Skip to main content

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

  • Article
  • Published:

Deciphering RNA–ligand binding specificity with GerNA-Bind

Nature Machine Intelligence volume 7, pages 1996–2008 (2025)Cite this article

Subjects

A preprint version of the article is available at bioRxiv.

Abstract

RNA molecules are essential regulators of biological processes and promising therapeutic targets for various diseases. Discovering small molecules that selectively bind to specific RNA conformations remains challenging due to RNA’s structural complexity and the limited availability of high-resolution data. Here we introduce GerNA-Bind, a geometric deep learning framework to predict RNA–ligand binding specificity by integrating multistate RNA–ligand representations and interactions. GerNA-Bind achieves state-of-the-art performance on multiple benchmark datasets and excels in predicting interactions for low-homology RNA–ligand pairs. It achieves a 20.8% improvement in precision for binding-site prediction compared with AlphaFold3. Furthermore, it offers informative, well-calibrated predictions with built-in uncertainty quantification. In a large-scale virtual screening application, GerNA-Bind identified 18 structurally diverse compounds targeting the oncogenic MALAT1 RNA, with experimentally confirmed submicromolar affinities. Among them, one leading compound selectively binds the MALAT1 triple helix, reduces its transcript levels and inhibits cancer cell migration. These findings highlight GerNA-Bind’s potential as a powerful tool for RNA-focused drug discovery, offering both accuracy and biological insight.

This is a preview of subscription content, access via your institution

Access options

Buy this article

  • Purchase on SpringerLink
  • Instant access to the full article PDF.

USD 39.95

Prices may be subject to local taxes which are calculated during checkout

Fig. 1: Overview of GerNA-Bind.
Fig. 2: GerNA-Bind in RNA–ligand binding specificity prediction.
Fig. 3: Uncertainty quantification of RNA–ligand interaction using GerNA-Bind.
Fig. 4: GerNA-Bind validation for RNA binding-site identification.
Fig. 5: GerNA-Bind assists wet-lab experiments for RNA-targeting drug discovery.

Similar content being viewed by others

Data availability

We utilized publicly accessible datasets as detailed in Methods. The preprocessed Robin and Biosensor datasets are available via Zenodo at https://doi.org/10.5281/zenodo.14808549 (ref. 61). RNA structural features were derived using three established tools: RNA-FM (via GitHub at https://github.com/ml4bio/RNA-FM) for 1D sequence embeddings, RNAfold (http://rna.tbi.univie.ac.at) for secondary structure predictions and RhoFold (via GitHub at https://github.com/ml4bio/RhoFold) for 3D structural modelling. Structural data for fine-tuning the model’s RNA–ligand interaction prediction capability were obtained from Hariboss (https://hariboss.pasteur.cloud/), with corresponding RNA tertiary structures retrieved from the RCSB Protein Data Bank (PDB) (https://www.rcsb.org/).

Code availability

The source code and the pretrained model weights of GerNA-Bind is freely available via GitHub at https://github.com/GENTEL-lab/GerNA-Bindand via Zenodo at https://doi.org/10.5281/zenodo.17509647 (ref. 62).

References

  1. Warner, K. D., Hajdin, C. E. & Weeks, K. M. Principles for targeting RNA with drug-like small molecules. Nat. Rev. Drug Discov. 17, 547–558 (2018).

    Article  Google Scholar 

  2. Childs-Disney, J. L. et al. Targeting RNA structures with small molecules. Nat. Rev. Drug Discov. 21, 736–762 (2022).

    Article  Google Scholar 

  3. Hopkins, A. L. & Groom, C. R. The druggable genome. Nat. Rev. Drug Discov. 1, 727–730 (2002).

    Article  Google Scholar 

  4. Corley, M., Burns, M. C. & Yeo, G. W. How RNA-binding proteins interact with RNA: molecules and mechanisms. Mol. Cell 78, 9–29 (2020).

    Article  Google Scholar 

  5. Xie, Y.-c, Eriksson, L. A. & Zhang, R.-b Molecular dynamics study of the recognition of ATP by nucleic acid aptamers. Nucleic Acids Res. 48, 6471–6480 (2020).

    Article  Google Scholar 

  6. Tong, Y. et al. Programming inactive RNA-binding small molecules into bioactive degraders. Nature 618, 169–179 (2023).

