phenotype caused by continuous depletion of sensory and autonomic neurons, progressive visual decline leading to blindness is one of the most problematic aspects of the disease, as it severely affects their quality of life. To better understand the disease mechanism as well as to test the in vivo efficacy of targeted therapies for FD, we have recently generated a novel phenotypic mouse model, TgFD9; IkbkapΔ20/flox. This mouse exhibits most of the clinical features of the disease and accurately recapitulates the tissue-specific splicing defect observed in FD patients. Driven by the dire need to develop therapies targeting retinal degeneration in FD, herein, we comprehensively characterized the progression of the retinal phenotype in this mouse, and we demonstrated that it is possible to correct ELP1 splicing defect in the retina using the splicing modulator compound (SMC) BPN-15477.

In silico Methods for Identification of Potential Therapeutic Targets

At the initial stage of drug discovery, identifying novel targets with maximal efficacy and minimal side effects can improve the success rate and portfolio value of drug discovery projects while simultaneously reducing cycle time and cost. However, harnessing the full potential of big data to narrow the range of plausible targets through existing computational methods remains a key issue in this field. This paper reviews two categories of in silico methods—comparative genomics and network-based methods—for finding potential therapeutic targets among cellular functions based on understanding their related biological processes. In addition to describing the principles, databases, software, and applications, we discuss some recent studies and prospects of the methods. While comparative genomics is mostly applied to infectious diseases, network-based methods can be applied to infectious and non-infectious diseases. Nonetheless, the methods often complement each other in their advantages and disadvantages. The information reported here guides toward improving the application of big data-driven computational methods for therapeutic target discovery.

Graphical abstract

Roles of Physicochemical and Structural Properties of RNA-Binding Proteins in Predicting the Activities of Trans-Acting Splicing Factors with Machine Learning

Trans-acting splicing factors play a pivotal role in modulating alternative splicing by specifically binding to cis-elements in pre-mRNAs. There are approximately 1500 RNA-binding proteins (RBPs) in the human genome, but the activities of these RBPs in alternative splicing are unknown. Since determining RBP activities through experimental methods is expensive and time consuming, the development of an efficient computational method for predicting the activities of RBPs in alternative splicing from their sequences is of great practical importance. Recently, a machine learning model for predicting the activities of splicing factors was built based on features of single and dual amino acid compositions. Here, we explored the role of physicochemical and structural properties in predicting their activities in alternative splicing using machine learning approaches and found that the prediction performance is significantly improved by including these properties. By combining the minimum redundancy–maximum relevance (mRMR) method and forward feature searching strategy, a promising feature subset with 24 features was obtained to predict the activities of RBPs. The feature subset consists of 16 dual amino acid compositions, 5 physicochemical features, and 3 structural features. The physicochemical and structural properties were as important as the sequence composition features for an accurate prediction of the activities of splicing factors. The hydrophobicity and distribution of coil are suggested to be the key physicochemical and structural features, respectively.

Reprogramming RNA processing: an emerging therapeutic landscape

The production of a mature mRNA requires coordination of multiple processing steps, which ultimately control its content, localization, and stability. These steps include some of the largest macromolecular machines in the cell, which were, until recently, considered undruggable due to their biological complexity. Building from an expanded understanding of the underlying mechanisms that drive these processes, a new wave of therapeutics is seeking to target RNA processing. With a focus on impacting gene regulation at the RNA level, such modalities offer potential for sequence-specific resolution in drug design. Here, we review our current understanding of RNA-processing events and their role in gene regulation, with a focus on the therapeutic opportunities that have emerged within this landscape.