Reportedly: USTC releases open-source RNA sequence foundation model structRFM, published in Nature
Reported by IT Home · not yet confirmed by the company or a second independent outlet. We update this page when it is.
University of Science and Technology of China (USTC) released open-source RNA sequence foundation model structRFM, improving performance on RNA structure prediction and function inference downstream tasks, IT Home reported. The research was published in Nature on November 5, 2026. This is USTC's 16th publication in Nature, Science, or Cell in 2026 as first affiliation and corresponding unit.
China context
- Original name
- 中国科学技术大学
- Outside China
- Not stated in the sources yet
- Claims
- Company-reported; not yet independently evaluated
- For builders
- Developers outside China can look for the model weights and code to be released; if available, they may integrate structRFM into RNA analysis pipelines, but must verify licensing terms.
- For investors
- Investors in biotech and AI-for-science should monitor whether structRFM gains adoption in drug discovery workflows, as that could indicate commercial value, but no licensing or partnership details are provided.
Translated from Chinese. Quotes and facts link to the original sources.
The MIRACLE lab at USTC, led by Professor Shaohua Zhou and Special Researcher Peng Xiong, developed structRFM, an open-source RNA sequence foundation model. The model incorporates structural knowledge into large-scale sequence learning, using approximately 21 million computationally annotated sequence-secondary structure pairs. It aims to improve performance on RNA structure prediction and function inference tasks, with potential applications in AI-driven scientific research, biological mechanism understanding, and drug discovery. The research was published in Nature on November 5, 2026.
structRFM uses a structure-guided pre-training framework that integrates RNA secondary structure information into sequence learning. The model was trained on about 21 million computationally annotated sequence-secondary structure pairs. The approach is designed to maintain adaptability and generalization across different structural and functional tasks.
Researchers and pharmaceutical companies working on RNA biology gain a new open-source tool that could reduce the cost and time of RNA structure prediction and function analysis, potentially accelerating drug discovery efforts. The publication in Nature indicates the model has passed peer review, but independent validation of its performance on diverse RNA families is still needed.
The open-source release of structRFM could lower barriers for academic and commercial teams to apply advanced RNA modeling in drug discovery and synthetic biology, potentially reducing R&D costs. However, the evidence does not specify licensing terms or commercial use restrictions.
The next verifiable signal is whether the model weights and code are made publicly available on a platform like Hugging Face or GitHub, and whether independent researchers reproduce the reported performance improvements on standard RNA benchmarks.