    Article  Google Scholar 

  7. Koehn, J. T., Felder, S. & Weeks, K. M. Innovations in targeting RNA by fragment-based ligand discovery. Curr. Opin. Struct. Biol. 79, 102550 (2023).

    Article  Google Scholar 

  8. Disney, M. D. et al. InfoRNA 2.0: a platform for the sequence-based design of small molecules targeting structured RNAs. ACS Chem. Biol. 11, 1720–1728 (2016).

    Article  Google Scholar 

  9. Sun, S., Yang, J. & Zhang, Z. RNAligands: a database and web server for RNA–ligand interactions. RNA 28, 115–122 (2022).

    Article  Google Scholar 

  10. Donlic, A. et al. R-bind 2.0: an updated database of bioactive RNA-targeting small molecules and associated RNA secondary structures. ACS Chem. Biol. 17, 1556–1566 (2022).

    Article  Google Scholar 

  11. Cai, Z., Zafferani, M., Akande, O. M. & Hargrove, A. E. Quantitative structure–activity relationship (qSAR) study predicts small-molecule binding to RNA structure. J. Med. Chem. 65, 7262–7277 (2022).

    Article  Google Scholar 

  12. Rekand, I. H. & Brenk, R. Drugpred_rna—a tool for structure-based druggability predictions for RNA binding sites. J. Chem. Inf. Model. 61, 4068–4081 (2021).

    Article  Google Scholar 

  13. Grimberg, H. et al. Machine learning approaches to optimize small-molecule inhibitors for RNA targeting. J. Cheminform. 14, 4 (2022).

    Article  Google Scholar 

  14. Deng, Z. et al. Predicting ligand–RNA binding using E3-equivariant network and pretraining. In Proc. Machine Learning for Structural Biology Workshop (NeurIPS, 2022).

  15. Yazdani, K. et al. Machine learning informs RNA-binding chemical space. Angewandte Chemie 135, e202211358 (2023).

    Article  Google Scholar 

  16. Sun, S. & Gao, L. Contrastive pre-training and 3D convolution neural network for RNA and small molecule bindingaffinity prediction. Bioinformatics 40, btae155 (2024).

    Article  Google Scholar 

  17. Baek, M. et al. Accurate prediction of protein–nucleic acid complexes using RoseTTAFoldNA. Nat. Methods 21, 117–121 (2024).

    Article  Google Scholar 

  18. Abramson, J. et al. Accurate structure prediction of biomolecular interactions with AlphaFold 3. Nature 630, 493–500 (2024).

  19. Sang, C. et al. The prediction of RNA–small molecule binding sites in RNA structures based on geometric deep learning. Int. J. Biol. Macromol. 310, 143308 (2025).

    Article  Google Scholar 

  20. Krishna, R. et al. Generalized biomolecular modeling and design with RoseTTAFold all-atom. Science 384, eadl2528 (2024).

    Article  Google Scholar 

  21. Patel, S. et al. Evoflow-RNA: generating and representing non-coding RNA with a language model. In Proc. ICLR Workshop on AI for Nucleic Acids (PMLR, 2025); https://openreview.net/forum?id=oDTBI7z52K

  22. Chen, T., Zhang, Y., Tang, S. & Chatterjee, P. Multi-objective-guided discrete flow matching for controllable biological sequence design. In Proc. ICML 2nd Generative AI for Biology Workshop (PMLR, 2025); https://openreview.net/forum?id=8YIMLoHP9J

  23. Tang, S., Zhang, Y., Tong, A. & Chatterjee, P. Gumbel-softmax score and flow matching for discrete biological sequence generation. In Proc. ICLR Deep Generation Model in Machine Learning: Theory, Principle and Efficacy (PMLR, 2025); https://openreview.net/forum?id=HnYSy4W0Vo

  24. Nguyen, E. et al. Sequence modeling and design from molecular to genome scale with evo. Science 386, eado9336 (2024).

    Article  Google Scholar 

  25. Brixi, G. et al. Genome modeling and design across all domains of life with Evo 2. Preprint at bioRxiv https://doi.org/10.1101/2025.02.18.638918 (2025).

