Nature Communications · 2025 · 14 citations · 60 references
Life SciencesBioinformaticsMolecular BiologyDrug DesignGene ExpressionProtein TranslationProtein SynthesisRibosome ProfilingRna EngineeringTranslation ControlTranslation RegulationTranslational RegulationCircular Messenger RnaMessenger Rna DesignMessenger Rna TherapeuticsNerve Growth Factor SignalingNoncoding Rna TherapeuticsRibosome Profiling Data AnalysisSequence Space ExplorationTherapeutic Mrna Optimization
Messenger RNA (mRNA) therapeutics show immense promise, but their efficacy is limited by suboptimal protein expression. Here, we present RiboDecode, a deep learning framework that generates mRNA codon sequences for enhanced mRNA translation. RiboDecode introduces several advances, including direct learning from large-scale ribosome profiling data and generative exploration of a large sequence space. In silico analysis demonstrates RiboDecode’s robust predictive accuracy for unseen genes and cellular environments. In vitro experiments showed substantial improvements in protein expression, significantly outperforming past methods. In addition, RiboDecode enables mRNA design with consideration of cellular context and demonstrates robust performance across different mRNA formats, including m1Ψ-modified and circular mRNAs, an important feature for mRNA therapeutics. In vivo mouse studies showed that optimized influenza hemagglutinin mRNAs induce ten times stronger neutralizing antibody responses against influenza virus compared to the unoptimized sequence. In an optic nerve crush model, optimized nerve growth factor mRNAs achieve equivalent neuroprotection of retinal ganglion cells at one-fifth the dose of the unoptimized sequence. Collectively, RiboDecode represents a paradigm shift from rule-based to a data-driven, context-aware approach for mRNA therapeutic applications, enabling the development of more potent and dose-efficient treatments. Messenger RNA therapeutics hold great promise but are limited by suboptimal protein expression. Here, the authors develop RiboDecode, a deep learning framework that generates optimized mRNA codon sequences to significantly enhance translation efficiency and therapeutic efficacy in both vaccines and protein replacement therapies.
60
Deep Residual Learning for Image Recognition
Kaiming He, Xiangyu Zhang, Shaoqing Ren et al. · 2016 · 222.1K citations · Full text
Fast gapped-read alignment with Bowtie 2
Ben Langmead, Steven L. Salzberg · Nature Methods · 2012 · 61.1K citations
STAR: ultrafast universal RNA-seq aligner
Alexander Dobin, Carrie Davis, Felix Schlesinger et al. · Bioinformatics · 2012 · 56.2K citations · Full text
Highly accurate protein structure prediction with AlphaFold
John Jumper, Alexander Pritzel, Tim Green et al. · Nature · 2021 · 45.2K citations · Full text
Cutadapt removes adapter sequences from high-throughput sequencing reads
Marcel Martin · EMBnet journal · 2011 · 35.8K citations · Full text