USTC Releases Open-Source RNA Sequence Foundation Model structRFM Published in Nature

On November 5, 2026, the University of Science and Technology of China (USTC) announced that a team from its MIRACLE lab, led by Professor Zhou Shaohua and special-term researcher Xiong Peng, has released an open-source RNA sequence foundation model called structRFM. According to the announcement, the model integrates structural knowledge into large-scale sequence learning, improving performance on a range of downstream tasks such as RNA structure prediction and function inference. The team says structRFM is expected to support AI-driven scientific research, the understanding of biological mechanisms, and drug discovery. The researchers published their results in the journal Nature in the early hours of November 5, 2026. The same day, USTC noted that this was its 16th paper published in 2026 with the university listed as the primary and corresponding institution in Nature, Science, or Cell. The announcement also explains the biological context behind the work: RNA participates in processes such as protein synthesis, gene expression regulation, and splicing, and its function depends on both sequence and folding structure. The model framework described in the paper is a structure-guided pretraining approach.