  26. Shen, T. et al. Accurate RNA 3D structure prediction using a language model-based deep learning approach. Nat. Methods 21, 2287–2298 (2024).

    Article  Google Scholar 

  27. Townshend, B., Kaplan, M. & Smolke, C. D. Highly multiplexed selection of RNA aptamers against a small molecule library. PLoS ONE 17, e0273381 (2022).

    Article  Google Scholar 

  28. Krishnan, S. R., Roy, A. & Gromiha, M. M. Reliable method for predicting the binding affinity of RNA–small molecule interactions using machine learning. Brief. Bioinform. 25, bbae002 (2024).

    Article  Google Scholar 

  29. Wen, M. et al. Deep-learning-based drug–target interaction prediction. J. Proteome Res. 16, 1401–1409 (2017).

    Article  Google Scholar 

  30. Lee, I., Keum, J. & Nam, H. Deepconv-dti: prediction of drug–target interactions via deep learning with convolution on protein sequences. PLoS Comput. Biol. 15, e1007129 (2019).

    Article  Google Scholar 

  31. Nguyen, T. et al. Graphdta: predicting drug–target binding affinity with graph neural networks. Bioinformatics 37, 1140–1147 (2021).

    Article  Google Scholar 

  32. Panei, F. P., Torchet, R., Menager, H., Gkeka, P. & Bonomi, M. Hariboss: a curated database of RNA–small molecules structures to aid rational drug design. Bioinformatics 38, 4185–4193 (2022).

    Article  Google Scholar 

  33. Su, H., Peng, Z. & Yang, J. Recognition of small molecule–RNA binding sites using RNA sequence and structure. Bioinformatics 37, 36–42 (2021).

    Article  Google Scholar 

  34. Discovery, C. et al. Chai-1: decoding the molecular interactions of life. Preprint at bioRxiv https://doi.org/10.1101/2024.10.10.615955 (2024).

  35. Gutschner, T. et al. The noncoding RNA MALAT1 is a critical regulator of the metastasis phenotype of lung cancer cells. Cancer Res. 73, 1180–1189 (2013).

    Article  Google Scholar 

  36. Ji, P. et al. Malat-1, a novel noncoding RNA, and thymosin β4 predict metastasis and survival in early-stage non-small cell lung cancer. Oncogene 22, 8031–8041 (2003).

    Article  Google Scholar 

  37. Guo, F. et al. Inhibition of metastasis-associated lung adenocarcinoma transcript 1 in caski human cervical cancer cells suppresses cell proliferation and invasion. Acta Biochim. Biophys. Sin. 42, 224–229 (2010).

    Article  Google Scholar 

  38. Tano, K. et al. Malat-1 enhances cell motility of lung adenocarcinoma cells by influencing the expression of motility-related genes. FEBS Lett. 584, 4575–4580 (2010).

    Article  Google Scholar 

  39. Arun, G. et al. Differentiation of mammary tumors and reduction in metastasis upon MALAT1 lncRNA loss. Genes Dev. 30, 34–51 (2016).

    Article  Google Scholar 

  40. Wilusz, J. E. et al. A triple helix stabilizes the 3 ends of long noncoding RNAs that lack poly(A) tails. Genes Dev. 26, 2392–2407 (2012).

    Article  Google Scholar 

  41. Brown, J. A. et al. Structural insights into the stabilization of MALAT1 noncoding RNA by a bipartite triple helix. Nat. Struct. Mol. Biol. 21, 633–640 (2014).

    Article  Google Scholar 

  42. Donlic, A. et al. Discovery of small molecule ligands for MALAT1 by tuning an RNA-binding scaffold. Angewandte Chemie 130, 13426–13431 (2018).

    Article  Google Scholar 

  43. Topscience Compound Library (Topscience Corporation, accessed May 2025); https://www.targetmol.cn/service/topscience-database

  44. Pernak, M. et al. Development of comprehensive screening and assessment assays for small-molecule ligands of MALAT1 lncRNA. ACS Chem. Biol. 20, 1068–1076 (2025).

    Article  Google Scholar 

  45. Abulwerdi, F. A. et al. Selective small-molecule targeting of a triple helix encoded by the long noncoding RNA, MALAT1. ACS Chem. Biol. 14, 223–235 (2019).

