Limits of deep-learning-based RNA prediction methods
- 1Science for Life Laboratory and Department of Biochemistry and Biophysics, Stockholm University , Solna 171 21 ,
Abstract
In recent years, tremendous advances have been made in predicting protein structures and protein-protein interactions, but progress on RNA structure, alone or in complex with other macromolecules, has lagged behind despite recent methodological developments. It remains unclear whether reported gains in accuracy generalize to novel RNA structures. Here, we benchmark 10 deep learning methods on an independent, non-redundant dataset of monomeric RNAs and RNA complexes. For monomeric RNA, AlphaFold3 and Boltz-1 achieve the best overall performance, but accuracy depends greatly on similarity to the training set: predictions are reliable mainly for RNAs with well-defined or recurring folds, such as L-shaped and simple loop structures, while structurally divergent RNAs remain poorly modelled. Extending the evaluation to RNA complexes, AlphaFold3 and Boltz-1 substantially outperform HelixFold3 and RosettaFoldNA, with no statistically significant difference between the two leading methods on most measures. Together, these results show that current methods are constrained by limited, structurally biased training data and recognize recurring motifs rather than generalize to novel RNA folds.
Related
- Antibody CDR-H3 loop flexibility: Insights from X-ray crystallography, structural bioinformatics, and the limits of current deep learning methods
- Anti-CRISPR prediction using deep learning reveals an inhibitor of Cas13b nucleases
- Frustration Quenching and Network Topology of the Energy Landscape as Primary Determinants of Protein-Ligand Binding Pose Prediction by Deep Learning Models
- Automated spheroid and organoid image analysis from rule-based methods to deep learning
- Prediction of late blight severity in a large panel of potato genotypes using low-altitude aerial images and machine learning methods
BibTeX
@article{ludaic2026limits,
title = {{Limits of deep-learning-based RNA prediction methods}},
author = {Marko Ludaic and Arne Elofsson},
journal = {Nucleic Acids Research},
year = {2026},
month = aug,
doi = {10.1093/nar/gkag813},
url = {https://doi.org/10.1093/nar/gkag813},
note = {CC BY},
}
RIS
TY - JOUR
TI - Limits of deep-learning-based RNA prediction methods
AU - Marko Ludaic
AU - Arne Elofsson
PY - 2026
DA - 2026/08/06
JO - Nucleic Acids Research
DO - 10.1093/nar/gkag813
UR - https://doi.org/10.1093/nar/gkag813
AB - In recent years, tremendous advances have been made in predicting protein structures and protein-protein interactions, but progress on RNA structure, alone or in complex with other macromolecules, has lagged behind despite recent methodological developments. It remains unclear whether reported gains in accuracy generalize to novel RNA structures. Here, we benchmark 10 deep learning methods on an independent, non-redundant dataset of monomeric RNAs and RNA complexes. For monomeric RNA, AlphaFold3 and Boltz-1 achieve the best overall performance, but accuracy depends greatly on similarity to the training set: predictions are reliable mainly for RNAs with well-defined or recurring folds, such as L-shaped and simple loop structures, while structurally divergent RNAs remain poorly modelled. Extending the evaluation to RNA complexes, AlphaFold3 and Boltz-1 substantially outperform HelixFold3 and RosettaFoldNA, with no statistically significant difference between the two leading methods on most measures. Together, these results show that current methods are constrained by limited, structurally biased training data and recognize recurring motifs rather than generalize to novel RNA folds.
SN - 0305-1048
SN - 1362-4954
SN - 1362-4962
ER -