Human 5' UTR design and variant effect prediction from a massively parallel translation assay

Characterization and optimization of 5´ untranslated region containing poly-adenine tracts in Kluyveromyces marxianus using machine-learning model

BACKGROUND

The 5´ untranslated region (5´ UTR) plays a key role in regulating translation efficiency and mRNA stability, making it a favored target in genetic engineering and synthetic biology. A common feature found in the 5´ UTR is the poly-adenine (poly(A)) tract. However, the effect of 5´ UTR poly(A) on protein production remains controversial. Machine-learning models are powerful tools for explaining the complex contributions of features, but models incorporating features of 5´ UTR poly(A) are currently lacking. Thus, our goal is to construct such a model, using natural 5´ UTRs from Kluyveromyces marxianus, a promising cell factory for producing heterologous proteins.

RESULTS

We constructed a mini-library consisting of 207 5´ UTRs harboring poly(A) and 34 5´ UTRs without poly(A) from K. marxianus. The effects of each 5´ UTR on the production of a GFP reporter were evaluated individually in vivo, and the resulting protein abundance spanned an approximately 450-fold range throughout. The data were used to train a multi-layer perceptron neural network (MLP-NN) model that incorporated the length and position of poly(A) as features. The model exhibited good performance in predicting protein abundance (average R2 = 0.7290). The model suggests that the length of poly(A) is negatively correlated with protein production, whereas poly(A) located between 10 and 30 nt upstream of the start codon (AUG) exhibits a weak positive effect on protein abundance. Using the model as guidance, the deletion or reduction of poly(A) upstream of 30 nt preceding AUG tended to improve the production of GFP and a feruloyl esterase. Deletions of poly(A) showed inconsistent effects on mRNA levels, suggesting that poly(A) represses protein production either with or without reducing mRNA levels.

CONCLUSION

The effects of poly(A) on protein production depend on its length and position. Integrating poly(A) features into machine-learning models improves simulation accuracy. Deleting or reducing poly(A) upstream of 30 nt preceding AUG tends to enhance protein production. This optimization strategy can be applied to enhance the yield of K. marxianus and other microbial cell factories.

Hold out the genome: a roadmap to solving the cis-regulatory code

Gene expression is regulated by transcription factors that work together to read cis-regulatory DNA sequences. The 'cis-regulatory code' - how cells interpret DNA sequences to determine when, where and how much genes should be expressed - has proven to be exceedingly complex. Recently, advances in the scale and resolution of functional genomics assays and machine learning have enabled substantial progress towards deciphering this code. However, the cis-regulatory code will probably never be solved if models are trained only on genomic sequences; regions of homology can easily lead to overestimation of predictive performance, and our genome is too short and has insufficient sequence diversity to learn all relevant parameters. Fortunately, randomly synthesized DNA sequences enable testing a far larger sequence space than exists in our genomes, and designed DNA sequences enable targeted queries to maximally improve the models. As the same biochemical principles are used to interpret DNA regardless of its source, models trained on these synthetic data can predict genomic activity, often better than genome-trained models. Here we provide an outlook on the field, and propose a roadmap towards solving the cis-regulatory code by a combination of machine learning and massively parallel assays using synthetic DNA.

Tailor made: the art of therapeutic mRNA design

mRNA medicine is a new and rapidly developing field in which the delivery of genetic information in the form of mRNA is used to direct therapeutic protein production in humans. This approach, which allows for the quick and efficient identification and optimization of drug candidates for both large populations and individual patients, has the potential to revolutionize the way we prevent and treat disease. A key feature of mRNA medicines is their high degree of designability, although the design choices involved are complex. Maximizing the production of therapeutic proteins from mRNA medicines requires a thorough understanding of how nucleotide sequence, nucleotide modification and RNA structure interplay to affect translational efficiency and mRNA stability. In this Review, we describe the principles that underlie the physical stability and biological activity of mRNA and emphasize their relevance to the myriad considerations that factor into therapeutic mRNA design.

Artificial Intelligence in battling infectious diseases: A transformative role

It is widely acknowledged that infectious diseases have wrought immense havoc on human society, being regarded as adversaries from which humanity cannot elude. In recent years, the advancement of Artificial Intelligence (AI) technology has ushered in a revolutionary era in the realm of infectious disease prevention and control. This evolution encompasses early warning of outbreaks, contact tracing, infection diagnosis, drug discovery, and the facilitation of drug design, alongside other facets of epidemic management. This article presents an overview of the utilization of AI systems in the field of infectious diseases, with a specific focus on their role during the COVID-19 pandemic. The article also highlights the contemporary challenges that AI confronts within this domain and posits strategies for their mitigation. There exists an imperative to further harness the potential applications of AI across multiple domains to augment its capacity in effectively addressing future disease outbreaks.

RNA-Based Vaccines and Therapeutics Against Intracellular Pathogens

RNA-based vaccines have sparked a paradigm shift in the treatment and prevention of diseases by nucleic acid medicines. There has been a notable surge in the development of nucleic acid therapeutics and vaccines following the global approval of the two messenger RNA-based COVID-19 vaccines. This growth is fueled by the exploration of numerous RNA products in preclinical stages, offering several advantages over conventional methods, i.e., safety, efficacy, scalability, and cost-effectiveness. In this chapter, we provide an overview of various types of RNA and their mechanisms of action for stimulating immune responses and inducing therapeutic effects. Furthermore, this chapter delves into the varying delivery systems, particularly emphasizing the use of nanoparticles to deliver RNA. The choice of delivery system is an intricate process involved in developing nucleic acid medicines that significantly enhances their stability, biocompatibility, and site-specificity. Additionally, this chapter sheds light on the current landscape of clinical trials of RNA therapeutics and vaccines against intracellular pathogens.

Monovalent SARS-COV-2 mRNA vaccine using optimal UTRs and LNPs is highly immunogenic and broadly protective against Omicron variants

The emergence of highly transmissible severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) variants of concern (VOCs) that are resistant to the current COVID-19 vaccines highlights the need for continued development of broadly protective vaccines for the future. Here, we developed two messenger RNA (mRNA)-lipid nanoparticle (LNP) vaccines, TU88mCSA and ALCmCSA, using the ancestral SARS-CoV-2 spike sequence, optimized 5' and 3' untranslated regions (UTRs), and LNP combinations. Our data showed that these nanocomplexes effectively activate CD4+ and CD8+ T cell responses and humoral immune response and provide complete protection against WA1/2020, Omicron BA.1 and BQ.1 infection in hamsters. Critically, in Omicron BQ.1 challenge hamster models, TU88mCSA and ALCmCSA not only induced robust control of virus load in the lungs but also enhanced protective efficacy in the upper respiratory airways. Antigen-specific immune analysis in mice revealed that the observed cross-protection is associated with superior UTRs [Carboxylesterase 1d (Ces1d)/adaptor protein-3β (AP3B1)] and LNP formulations that elicit robust lung tissue-resident memory T cells. Strong protective effects of TU88mCSA or ALCmCSA against both WA1/2020 and VOCs suggest that this mRNA-LNP combination can be a broadly protective vaccine platform in which mRNA cargo uses the ancestral antigen sequence regardless of the antigenic drift. This approach could be rapidly adapted for clinical use and timely deployment of vaccines against emerging and reemerging VOCs.

