AI-Driven Model Enhances mRNA Therapeutics Through Protein Prediction

In a breakthrough for the field of biomedicine, researchers from The University of Texas at Austin (UT Austin) and Sanofi have developed an innovative artificial intelligence (AI) model, known as RiboNN, which enhances the design and efficiency of messenger RNA (mRNA) therapeutics by predicting protein production across various cell types. This significant advancement, published in the journal Nature Biotechnology on July 28, 2025, addresses long-standing challenges in drug and vaccine development, particularly in optimizing the therapeutic efficacy of mRNA-based treatments.

Messenger RNA plays a crucial role in directing protein synthesis in cells, which is fundamental in the development of mRNA vaccines and therapies aimed at combatting diseases such as cancer, genetic disorders, and viral infections. The new model predicts how efficiently specific mRNA sequences will translate into proteins, thereby minimizing the need for extensive trial-and-error experimentation typically associated with therapeutic development. This not only expedites the process but also enhances the precision of mRNA-based treatments.

According to Dr. Can Cenik, an associate professor of molecular biosciences at UT Austin and co-lead of the research, the model was born from a curiosity-driven research initiative that began over six years ago. "When we started this project, there was no obvious application. We were curious about whether cells coordinate which mRNAs they produce and how efficiently they are translated into proteins," Dr. Cenik explained. His collaborator, Vikram Agarwal, head of mRNA platform design data science at Sanofi’s mRNA Center of Excellence, highlighted that RiboNN represents a significant leap forward in understanding mRNA translation efficiency.

The development of RiboNN was supported by substantial funding from the National Institutes of Health, The Welch Foundation, and the Lonestar6 supercomputer at UT Austin's Texas Advanced Computing Center. In extensive testing involving over 140 human and mouse cell types, RiboNN demonstrated approximately double the accuracy in predicting translation efficiency compared to earlier models. This capability could potentially transform how researchers approach the development of treatments for cancer and infectious diseases.

The implications of this research extend far beyond mere protein synthesis. As Dr. Cenik noted, the goal of RiboNN is not simply to enhance production but also to allow for targeted therapies designed for specific cell types, such as those in the liver or immune system. This tailored approach opens up new possibilities for addressing a variety of medical conditions with unprecedented precision.

The study also sheds light on the coordination of gene transcription and translation processes. A companion paper in Nature Biotechnology reveals that mRNAs with similar biological functions are translated into proteins at comparable levels across different cell types, affirming the notion that these processes are intricately linked. This finding could further bolster the development of therapies that are precisely calibrated to elicit desired biological responses.

As the landscape of mRNA therapeutics continues to evolve, the RiboNN model stands poised to play a pivotal role in the future of drug development, potentially leading to faster and more effective treatments for a range of diseases. The integration of AI-driven predictive models into biomedical research heralds a new era of innovation, demonstrating the power of interdisciplinary collaboration in solving complex health challenges.

### Conclusion The development of the RiboNN model marks a significant milestone in the field of mRNA therapeutics, indicating a promising future for accelerated and targeted treatments. As researchers continue to refine these methodologies, the potential for AI to revolutionize drug discovery and vaccine development becomes increasingly apparent, paving the way for next-generation therapies that can respond adeptly to the evolving nature of health challenges.

### References 1. Zheng D, Persyn L, Wang J, et al. (2025). Predicting the translation efficiency of messenger RNA in mammalian cells. Nature Biotechnology. doi: 10.1038/s41587-025-02712-x.