    Article  Google Scholar 

  46. Rocca, R. et al. Hit identification of novel small molecules interfering with MALAT1 triplex by a structure-based virtual screening. Archiv der Pharmazie 356, 2300134 (2023).

    Article  Google Scholar 

  47. Krishnamurthy, M., Schirle, N. T. & Beal, P. A. Screening helix-threading peptides for RNA binding using a thiazole orange displacement assay. Bioorganic Med. Chem. 16, 8914–8921 (2008).

    Article  Google Scholar 

  48. Wang, Z.-F. et al. The hairpin form of r(G4C2)expin c9als/FTD is repeat-associated non-ATG translated and a target for bioactive small molecules. Cell Chem. Biol. 26, 179–190 (2019).

    Article  Google Scholar 

  49. Tran, T. & Disney, M. D. Identifying the preferred RNA motifs and chemotypes that interact by probing millions of combinations. Nat. Commun. 3, 1125 (2012).

    Article  Google Scholar 

  50. Liu, X. et al. Targeted degradation of the oncogenic microRNA 17-92 cluster by structure-targeting ligands. J. Am. Chem. Soc. 142, 6970–6982 (2020).

    Article  Google Scholar 

  51. Tong, Y. et al. Transcriptome-wide mapping of small-molecule RNA-binding sites in cells informs an isoform-specific degrader of qsox1 mRNA. J. Am. Chem. Soc. 144, 11620–11625 (2022).

    Article  Google Scholar 

  52. Zhang, P. et al. Reprogramming of protein-targeted small-molecule medicines to RNA by ribonuclease recruitment. J. Am. Chem. Soc. 143, 13044–13055 (2021).

    Article  Google Scholar 

  53. Gupta, R. A. et al. Long non-coding RNA hotair reprograms chromatin state to promote cancer metastasis. Nature 464, 1071–1076 (2010).

    Article  Google Scholar 

  54. Ruszkowska, A., Ruszkowski, M., Hulewicz, J. P., Dauter, Z. & Brown, J. A. Molecular structure of a U•AU-rich RNA triple helix with 11 consecutive base triples. Nucleic Acids Res. 48, 3304–3314 (2020).

    Article  Google Scholar 

  55. Adasme, M. F. et al. Plip 2021: expanding the scope of the protein–ligand interaction profiler to DNA and RNA. Nucleic Acids Res. 49, W530–W534 (2021).

    Article  Google Scholar 

  56. Chen, J. et al. Interpretable RNA foundation model from unannotated data for highly accurate RNA structure and function predictions. Preprint at https://arxiv.org/abs/2204.00300 (2022).

  57. Lorenz, R. et al. ViennaRNA package 2.0. Algorithms Mol. Biol. 6, 1–14 (2011).

    Article  Google Scholar 

  58. Kipf, T. N. & Welling, M. Semi-supervised classification with graph convolutional networks. In Proc. International Conference on Learning Representations (ICLR, 2017); https://openreview.net/forum?id=SJU4ayYgl

  59. Liao, Y.-L. & Smidt, T. Equiformer: equivariant graph attention transformer for 3D atomistic graphs. In Proc. International Conference on Learning Representations (ICLR, 2023); https://openreview.net/forum?id=KwmPfARgOTD

  60. Sensoy, M., Kaplan, L. & Kandemir, M. Evidential deep learning to quantify classification uncertainty. In Proc. Advances in Neural Information Processing Systems (NeurIPS, 2018); https://dl.acm.org/doi/pdf/10.5555/3327144.3327239

  61. Yunpeng, X. Data of GerNA-Bind. Zenodo https://doi.org/10.5281/zenodo.14808549 (2025).

  62. Yunpeng, X. Gentel-lab/gerna-bind: v.1.0.0. Zenodo https://doi.org/10.5281/zenodo.17509647 (2025).

Download references

Acknowledgements

This study has been supported by the National Natural Science Foundation of China (grant numbers 62402314, 62372234, 62072243 and 22207135), Lingang Laboratory (grant number LG8888), Natural Science Foundation of Shanghai (grant number 24ZR1440600), the Guangdong Basic and Applied Basic Research Foundation (grant number 2023A1515012616), the Young Elite Scientists Sponsorship Program by CAST (grant number 2023QNRC001), the Science and Technology Commission of Shanghai Municipality (grant number 24510714300) and the project from Smart Medical Innovation Technology Center-GDUT (grant number ZYZX24-011). S.Z. acknowledges funding from the Asian Young Scientist Fellowship.