Discovery of regulatory motifs in 5' untranslated regions using interpretable multi-task learning models

The sequence in the 5' untranslated regions (UTRs) is known to affect mRNA translation rates. However, the underlying regulatory grammar remains elusive. Here, we propose MTtrans, a multi-task translation rate predictor capable of learning common sequence patterns from datasets across various experimental techniques. The core premise is that common motifs are more likely to be genuinely involved in translation control. MTtrans outperforms existing methods in both accuracy and the ability to capture transferable motifs across species, highlighting its strength in identifying evolutionarily conserved sequence motifs. Our independent fluorescence-activated cell sorting coupled with deep sequencing (FACS-seq) experiment validates the impact of most motifs identified by MTtrans. Additionally, we introduce "GRU-rewiring," a technique to interpret the hidden states of the recurrent units. Gated recurrent unit (GRU)-rewiring allows us to identify regulatory element-enriched positions and examine the local effects of 5' UTR mutations. MTtrans is a powerful tool for deciphering the translation regulatory motifs.

The anticancer compound JTE-607 reveals hidden sequence specificity of the mRNA 3' processing machinery

JTE-607 is an anticancer and anti-inflammatory compound and its active form, compound 2, directly binds to and inhibits CPSF73, the endonuclease for the cleavage step in pre-messenger RNA (pre-mRNA) 3' processing. Surprisingly, compound 2-mediated inhibition of pre-mRNA cleavage is sequence specific and the drug sensitivity is predominantly determined by sequences flanking the cleavage site (CS). Using massively parallel in vitro assays, we identified key sequence features that determine drug sensitivity. We trained a machine learning model that can predict poly(A) site (PAS) relative sensitivity to compound 2 and provide the molecular basis for understanding the impact of JTE-607 on PAS selection and transcription termination genome wide. We propose that CPSF73 and associated factors bind to the CS region in a sequence-dependent manner and the interaction affinity determines compound 2 sensitivity. These results have not only elucidated the mechanism of action of JTE-607, but also unveiled an evolutionarily conserved sequence specificity of the mRNA 3' processing machinery.

The regulatory landscape of 5' UTRs in translational control during zebrafish embryogenesis

The 5' UTRs of mRNAs are critical for translation regulation, but their in vivo regulatory features are poorly characterized. Here, we report the regulatory landscape of 5' UTRs during early zebrafish embryogenesis using a massively parallel reporter assay of 18,154 sequences coupled to polysome profiling. We found that the 5' UTR is sufficient to confer temporal dynamics to translation initiation, and identified 86 motifs enriched in 5' UTRs with distinct ribosome recruitment capabilities. A quantitative deep learning model, DaniO5P, revealed a combined role for 5' UTR length, translation initiation site context, upstream AUGs and sequence motifs on in vivo ribosome recruitment. DaniO5P predicts the activities of 5' UTR isoforms and indicates that modulating 5' UTR length and motif grammar contributes to translation initiation dynamics. This study provides a first quantitative model of 5' UTR-based translation regulation in early vertebrate development and lays the foundation for identifying the underlying molecular effectors.

Highlights

In vivo MPRA systematically interrogates the regulatory potential of endogenous 5' UTRs The 5' UTR alone is sufficient to regulate the dynamics of ribosome recruitment during early embryogenesis The MPRA identifies 5' UTR cis -regulatory motifs for translation initiation control 5' UTR length, upstream AUGs and motif grammar contribute to the differential regulatory capability of 5' UTR switching isoforms.

Effects of Combinations of Untranslated-Region Sequences on Translation of mRNA

mRNA-based therapeutics have been found to be a promising treatment strategy in immunotherapy, gene therapy, and cancer treatments. Effectiveness of mRNA therapeutics depends on the level and duration of a desired protein's expression, which is determined by various cis- and trans-regulatory elements of the mRNA. Sequences of 5' and 3' untranslated regions (UTRs) are responsible for translational efficiency and stability of mRNA. An optimal combination of the regulatory sequences allows researchers to significantly increase the target protein's expression. Using both literature data and previously obtained experimental data, we chose six sequences of 5'UTRs (adenoviral tripartite leader [TPL], HBB, rabbit β-globin [Rabb], H4C2, Moderna, and Neo2) and five sequences of 3'UTRs (mtRNR-EMCV, mtRNR-AES, mtRNR-mtRNR, BioNTech, and Moderna). By combining them, we constructed 30 in vitro transcribed RNAs encoding firefly luciferase with various combinations of 5'- and 3'UTRs, and the resultant bioluminescence was assessed in the DC2.4 cell line at 4, 8, 24, and 72 h after transfection. The cellular data enabled us to identify the best seven combinations of 5'- and 3'UTRs, whose translational efficiency was then assessed in BALB/c mice. Two combinations of 5'- and 3'UTRs (5'Rabb-3'mtRNR-EMCV and 5'TPL-3'Biontech) led to the most pronounced increase in the luciferase amount in the in vivo experiment in mice. Subsequent analysis of the stability of the mRNA indicated that the increase in luciferase expression is explained primarily by the efficiency of translation, not by the number of RNA molecules. Altogether, these findings suggest that 5'UTR-and-3'UTR combinations 5'Rabb-3'mtRNR- EMCV and 5'TPL-3'Biontech lead to high expression of target proteins and may be considered for use in preventive and therapeutic modalities based on mRNA.

