DRfold2 is a deep learning-based tool that enables efficient and accurate RNA structure prediction
PLOS Biology · 17.02.2026 · 24 (2) · e3003659
DOI: 10.1371/journal.pbio.3003659 ↗Kurz gefasst
Fachartikel in der Zeitschrift PLOS Biology, online erschienen am 17.02.2026. Verfasst von Yang Li und 5 weiteren Autor:innen. Beteiligte Einrichtung: Cancer Science Institute of Singapore, National University of Singapore, Singapore, Singapore. Hauptthemen laut MEDLINE: Deep Learning, RNA, Nucleic Acid Conformation, Computational Biology. Die Arbeit wurde bisher 15-mal zitiert (Crossref) und nennt 58 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Ministry of Education - Singapore, Agency for Science, Technology and Research (A*STAR), Singapore, Natural Science Foundation of Ningxia Province.
Abstract (Originalsprache)
RNA structures are essential for understanding their biological functions and developing RNA-targeted therapeutics. However, accurate RNA structure prediction from sequence remains a crucial challenge. We introduce DRfold2, a deep learning framework that integrates a novel pre-trained RNA Composite Language Model (RCLM) with a denoising structure module for end-to-end RNA structure prediction. Based solely on single sequence, DRfold2 achieves superior performance in both global topology and secondary structure predictions over other state-of-the-art approaches across multiple benchmark tests from diverse species. Detailed analyses reveal that the improvements primarily stem from the RCLM’s ability to capture co-evolutionary pattern and the effective denoising process, with a more than 100% increase in contact prediction precision compared to existing methods. Furthermore, DRfold2 demonstrates high complementarity with AlphaFold3, achieving statistically significant accuracy gains when integrated into our optimization framework. By uniquely combining composite language modeling, denoising-based end-to-end learning, and deep learning-guided post-optimization, DRfold2 establishes a distinct direction for advancing ab initio RNA structure prediction.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Substanzen
Autor:innen und Institutionen
- Cancer Science Institute of Singapore, National University of Singapore, Singapore, Singapore.
- School of Science, Ningxia Medical University, Yinchuan, China.
- Center for AI and Computational Biology, Suzhou Institute of Systems Medicine, Chinese Academy of Medical Sciences & Peking Union Medical College, Suzhou, China.
- Department of Computer Science, School of Computing, National University of Singapore, Singapore, Singapore.
- School of Life Science and Technology, Institute of Science Tokyo, Tokyo, Japan.
- Department of Biochemistry, Yong Loo Lin School of Medicine, National University of Singapore, Singapore, Singapore.
ORCID-Profile: Yang Li, Yang Zhang
Förderung
- Ministry of Education - Singapore
- Agency for Science, Technology and Research (A*STAR), Singapore
- Natural Science Foundation of Ningxia Province
Zitieren
APA
Li, Y., Feng, C., Zhang, X., Tsukiyama, S., Feng, D., & Zhang, Y. (2026). DRfold2 is a deep learning-based tool that enables efficient and accurate RNA structure prediction. PLOS Biology, 24(2), e3003659. https://doi.org/10.1371/journal.pbio.3003659
Vancouver
Li Y, Feng C, Zhang X, Tsukiyama S, Feng D, Zhang Y. DRfold2 is a deep learning-based tool that enables efficient and accurate RNA structure prediction. PLOS Biology. 2026;24(2):e3003659. doi:10.1371/journal.pbio.3003659
BibTeX
@article{Li2026,
author = {Li, Yang and Feng, Chenjie and Zhang, Xi and Tsukiyama, Sho and Feng, Duanyu and Zhang, Yang},
title = {DRfold2 is a deep learning-based tool that enables efficient and accurate RNA structure prediction},
journal = {PLOS Biology},
year = {2026},
volume = {24},
number = {2},
pages = {e3003659},
doi = {10.1371/journal.pbio.3003659}
}
EUVIMED