Author information

Author notes

  1. These authors contributed equally: Yunpeng Xia, Jiayi Li, Yi-Ting Chu.

Authors and Affiliations

  1. Global Institute of Future Technology, Shanghai Jiao Tong University, Shanghai, China

    Yunpeng Xia & Shuangjia Zheng

  2. School of Computer Science and Technology, Nanjing University of Science and Technology, Nanjing, China

    Yunpeng Xia & Dong-Jun Yu

  3. School of Mathematics and Physics, Xi’an Jiaotong-Liverpool University, Suzhou, China

    Jiayi Li

  4. Department of Biostatistics, Harvard University, Boston, MA, USA

    Jiayi Li

  5. School of Biomedical and Pharmaceutical Sciences, Guangdong University of Technology, Guangzhou, China

    Yi-Ting Chu, Jing Chen & Xiu-Cai Chen

  6. School of Computer Science and Engineering, Sun Yat-sen University, Guangzhou, China

    Jiahua Rao

  7. School of Computer Science, McGill University, Montreal, Quebec, Canada

    Chenqing Hua

  8. Smart Medical Innovation Technology Center, Guangdong University of Technology, Guangzhou, China

    Xiu-Cai Chen

  9. Shanghai Key Laboratory for Antibody–Drug Conjugates with Innovative Target, Shanghai Jiao Tong University, Shanghai, China

    Shuangjia Zheng

  10. School of Artificial Intelligence, Shanghai Jiao Tong University, Shanghai, China

    Shuangjia Zheng

Authors
  1. Yunpeng Xia
    View author publications

    Search author on:PubMed Google Scholar

  2. Jiayi Li
    View author publications

    Search author on:PubMed Google Scholar

  3. Yi-Ting Chu
    View author publications

    Search author on:PubMed Google Scholar

  4. Jiahua Rao
    View author publications

    Search author on:PubMed Google Scholar

  5. Jing Chen
    View author publications

    Search author on:PubMed Google Scholar

  6. Chenqing Hua
    View author publications

    Search author on:PubMed Google Scholar

  7. Dong-Jun Yu
    View author publications

    Search author on:PubMed Google Scholar

  8. Xiu-Cai Chen
    View author publications

    Search author on:PubMed Google Scholar

  9. Shuangjia Zheng
    View author publications

    Search author on:PubMed Google Scholar

Contributions

S.Z. conceived and supervised the project. Y.X., J.R. and S.Z. contributed to the algorithm implementation. J.L. and Y.X. performed the data preprocessing. Y.X., S.Z., J.L. and D.-J.Y. contributed to the visualization implementation. Y.T.-C., J.C. and X.-C.C. conducted the wet-lab experiments. S.Z., Y.X., X.-C.C., C.H. and J.L. wrote the paper. All authors were involved in the discussion and proofreading.

Corresponding authors

Correspondence to Xiu-Cai Chen or Shuangjia Zheng.

Ethics declarations

Competing interests

The authors declare no competing interests.

Peer review

Peer review information

Nature Machine Intelligence thanks Yiliang Ding and the other, anonymous, reviewer(s) for their contribution to the peer review of this work.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Supplementary Information

Supplementary Notes 1–7, Figs. 1–15 and Tables 1–11.

Reporting Summary

Rights and permissions

Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Xia, Y., Li, J., Chu, YT. et al. Deciphering RNA–ligand binding specificity with GerNA-Bind. Nat Mach Intell 7, 1996–2008 (2025). https://doi.org/10.1038/s42256-025-01154-z

Download citation

  • Received:

  • Accepted:

  • Published:

  • Version of record:

  • Issue date:

  • DOI: https://doi.org/10.1038/s42256-025-01154-z

Search

Advanced search

Quick links

Nature Briefing: Translational Research

Sign up for the Nature Briefing: Translational Research newsletter — top stories in biotechnology, drug discovery and pharma.

Get what matters in translational research, free to your inbox weekly. Sign up for Nature Briefing: Translational Research