Integrated multiplexed assays of variant effect reveal cis-regulatory determinants of catechol-O-methyltransferase gene expression

Multiplexed assays of variant effect are powerful methods to profile the consequences of rare variants on gene expression and organismal fitness. Yet, few studies have integrated several multiplexed assays to map variant effects on gene expression in coding sequences. Here, we pioneered a multiplexed assay based on polysome profiling to measure variant effects on translation at scale, uncovering single-nucleotide variants that increase and decrease ribosome load. By combining high-throughput ribosome load data with multiplexed mRNA and protein abundance readouts, we mapped the cis-regulatory landscape of thousands of catechol-O-methyltransferase (COMT) variants from RNA to protein and found numerous coding variants that alter COMT expression. Finally, we trained machine learning models to map signatures of variant effects on COMT gene expression and uncovered both directional and divergent impacts across expression layers. Our analyses reveal expression phenotypes for thousands of variants in COMT and highlight variant effects on both single and multiple layers of expression. Our findings prompt future studies that integrate several multiplexed assays for the readout of gene expression.

A Massively Parallel Screen of 5'UTR Mutations Identifies Variants Impacting Translation and Protein Production in Neurodevelopmental Disorder Genes

De novo mutations cause a variety of neurodevelopmental disorders including autism. Recent whole genome sequencing from individuals with autism has shown that many de novo mutations also occur in untranslated regions (UTRs) of genes, but it is difficult to predict from sequence alone which mutations are functional, let alone causal. Therefore, we developed a high throughput assay to screen the transcriptional and translational effects of 997 variants from 5'UTR patient mutations. This assay successfully enriched for elements that alter reporter translation, identifying over 100 potentially functional mutations from probands. Studies in patient-derived cell lines further confirmed that these mutations can alter protein production in individuals with autism, and some variants fall in genes known to cause syndromic forms of autism, suggesting a diagnosis for these individual patients. Since UTR function varies by cell type, we further optimized this high throughput assay to enable assessment of mutations in neurons in vivo. First, comparing in cellulo to in vivo results, we demonstrate neurons have different principles of regulation by 5'UTRs, consistent with a more robust mechanism for reducing the impact of RNA secondary structure. Finally, we discovered patient mutations specifically altering the translational activity of additional known syndromic genes LRRC4 and ZNF644 in neurons of the brain. Overall our results highlight a new approach for assessing the impact of 5'UTR mutations across cell types and suggest that some cases of neurodevelopmental disorder may be caused by such variants.

Biochemical characterization of mRNA capping enzyme from Faustovirus

The mammalian mRNA 5' cap structures play important roles in cellular processes such as nuclear export, efficient translation, and evading cellular innate immune surveillance and regulating 5'-mediated mRNA turnover. Hence, installation of the proper 5' cap is crucial in therapeutic applications of synthetic mRNA. The core 5' cap structure, Cap-0, is generated by three sequential enzymatic activities: RNA 5' triphosphatase, RNA guanylyltransferase, and cap N7-guanine methyltransferase. Vaccinia virus RNA capping enzyme (VCE) is a heterodimeric enzyme that has been widely used in synthetic mRNA research and manufacturing. The large subunit of VCE D1R exhibits a modular structure where each of the three structural domains possesses one of the three enzyme activities, whereas the small subunit D12L is required to activate the N7-guanine methyltransferase activity. Here, we report the characterization of a single-subunit RNA capping enzyme from an amoeba giant virus. Faustovirus RNA capping enzyme (FCE) exhibits a modular array of catalytic domains in common with VCE and is highly efficient in generating the Cap-0 structure without an activation subunit. Phylogenetic analysis suggests that FCE and VCE are descended from a common ancestral capping enzyme. We found that compared to VCE, FCE exhibits higher specific activity, higher activity toward RNA containing secondary structures and a free 5' end, and a broader temperature range, properties favorable for synthetic mRNA manufacturing workflows.

In Vitro Transcribed RNA-Based Platform Vaccines: Past, Present, and Future

mRNA was discovered in 1961, but it was not used as a vaccine until after three decades. Recently, the development of mRNA vaccine technology gained great impetus from the pursuit of vaccines against COVID-19. To improve the properties of RNA vaccines, and primarily their circulation time, self-amplifying mRNA and trans-amplifying mRNA were developed. A separate branch of mRNA technology is circular RNA vaccines, which were developed with the discovery of the possibility of translation on their protein matrix. Circular RNA has several advantages over mRNA vaccines and is considered a fairly promising platform, as is trans-amplifying mRNA. This review presents an overview of the mRNA platform and a critical discussion of the more modern self-amplifying mRNA, trans-amplifying mRNA, and circular RNA platforms created on its basis. Finally, the main features, advantages, and disadvantages of each of the presented mRNA platforms are discussed. This discussion will facilitate the decision-making process in selecting the most appropriate platform for creating RNA vaccines against cancer or viral diseases.

A fast non-parametric test of association for multiple traits

The increasing availability of multidimensional phenotypic data in large cohorts of genotyped individuals requires efficient methods to identify genetic effects on multiple traits. Permutational multivariate analysis of variance (PERMANOVA) offers a powerful non-parametric approach. However, it relies on permutations to assess significance, which hinders the analysis of large datasets. Here, we derive the limiting null distribution of the PERMANOVA test statistic, providing a framework for the fast computation of asymptotic p values. Our asymptotic test presents controlled type I error and high power, often outperforming parametric approaches. We illustrate its applicability in the context of QTL mapping and GWAS.

Human-specific genetics: new tools to explore the molecular and cellular basis of human evolution

Our ancestors acquired morphological, cognitive and metabolic modifications that enabled humans to colonize diverse habitats, develop extraordinary technologies and reshape the biosphere. Understanding the genetic, developmental and molecular bases for these changes will provide insights into how we became human. Connecting human-specific genetic changes to species differences has been challenging owing to an abundance of low-effect size genetic changes, limited descriptions of phenotypic differences across development at the level of cell types and lack of experimental models. Emerging approaches for single-cell sequencing, genetic manipulation and stem cell culture now support descriptive and functional studies in defined cell types with a human or ape genetic background. In this Review, we describe how the sequencing of genomes from modern and archaic hominins, great apes and other primates is revealing human-specific genetic changes and how new molecular and cellular approaches - including cell atlases and organoids - are enabling exploration of the candidate causal factors that underlie human-specific traits.

