Deep generative optimization of mRNA codon sequences for enhanced mRNA translation and therapeutic efficacy
Open accessPublished version of a preprint
Abstract
ABSTRACT 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 m 1 Ψ-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.
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Integrity signals
Registry
Cites
Statistics
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Cites
Deep Residual Learning for Image Recognition
He, K., Zhang, X., Ren, S., & Sun, J. (2016) · 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2016.90 · Cited by 235,474 · see citing works
ImageNet classification with deep convolutional neural networks
Krizhevsky, A., Sutskever, I., & Hinton, G. E. (2017) · Communications of the ACM · doi:10.1145/3065386 · Cited by 112,155 · see citing works
Highly accurate protein structure prediction with AlphaFold
Jumper, J., Evans, R., Pritzel, A., Green, T., Figurnov, M., Ronneberger, O., Tunyasuvunakool, K., Bates, R., Žídek, A., Potapenko, A., et al. (2021) · Nature · doi:10.1038/s41586-021-03819-2 · Cited by 43,764 · see citing works
Efficacy and Safety of the mRNA-1273 SARS-CoV-2 Vaccine
Baden, L. R., El Sahly, H. M., Essink, B., Kotloff, K., Frey, S., Novak, R., Diemert, D., Spector, S. A., Rouphael, N., Creech, C. B., et al. (2021) · New England Journal of Medicine · doi:10.1056/nejmoa2035389 · Cited by 10,119 · see citing works
Self-citationGlobal quantification of mammalian gene expression control
Schwanhäusser, B., Busse, D., Li, N., Dittmar, G., Schuchhardt, J., Wolf, J., Chen, W., & Selbach, M. (2011) · Nature · doi:10.1038/nature10098 · Cited by 6,566 · cited 2× · see citing works
“These fi ndings also align with the broader framework of gene expression regulation: while mRNA abundance is a prerequisite for protein synthesis (consistent with its predictive dominance), our approach directly targets translational control, a critical layer high-lighted by Schwanhäusser et al. 45 , by leveraging Ribo…”
Robust deep learning based protein sequence design using ProteinMPNN
Dauparas, J., Anishchenko, I., Bennett, N., Bai, H., Ragotte, R. J., Milles, L. F., Wicky, B. I. M., Courbet, A., de Haas, R. J., Bethel, N., et al. (2022) · doi:10.1101/2022.06.03.494563 · Cited by 2,107 · see citing works
Robust deep learning–based protein sequence design using ProteinMPNN
Dauparas, J., Anishchenko, I., Bennett, N., Bai, H., Ragotte, R. J., Milles, L. F., Wicky, B. I. M., Courbet, A., de Haas, R. J., Bethel, N., et al. (2022) · Science · doi:10.1126/science.add2187 · Cited by 2,070 · see citing works
Lu, P., Vogel, C., Wang, R., Yao, X., & Marcotte, E. M. (2007) · Nature Biotechnology · doi:10.1038/nbt1270 · Cited by 1,139 · see citing works
Codon optimality, bias and usage in translation and mRNA decay
Hanson, G., & Coller, J. (2018) · Nature Reviews Molecular Cell Biology · doi:10.1038/nrm.2017.91 · Cited by 825 · see citing works
mRNA structure regulates protein expression through changes in functional half-life
Mauger, D. M., Cabral, B. J., Presnyak, V., Su, S. V., Reid, D. W., Goodman, B., Link, K., Khatwani, N., Reynders, J., Moore, M. J., et al. (2019) · Proceedings of the National Academy of Sciences · doi:10.1073/pnas.1908052116 · Cited by 555 · see citing works
Algorithm for optimized mRNA design improves stability and immunogenicity
Zhang, H., Zhang, L., Lin, A., Xu, C., Li, Z., Liu, K., Liu, B., Ma, X., Zhao, F., Jiang, H., et al. (2023) · Nature · doi:10.1038/s41586-023-06127-z · Cited by 498 · cited 16× · see citing works
“LinearDesign uses a linear programming approach to explore a wider space of sequence variants compared to previous methods and showed superior performance over the previous codon optimization methods.”
