Predicting RNA 3D structure and conformers using a pre-trained secondary structure model and structure-aware attention

ABSTRACT Determining RNA 3D structure and conformers remains a grand challenge in structural biology. In this work, we propose trRosettaRNA2, a deep learning-based end-to-end approach to this problem. Considering the scarcity of RNA 3D structure data, trRosettaRNA2 integrates an auxiliary secondary structure (SS) prior module, pre-trained on extensive SS data, to generate informative base-pairing priors. This module also serves as an independent RNA secondary structure prediction method, trRNA2-SS, and achieves state-of-the-art performance. To enable end-to-end prediction, trRosettaRNA2 employs SS-aware attention to generate RNA 3D structure and conformers. Rigorous benchmarks demonstrate that trRosettaRNA2 outperforms other RNA 3D structure prediction methods, despite using significantly fewer parameters and computational resources. Notably, its flexibility in leveraging diverse secondary structure inputs provides a pathway to generate accurate 3D structure and explore the RNA conformers. Based on trRosettaRNA2, our group, Yang-Server, ranks as the top automated server for RNA structure prediction in the CASP16 blind test, surpassing AlphaFold3. This performance highlights that trRosettaRNA2 represents a solid step towards the “AlphaFold moment” for RNA structure prediction. Application to the RNase P RNA demonstrates that trRosettaRNA2 successfully captures its structural heterogeneity even without requiring experimental data, showing its potential to predict RNA conformational ensembles.