Using Synthetic DNA Libraries to Investigate Chromatin and Gene Regulation

Despite the recent explosion in genome-wide studies in chromatin and gene regulation, we are still far from extracting a set of genetic rules that can predict the function of the regulatory genome. One major reason for this deficiency is that gene regulation is a multi-layered process that involves an enormous variable space, which cannot be fully explored using native genomes. This problem can be partially solved by introducing synthetic DNA libraries into cells, a method that can test the regulatory roles of thousands to millions of sequences with limited variables. Here, we review recent applications of this method to study transcription factor (TF) binding, nucleosome positioning, and transcriptional activity. We discuss the design principles, experimental procedures, and major findings from these studies and compare the pros and cons of different approaches.

mRNA therapeutics: New vaccination and beyond

The idea of mRNA therapy had been conceived for decades before it came into reality during the Covid-19 pandemic. The mRNA vaccine emerges as a powerful and general tool against new viral infections, largely due to its versatility and rapid development. In addition to prophylactic vaccines, mRNA technology also offers great promise for new applications as a versatile drug modality. However, realizing the conceptual potential faces considerable challenges, such as minimal immune stimulation, high and long-term expression, and efficient delivery to target cells and tissues. Here we review the applications of mRNA-based therapeutics, with emphasis on the innovative design and future challenges/solutions. In addition, we also discuss the next generation of mRNA therapy, including circular mRNA and self-amplifying RNAs. We aim to provide a conceptual overview and outlook on mRNA therapeutics beyond prophylactic vaccines.

Recent Advances and Innovations in the Preparation and Purification of In Vitro-Transcribed-mRNA-Based Molecules

The coronavirus disease 2019 (COVID-19) pandemic poses a disruptive impact on public health and the global economy. Fortunately, the development of COVID-19 vaccines based on in vitro-transcribed messenger RNA (IVT mRNA) has been a breakthrough in medical history, benefiting billions of people with its high effectiveness, safety profile, and ease of large-scale production. This success is the result of decades of continuous RNA research, which has led to significant improvements in the stability and expression level of IVT mRNA through various approaches such as sequence optimization and improved preparation processes. IVT mRNA sequence optimization has been shown to have a positive effect on enhancing the mRNA expression level. The innovation of IVT mRNA purification technology is also indispensable, as the purity of IVT mRNA directly affects the success of downstream vaccine preparation processes and the potential for inducing unwanted side effects in therapeutic applications. Despite the progress made, challenges related to IVT mRNA sequence design and purification still require further attention to enhance the quality of IVT mRNA in the future. In this review, we discuss the latest innovative progress in IVT mRNA design and purification to further improve its clinical efficacy.

Machine-guided design of synthetic cell type-specific cis-regulatory elements

Cis-regulatory elements (CREs) control gene expression, orchestrating tissue identity, developmental timing, and stimulus responses, which collectively define the thousands of unique cell types in the body. While there is great potential for strategically incorporating CREs in therapeutic or biotechnology applications that require tissue specificity, there is no guarantee that an optimal CRE for an intended purpose has arisen naturally through evolution. Here, we present a platform to engineer and validate synthetic CREs capable of driving gene expression with programmed cell type specificity. We leverage innovations in deep neural network modeling of CRE activity across three cell types, efficient in silico optimization, and massively parallel reporter assays (MPRAs) to design and empirically test thousands of CREs. Through in vitro and in vivo validation, we show that synthetic sequences outperform natural sequences from the human genome in driving cell type-specific expression. Synthetic sequences leverage unique sequence syntax to promote activity in the on-target cell type and simultaneously reduce activity in off-target cells. Together, we provide a generalizable framework to prospectively engineer CREs and demonstrate the required literacy to write regulatory code that is fit-for-purpose in vivo across vertebrates.

Advancements of in vitro transcribed mRNA (IVT mRNA) to enable translation into the clinics

The accelerated progress and approval of two mRNA-based vaccines to address the SARS-CoV-2 virus were unprecedented. This record-setting feat was made possible through the solid foundation of research on in vitro transcribed mRNA (IVT mRNA) which could be utilized as a therapeutic modality. Through decades of thorough research to overcome barriers to implementation, mRNA-based vaccines or therapeutics offer many advantages to rapidly address a broad range of applications including infectious diseases, cancers, and gene editing. Here, we describe the advances that have supported the adoption of IVT mRNA in the clinics, including optimization of the IVT mRNA structural components, synthesis, and lastly concluding with different classes of IVT RNA. Continuing interest in driving IVT mRNA technology will enable a safer and more efficacious therapeutic modality to address emerging and existing diseases.

Polyplex designs for improving the stability and safety of RNA therapeutics

Nanoparticle-based delivery systems have contributed to the recent clinical success of RNA therapeutics, including siRNA and mRNA. RNA delivery using polymers has several distinct properties, such as enabling RNA delivery into extra-hepatic organs, modulation of immune responses to RNA, and regulation of intracellular RNA release. However, delivery systems should overcome safety and stability issues to achieve widespread therapeutic applications. Safety concerns include direct damage to cellular components, innate and adaptive immune responses, complement activation, and interaction with surrounding molecules and cells in the blood circulation. The stability of the delivery systems should balance extracellular RNA protection and controlled intracellular RNA release, which requires optimization for each RNA species. Further, polymer designs for improving safety and stability often conflict with each other. This review covers advances in polymer-based approaches to address these issues over several years, focusing on biological understanding and design concepts for delivery systems rather than material chemistry.

Recent trends in RNA informatics: a review of machine learning and deep learning for RNA secondary structure prediction and RNA drug discovery

Computational analysis of RNA sequences constitutes a crucial step in the field of RNA biology. As in other domains of the life sciences, the incorporation of artificial intelligence and machine learning techniques into RNA sequence analysis has gained significant traction in recent years. Historically, thermodynamics-based methods were widely employed for the prediction of RNA secondary structures; however, machine learning-based approaches have demonstrated remarkable advancements in recent years, enabling more accurate predictions. Consequently, the precision of sequence analysis pertaining to RNA secondary structures, such as RNA-protein interactions, has also been enhanced, making a substantial contribution to the field of RNA biology. Additionally, artificial intelligence and machine learning are also introducing technical innovations in the analysis of RNA-small molecule interactions for RNA-targeted drug discovery and in the design of RNA aptamers, where RNA serves as its own ligand. This review will highlight recent trends in the prediction of RNA secondary structure, RNA aptamers and RNA drug discovery using machine learning, deep learning and related technologies, and will also discuss potential future avenues in the field of RNA informatics.

Leveraging artificial intelligence in the fight against infectious diseases

Despite advances in molecular biology, genetics, computation, and medicinal chemistry, infectious disease remains an ominous threat to public health. Addressing the challenges posed by pathogen outbreaks, pandemics, and antimicrobial resistance will require concerted interdisciplinary efforts. In conjunction with systems and synthetic biology, artificial intelligence (AI) is now leading to rapid progress, expanding anti-infective drug discovery, enhancing our understanding of infection biology, and accelerating the development of diagnostics. In this Review, we discuss approaches for detecting, treating, and understanding infectious diseases, underscoring the progress supported by AI in each case. We suggest future applications of AI and how it might be harnessed to help control infectious disease outbreaks and pandemics.