InfluentialA Quantitative Proteome Map of the Human Body
Jiang, L., Wang, M., Lin, S., Jian, R., Li, X., Chan, J., Dong, G., Fang, H., Robinson, A. E., Aguet, F., et al. (2020) · Cell · doi:10.1016/j.cell.2020.08.036 · Cited by 458 · see citing works
Self-citationCombinatorial optimization of mRNA structure, stability, and translation for RNA-based therapeutics
Leppek, K., Byeon, G. W., Kladwang, W., Wayment-Steele, H. K., Kerr, C. H., Xu, A. F., Kim, D. S., Topkar, V. V., Choe, C., Rothschild, D., et al. (2021) · doi:10.1101/2021.03.29.437587 · Cited by 399 · see citing works
Combinatorial optimization of mRNA structure, stability, and translation for RNA-based therapeutics
Leppek, K., Byeon, G. W., Kladwang, W., Wayment-Steele, H. K., Kerr, C. H., Xu, A. F., Kim, D. S., Topkar, V. V., Choe, C., Rothschild, D., et al. (2022) · Nature Communications · doi:10.1038/s41467-022-28776-w · Cited by 399 · see citing works
Predicting RNA-seq coverage from DNA sequence as a unifying model of gene regulation
Linder, J., Srivastava, D., Yuan, H., Agarwal, V., & Kelley, D. R. (2025) · Nature Genetics · doi:10.1038/s41588-024-02053-6 · Cited by 265 · see citing works
LinearFold: linear-time approximate RNA folding by 5'-to-3' dynamic programming and beam search
Huang, L., Zhang, H., Deng, D., Zhao, K., Liu, K., Hendrix, D. A., & Mathews, D. H. (2019) · Bioinformatics · doi:10.1093/bioinformatics/btz375 · Cited by 207 · cited 1× · see citing works
“Current MFE prediction tools, such as RNAfold 35 and Linearfold 36 , use dynamic programming.”
Zrimec, J., Börlin, C. S., Buric, F., Muhammad, A. S., Chen, R., Siewers, V., Verendel, V., Nielsen, J., Töpel, M., & Zelezniak, A. (2020) · Nature Communications · doi:10.1038/s41467-020-19921-4 · Cited by 174 · see citing works
Gene expression is encoded in all parts of a co-evolving interacting gene regulatory structure
Zrimec, J., Buric, F., Muhammad, A. S., Chen, R., Verendel, V., Töpel, M., & Zelezniak, A. (2019) · doi:10.1101/792531 · Cited by 173 · see citing works
Zhuang, X., Qi, Y., Wang, M., Yu, N., Nan, F., Zhang, H., Tian, M., Li, C., Lu, H., & Jin, N. (2020) · Vaccines · doi:10.3390/vaccines8010123 · Cited by 132 · see citing works
Optimization of non-coding regions for a non-modified mRNA COVID-19 vaccine
Gebre, M. S., Rauch, S., Roth, N., Yu, J., Chandrashekar, A., Mercado, N. B., He, X., Liu, J., McMahan, K., Martinot, A., et al. (2022) · Nature · doi:10.1038/s41586-021-04231-6 · Cited by 127 · see citing works
Self-citation
Flagged first, then by citations
- Authors
- Li, Y., Wang, F., Yang, J., Han, Z., Chen, L., Jiang, W., Zhou, H., Li, T., Tang, Z., Deng, J., He, X., Zha, G., Hu, Z., Hu, Y., Wu, L., Zhan, C., Sun, C., He, Y., & Xie, Z.
- Affiliations
- State Key Laboratory of Ophthalmology (CN)
- Sun Yat-sen University (CN)
- Chinese Academy of Sciences (CN)
- Shanghai Medical College of Fudan University (CN)
- Corresponding author
- He, Yao
- Xie, Zhi
- ORCID
- Venue
- Nature Communications
- Published
- 2025-11-12
- Volume
- 16
- Issue
- 1
- Pages
- 9957-9957
- ISSN
- 2041-1723 (ISSN-L)
- PMC
- PMC12612108
- OpenAlex
- W4416142807
- Cited by
- 37
- References
- 20
- Citations by year
- 2026: 33 · 2025: 3 · 2024: 1
- FWCI
- 10.82
- Type
- journal article
- Language
- en
- Open access
- gold
- doi.org
- Licence
- Locations
- Nature Communications (publishedVersion, cc-by-nc-nd)
- bioRxiv (Cold Spring Harbor Laboratory) (acceptedVersion)
- Research Square (acceptedVersion, cc-by)
- Europe PMC (PubMed Central) (publishedVersion, cc-by-nc-nd)
- PubMed (publishedVersion)
- DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
- Funders
- National Natural Science Foundation of China (32470705, 2022YFF1203100)
- Topics
- Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › RNA and protein synthesis mechanisms
- Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › RNA Interference and Gene Delivery
- Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › RNA modifications and cancer
- Keywords
- mRNA therapeutics, deep generative models, ribosome profiling, mRNA translation efficiency, protein expression, neutralizing antibodies, nerve growth factor, optic nerve crush, retinal ganglion cell neuroprotection, mRNA vaccines, protein replacement therapy
- MeSH terms
- Animals, Codon, Humans, Mice, Inbred C57BL, Orthomyxoviridae Infections, Ribosomes, RNA, Messenger, Protein Biosynthesis, Mice, Antibodies, Neutralizing
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