Generative and predictive neural networks for the design of functional RNA molecules

RNA is a remarkably versatile molecule that has been engineered for applications in therapeutics, diagnostics, and in vivo information-processing systems. However, the complex relationship between the sequence and structural properties of an RNA molecule and its ability to perform specific functions often necessitates extensive experimental screening of candidate sequences. Here we present a generalized neural network architecture that utilizes the sequence and structure of RNA molecules (SANDSTORM) to inform functional predictions. We demonstrate that this approach achieves state-of-the-art performance across several distinct RNA prediction tasks, while learning interpretable abstractions of RNA secondary structure. We paired these predictive models with generative adversarial RNA design networks (GARDN), allowing the generative modelling of novel mRNA 5' untranslated regions and toehold switch riboregulators exhibiting a predetermined fitness. This approach enabled the design of novel toehold switches with a 43-fold increase in experimentally characterized dynamic range compared to those designed using classic thermodynamic algorithms. SANDSTORM and GARDN thus represent powerful new predictive and generative tools for the development of diagnostic and therapeutic RNA molecules with improved function.

An RNA-immunoprecipitation via CRISPR/dCas13 reveals an interaction between the SARS-CoV-2 5'UTR RNA and the process of human lipid metabolism

We herein elucidate the function of SARS-CoV-2derived 5'UTR in the human cells. 5'UTR bound host cellular RNAs were immunoprecipitated by gRNA-dCas13 (targeting luciferase RNA fused to SARS-CoV-2 5'UTR) in HEK293T and A549 cells. The 5'UTR bound RNA extractions were predominantly enriched for regulating lipid metabolism. Overexpression of SARS-CoV-2 5'UTR RNA altered the expression of factors involved in the process of the human Mevalonate pathway. In addition, we found that HMG-CoA reductase inhibitors were shown to suppress SARS-CoV-2 5'UTR-mediated translation activities. In conclusion, we deduce the array of host RNAs interacting with SARS-CoV-2 5'UTR that drives SARS-CoV-2 translation and influences host metabolic pathways.

Functional characterization of human genomic variation linked to polygenic diseases

The burden of human disease lies predominantly in polygenic diseases. Since the early 2000s, genome-wide association studies (GWAS) have identified genetic variants and loci associated with complex traits. These have ranged from variants in coding sequences to mutations in regulatory regions, such as promoters and enhancers, as well as mutations affecting mediators of mRNA stability and other downstream regulators, such as 5' and 3'-untranslated regions (UTRs), long noncoding RNA (lncRNA), and miRNA. Recent research advances in genetics have utilized a combination of computational techniques, high-throughput in vitro and in vivo screening modalities, and precise genome editing to impute the function of diverse classes of genetic variants identified through GWAS. In this review, we highlight the vastness of genomic variants associated with polygenic disease risk and address recent advances in how genetic tools can be used to functionally characterize them.

Translation initiation consistency between in vivo and in vitro bacterial protein expression systems

Strict on-demand control of protein synthesis is a crucial aspect of synthetic biology. The 5'-terminal untranslated region (5'-UTR) is an essential bacterial genetic element that can be designed for the regulation of translation initiation. However, there is insufficient systematical data on the consistency of 5'-UTR function among various bacterial cells and in vitro protein synthesis systems, which is crucial for the standardization and modularization of genetic elements in synthetic biology. Here, more than 400 expression cassettes comprising the GFP gene under the regulation of various 5'-UTRs were systematically characterized to evaluate the protein translation consistency in the two popular Escherichia coli strains of JM109 and BL21, as well as an in vitro protein expression system based on cell lysate. In contrast to the very strong correlation between the two cellular systems, the consistency between in vivo and in vitro protein translation was lost, whereby both in vivo and in vitro translation evidently deviated from the estimation of the standard statistical thermodynamic model. Finally, we found that the absence of nucleotide C and complex secondary structure in the 5'-UTR significantly improve the efficiency of protein translation, both in vitro and in vivo.

Unraveling the influences of sequence and position on yeast uORF activity using massively parallel reporter systems and machine learning

Upstream open-reading frames (uORFs) are potent cis-acting regulators of mRNA translation and nonsense-mediated decay (NMD). While both AUG- and non-AUG initiated uORFs are ubiquitous in ribosome profiling studies, few uORFs have been experimentally tested. Consequently, the relative influences of sequence, structural, and positional features on uORF activity have not been determined. We quantified thousands of yeast uORFs using massively parallel reporter assays in wildtype and ∆upf1 yeast. While nearly all AUG uORFs were robust repressors, most non-AUG uORFs had relatively weak impacts on expression. Machine learning regression modeling revealed that both uORF sequences and locations within transcript leaders predict their effect on gene expression. Indeed, alternative transcription start sites highly influenced uORF activity. These results define the scope of natural uORF activity, identify features associated with translational repression and NMD, and suggest that the locations of uORFs in transcript leaders are nearly as predictive as uORF sequences.

Deep learning for optimization of protein expression

Recent progress in high-throughput DNA synthesis and sequencing has enabled the development of massively parallel reporter assays for strain characterization. These datasets map a large number of DNA sequences to protein expression levels, sparking increased interest in data-driven methods for sequence-to-expression modeling. Here, we highlight advances in deep learning models of protein expression and their potential for optimizing strains engineered to produce recombinant proteins. We review recent works that built highly accurate models and discuss challenges that hinder adoption by end users. There is a need to better align this technology with the constraints encountered in strain engineering, particularly the cost of acquiring large amounts of data and the requirement for interpretable models that generalize beyond the training data. Overcoming these barriers will help to incentivize academic and industrial laboratories to tap into a new era of data-centric strain engineering.

satmut_utils: a simulation and variant calling package for multiplexed assays of variant effect

The impact of millions of individual genetic variants on molecular phenotypes in coding sequences remains unknown. Multiplexed assays of variant effect (MAVEs) are scalable methods to annotate relevant variants, but existing software lacks standardization, requires cumbersome configuration, and does not scale to large targets. We present satmut_utils as a flexible solution for simulation and variant quantification. We then benchmark MAVE software using simulated and real MAVE data. We finally determine mRNA abundance for thousands of cystathionine beta-synthase variants using two experimental methods. The satmut_utils package enables high-performance analysis of MAVEs and reveals the capability of variants to alter mRNA abundance.

The anti-cancer compound JTE-607 reveals hidden sequence specificity of the mRNA 3' processing machinery

JTE-607 is a small molecule compound with anti-inflammation and anti-cancer activities. Upon entering the cell, it is hydrolyzed to Compound 2, which directly binds to and inhibits CPSF73, the endonuclease for the cleavage step in pre-mRNA 3' processing. Although CPSF73 is universally required for mRNA 3' end formation, we have unexpectedly found that Compound 2- mediated inhibition of pre-mRNA 3' processing is sequence-specific and that the sequences flanking the cleavage site (CS) are a major determinant for drug sensitivity. By using massively parallel in vitro assays, we have measured the Compound 2 sensitivities of over 260,000 sequence variants and identified key sequence features that determine drug sensitivity. A machine learning model trained on these data can predict the impact of JTE-607 on poly(A) site (PAS) selection and transcription termination genome-wide. We propose a biochemical model in which CPSF73 and other mRNA 3' processing factors bind to RNA of the CS region in a sequence-specific manner and the affinity of such interaction determines the Compound 2 sensitivity of a PAS. As the Compound 2-resistant CS sequences, characterized by U/A-rich motifs, are prevalent in PASs from yeast to human, the CS region sequence may have more fundamental functions beyond determining drug resistance. Together, our study not only characterized the mechanism of action of a compound with clinical implications, but also revealed a previously unknown and evolutionarily conserved sequence-specificity of the mRNA 3' processing machinery.

The non-coding genome in Autism Spectrum Disorders

Autism Spectrum Disorders (ASD) are a group of neurodevelopmental disorders (NDDs) characterized by difficulties in social interaction and communication, repetitive behavior, and restricted interests. While ASD have been proven to have a strong genetic component, current research largely focuses on coding regions of the genome. However, non-coding DNA, which makes up for ∼99% of the human genome, has recently been recognized as an important contributor to the high heritability of ASD, and novel sequencing technologies have been a milestone in opening up new directions for the study of the gene regulatory networks embedded within the non-coding regions. Here, we summarize current progress on the contribution of non-coding alterations to the pathogenesis of ASD and provide an overview of existing methods allowing for the study of their functional relevance, discussing potential ways of unraveling ASD's "missing heritability".

Optimization of 5'UTR to evade SARS-CoV-2 Nonstructural protein 1-directed inhibition of protein synthesis in cells

Maximizing the expression level of therapeutic proteins in cells is the general goal for DNA/mRNA therapies. It is particularly challenging to achieve efficient protein expression in the cellular contexts with inhibited translation machineries, such as in the presence of cellular Nonstructural protein 1 (Nsp1) of coronaviruses (CoVs) that has been reported to inhibit overall protein synthesis of host genes and exogenously delivered mRNAs/DNAs. In this study, we thoroughly examined the sequence and structure contexts of viral and non-viral 5'UTRs that determine the protein expression levels of exogenously delivered DNAs and mRNAs in cells expressing SARS-CoV-2 Nsp1. It was found that high 5'-proximal A/U content promotes an escape from Nsp1-directed inhibition of protein synthesis and results in selective protein expression. Furthermore, 5'-proximal Cs were found to significantly enhance the protein expression in an Nsp1-dependent manner, while Gs located at a specific window close to the 5'-end counteract such enhancement. The distinct protein expression levels resulted from different 5'UTRs were found correlated to Nsp1-induced mRNA degradations. These findings ultimately enabled rational designs for optimized 5'UTRs that lead to strong expression of exogenous proteins regardless of the translationally repressive Nsp1. On the other hand, we have also identified several 5'-proximal sequences derived from host genes that are capable of mediating the escapes. These results provided novel perspectives to the optimizations of 5'UTRs for DNA/mRNA therapies and/or vaccinations, as well as shedding light on the potential host escapees from Nsp1-directed translational shutoffs. KEY POINTS: • The 5'-proximal SL1 and 5a/b derived from SARS-CoV-2 genomic RNA promote exogenous protein synthesis in cells expressing Nsp1 comparing with non-specific 5'UTRs. • Specific 5'-proximal sequence contexts are the key determinants of the escapes from Nsp1-directed translational repression and thereby enhance protein expressions. • Systematic mutagenesis identified optimized 5'UTRs that strongly enhance protein expression and promote resistance to Nsp1-induced translational repression and RNA degradation.

Genetic architecture of human brain evolution

Comparative studies of hominids have long sought to identify mutational events that shaped the evolution of the human nervous system. However, functional genetic differences are outnumbered by millions of nearly neutral mutations, and the developmental mechanisms underlying human nervous system specializations are difficult to model and incompletely understood. Candidate-gene studies have attempted to map select human-specific genetic differences to neurodevelopmental functions, but it remains unclear how to contextualize the relative effects of genes that are investigated independently. Considering these limitations, we discuss scalable approaches for probing the functional contributions of human-specific genetic differences. We propose that a systems-level view will enable a more quantitative and integrative understanding of the genetic, molecular and cellular underpinnings of human nervous system evolution.

Modified mRNA as a Treatment for Myocardial Infarction

Myocardial infarction (MI) is a severe disease with high mortality worldwide. However, regenerative approaches remain limited and with poor efficacy. The major difficulty during MI is the substantial loss of cardiomyocytes (CMs) with limited capacity to regenerate. As a result, for decades, researchers have been engaged in developing useful therapies for myocardial regeneration. Gene therapy is an emerging approach for promoting myocardial regeneration. Modified mRNA (modRNA) is a highly potential delivery vector for gene transfer with its properties of efficiency, non-immunogenicity, transiency, and relative safety. Here, we discuss the optimization of modRNA-based therapy, including gene modification and delivery vectors of modRNA. Moreover, the effective of modRNA in animal MI treatment is also discussed. We conclude that modRNA-based therapy with appropriate therapeutical genes can potentially treat MI by directly promoting proliferation and differentiation, inhibiting apoptosis of CMs, as well as enhancing paracrine effects in terms of promoting angiogenesis and inhibiting fibrosis in heart milieu. Finally, we summarize the current challenges of modRNA-based cardiac treatment and look forward to the future direction of such treatment for MI. Further advanced clinical trials incorporating more MI patients should be conducted in order for modRNA therapy to become practical and feasible in real-world treatment.

mRNA-Based Vaccine for COVID-19: They Are New but Not Unknown!

mRNA vaccines take advantage of the mechanism that our cells use to produce proteins. Our cells produce proteins based on the knowledge contained in our DNA; each gene encodes a unique protein. The genetic information is essential, but cells cannot use it until mRNA molecules convert it into instructions for producing specific proteins. mRNA vaccinations provide ready-to-use mRNA instructions for constructing a specific protein. BNT162b2 (Pfizer-BioNTech) and mRNA-1273 (Moderna) both are newly approved mRNA-based COVID-19 vaccines that have shown excellent protection and efficacy. In total, there are five more mRNA-based vaccine candidates for COVID-19 under different phases of clinical development. This review is specifically focused on mRNA-based vaccines for COVID-19 covering its development, mechanism, and clinical aspects.

mRNA-Based Therapeutics in Cancer Treatment

Over the past two decades, significant technological innovations have led to messenger RNA (mRNA) becoming a promising option for developing prophylactic and therapeutic vaccines, protein replacement therapies, and genome engineering. The success of the two COVID-19 mRNA vaccines has sparked new enthusiasm for other medical applications, particularly in cancer treatment. In vitro-transcribed (IVT) mRNAs are structurally designed to resemble naturally occurring mature mRNA. Delivery of IVT mRNA via delivery platforms such as lipid nanoparticles allows host cells to produce many copies of encoded proteins, which can serve as antigens to stimulate immune responses or as additional beneficial proteins for supplements. mRNA-based cancer therapeutics include mRNA cancer vaccines, mRNA encoding cytokines, chimeric antigen receptors, tumor suppressors, and other combination therapies. To better understand the current development and research status of mRNA therapies for cancer treatment, this review focused on the molecular design, delivery systems, and clinical indications of mRNA therapies in cancer.

Translation of in vitro-transcribed RNA therapeutics

In vitro transcribed, modified messenger RNAs (IVTmRNAs) have been used to vaccinate billions of individuals against the SARS-CoV-2 virus, and are currently being developed for many additional therapeutic applications. IVTmRNAs must be translated into proteins with therapeutic activity by the same cellular machinery that also translates native endogenous transcripts. However, different genesis pathways and routes of entry into target cells as well as the presence of modified nucleotides mean that the way in which IVTmRNAs engage with the translational machinery, and the efficiency with which they are being translated, differs from native mRNAs. This review summarises our current knowledge of commonalities and differences in translation between IVTmRNAs and cellular mRNAs, which is key for the development of future design strategies that can generate IVTmRNAs with improved activity in therapeutic applications.

mRNA ageing shapes the Cap2 methylome in mammalian mRNA

The mRNA cap structure is a major site of dynamic mRNA methylation. mRNA caps exist in either the Cap1 or Cap2 form, depending on the presence of 2'-O-methylation on the first transcribed nucleotide or both the first and second transcribed nucleotides, respectively1,2. However, the identity of Cap2-containing mRNAs and the function of Cap2 are unclear. Here we describe CLAM-Cap-seq, a method for transcriptome-wide mapping and quantification of Cap2. We find that unlike other epitranscriptomic modifications, Cap2 can occur on all mRNAs. Cap2 is formed through a slow continuous conversion of mRNAs from Cap1 to Cap2 as mRNAs age in the cytosol. As a result, Cap2 is enriched on long-lived mRNAs. Large increases in the abundance of Cap1 leads to activation of RIG-I, especially in conditions in which expression of RIG-I is increased. The methylation of Cap1 to Cap2 markedly reduces the ability of RNAs to bind to and activate RIG-I. The slow methylation rate of Cap2 allows Cap2 to accumulate on host mRNAs, yet ensures that low levels of Cap2 occur on newly expressed viral RNAs. Overall, these results reveal an immunostimulatory role for Cap1, and that Cap2 functions to reduce activation of the innate immune response.

Ionizable lipid nanoparticles deliver mRNA to pancreatic β cells via macrophage-mediated gene transfer

Systemic messenger RNA (mRNA) delivery to organs outside the liver, spleen, and lungs remains challenging. To overcome this issue, we hypothesized that altering nanoparticle chemistry and administration routes may enable mRNA-induced protein expression outside of the reticuloendothelial system. Here, we describe a strategy for delivering mRNA potently and specifically to the pancreas using lipid nanoparticles. Our results show that delivering lipid nanoparticles containing cationic helper lipids by intraperitoneal administration produces robust and specific protein expression in the pancreas. Most resultant protein expression occurred within insulin-producing β cells. Last, we found that pancreatic mRNA delivery was dependent on horizontal gene transfer by peritoneal macrophage exosome secretion, an underappreciated mechanism that influences the delivery of mRNA lipid nanoparticles. We anticipate that this strategy will enable gene therapies for intractable pancreatic diseases such as diabetes and cancer.

Integrated mRNA sequence optimization using deep learning

The coronavirus disease of 2019 pandemic has catalyzed the rapid development of mRNA vaccines, whereas, how to optimize the mRNA sequence of exogenous gene such as severe acute respiratory syndrome coronavirus 2 spike to fit human cells remains a critical challenge. A new algorithm, iDRO (integrated deep-learning-based mRNA optimization), is developed to optimize multiple components of mRNA sequences based on given amino acid sequences of target protein. Considering the biological constraints, we divided iDRO into two steps: open reading frame (ORF) optimization and 5' untranslated region (UTR) and 3'UTR generation. In ORF optimization, BiLSTM-CRF (bidirectional long-short-term memory with conditional random field) is employed to determine the codon for each amino acid. In UTR generation, RNA-Bart (bidirectional auto-regressive transformer) is proposed to output the corresponding UTR. The results show that the optimized sequences of exogenous genes acquired the pattern of human endogenous gene sequence. In experimental validation, the mRNA sequence optimized by our method, compared with conventional method, shows higher protein expression. To the best of our knowledge, this is the first study by introducing deep-learning methods to integrated mRNA sequence optimization, and these results may contribute to the development of mRNA therapeutics.

Race with virus evolution: The development and application of mRNA vaccines against SARS-CoV-2

Since the COVID-19 pandemic was declared, vaccines against SARS-CoV-2 have been urgently developed around the world. On the basis of the mRNA vaccine technology developed previously, COVID-19 mRNA vaccines were promptly tested in animals, advanced to clinical trials, and then authorized for emergency use in humans. The administration of COVID-19 mRNA vaccines has successfully reduced the hospitalization and mortality caused by the viral infection, although the virus continuously evolves with its transmission. Therefore, the development of mRNA vaccine technology, including RNA modification and delivery systems, is well recognized for its contribution to moderating the harms caused by the COVID-19 pandemic. The scientists who developed these technologies, Katalin Karikó, Drew Weissman, and Pieter Cullis, were awarded the 2022 Tang Prize in Biopharmaceutical Science. In this review, we summarize the principles, safety and efficacy of as well as the immune response to COVID-19 mRNA vaccines. Since mRNA vaccine approaches could be practical for the prevention of infectious diseases, we also briefly describe mRNA vaccines against other human viral pathogens in clinical trials.

Аpplication of massive parallel reporter analysis in biotechnology and medicine

The development and functioning of an organism relies on tissue-specific gene programs. Genome regulatory elements play a key role in the regulation of such programs, and disruptions in their function can lead to the development of various pathologies, including cancers, malformations and autoimmune diseases. The emergence of high-throughput genomic studies has led to massively parallel reporter analysis (MPRA) methods, which allow the functional verification and identification of regulatory elements on a genome-wide scale. Initially MPRA was used as a tool to investigate fundamental aspects of epigenetics, but the approach also has great potential for clinical and practical biotechnology. Currently, MPRA is used for validation of clinically significant mutations, identification of tissue-specific regulatory elements, search for the most promising loci for transgene integration, and is an indispensable tool for creating highly efficient expression systems, the range of application of which extends from approaches for protein development and design of next-generation therapeutic antibody superproducers to gene therapy. In this review, the main principles and areas of practical application of high-throughput reporter assays will be discussed.

Ribo-uORF: a comprehensive data resource of upstream open reading frames (uORFs) based on ribosome profiling

Upstream open reading frames (uORFs) are typically defined as translation sites located within the 5' untranslated region upstream of the main protein coding sequence (CDS) of messenger RNAs (mRNAs). Although uORFs are prevalent in eukaryotic mRNAs and modulate the translation of downstream CDSs, a comprehensive resource for uORFs is currently lacking. We developed Ribo-uORF (http://rnainformatics.org.cn/RiboUORF) to serve as a comprehensive functional resource for uORF analysis based on ribosome profiling (Ribo-seq) data. Ribo-uORF currently supports six species: human, mouse, rat, zebrafish, fruit fly, and worm. Ribo-uORF includes 501 554 actively translated uORFs and 107 914 upstream translation initiation sites (uTIS), which were identified from 1495 Ribo-seq and 77 quantitative translation initiation sequencing (QTI-seq) datasets, respectively. We also developed mRNAbrowse to visualize items such as uORFs, cis-regulatory elements, genetic variations, eQTLs, GWAS-based associations, RNA modifications, and RNA editing. Ribo-uORF provides a very intuitive web interface for conveniently browsing, searching, and visualizing uORF data. Finally, uORFscan and UTR5var were developed in Ribo-uORF to precisely identify uORFs and analyze the influence of genetic mutations on uORFs using user-uploaded datasets. Ribo-uORF should greatly facilitate studies of uORFs and their roles in mRNA translation and posttranscriptional control of gene expression.

Accuracy and data efficiency in deep learning models of protein expression

Synthetic biology often involves engineering microbial strains to express high-value proteins. Thanks to progress in rapid DNA synthesis and sequencing, deep learning has emerged as a promising approach to build sequence-to-expression models for strain optimization. But such models need large and costly training data that create steep entry barriers for many laboratories. Here we study the relation between accuracy and data efficiency in an atlas of machine learning models trained on datasets of varied size and sequence diversity. We show that deep learning can achieve good prediction accuracy with much smaller datasets than previously thought. We demonstrate that controlled sequence diversity leads to substantial gains in data efficiency and employed Explainable AI to show that convolutional neural networks can finely discriminate between input DNA sequences. Our results provide guidelines for designing genotype-phenotype screens that balance cost and quality of training data, thus helping promote the wider adoption of deep learning in the biotechnology sector.

Nurturing Deep Tech to Solve Social Problems: Learning from COVID-19 mRNA Vaccine Development

In mRNA vaccines against COVID-19, a new technology that had never been used for approved drugs was applied and succeeded in rapid clinical use. The development and application of new technologies are critical to solving emerging public health problems therefore it is important to understand which factors enabled the rapid development of the COVID-19 mRNA vaccines. This review discusses administrative and technological aspects of rapid vaccine development. In the technological aspects, I carefully examined the technology and clinical development histories of BioNTech and Moderna by searching their publication, patent application and clinical trials. Compared to the case of Japanese company that has not succeeded in the rapid development of mRNA vaccine, years of in-depth technology research and clinical development experience with other diseases and viruses were found to have enhanced BioNTech and Moderna's technological readiness and contributed to rapid vaccine development against COVID-19 in addition to government administrative support. An aspect of the investments that supported the long-term research and development of mRNA vaccines is also discussed.

Scalable Functional Assays for the Interpretation of Human Genetic Variation

Scalable sequence-function studies have enabled the systematic analysis and cataloging of hundreds of thousands of coding and noncoding genetic variants in the human genome. This has improved clinical variant interpretation and provided insights into the molecular, biophysical, and cellular effects of genetic variants at an astonishing scale and resolution across the spectrum of allele frequencies. In this review, we explore current applications and prospects for the field and outline the principles underlying scalable functional assay design, with a focus on the study of single-nucleotide coding and noncoding variants.

The genetic and biochemical determinants of mRNA degradation rates in mammals

BACKGROUND

Degradation rate is a fundamental aspect of mRNA metabolism, and the factors governing it remain poorly characterized. Understanding the genetic and biochemical determinants of mRNA half-life would enable more precise identification of variants that perturb gene expression through post-transcriptional gene regulatory mechanisms.

RESULTS

We establish a compendium of 39 human and 27 mouse transcriptome-wide mRNA decay rate datasets. A meta-analysis of these data identified a prevalence of technical noise and measurement bias, induced partially by the underlying experimental strategy. Correcting for these biases allowed us to derive more precise, consensus measurements of half-life which exhibit enhanced consistency between species. We trained substantially improved statistical models based upon genetic and biochemical features to better predict half-life and characterize the factors molding it. Our state-of-the-art model, Saluki, is a hybrid convolutional and recurrent deep neural network which relies only upon an mRNA sequence annotated with coding frame and splice sites to predict half-life (r=0.77). The key novel principle learned by Saluki is that the spatial positioning of splice sites, codons, and RNA-binding motifs within an mRNA is strongly associated with mRNA half-life. Saluki predicts the impact of RNA sequences and genetic mutations therein on mRNA stability, in agreement with functional measurements derived from massively parallel reporter assays.

CONCLUSIONS

Our work produces a more robust ground truth for transcriptome-wide mRNA half-lives in mammalian cells. Using these revised measurements, we trained Saluki, a model that is over 50% more accurate in predicting half-life from sequence than existing models. Saluki succinctly captures many of the known determinants of mRNA half-life and can be rapidly deployed to predict the functional consequences of arbitrary mutations in the transcriptome.