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:

A unified pre-trained deep learning framework for cross-task reaction performance prediction and synthesis planning

Nature Machine Intelligence volume 7, pages 1561–1571 (2025)Cite this article

Subjects

A preprint version of the article is available at Research Square.

Abstract

Artificial intelligence has transformed the field of precise organic synthesis. Data-driven methods, including machine learning and deep learning, have shown great promise in predicting reaction performance and synthesis planning. However, the inherent methodological divergence between numerical regression-driven reaction performance prediction and sequence generation-based synthesis planning creates formidable challenges in constructing a unified deep learning architecture. Here we present RXNGraphormer, a framework to jointly address these tasks through a unified pre-training approach. By synergizing graph neural networks for intramolecular pattern recognition with Transformer-based models for intermolecular interaction modelling, and training on 13 million reactions via a carefully designed strategy, RXNGraphormer achieves state-of-the-art performance across eight benchmark datasets for reactivity or selectivity prediction and forward-synthesis or retrosynthesis planning, as well as three external realistic datasets for reactivity and selectivity prediction. Notably, the model generates chemically meaningful embeddings that spontaneously cluster reactions by type without explicit supervision. This work bridges the critical gap between performance prediction and synthesis planning tasks in chemical AI, offering a versatile tool for accurate reaction prediction and synthesis design.

Access through your institution
Buy or subscribe

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

Access options

Access through your institution

Buy this article

  • Purchase on SpringerLink
  • Instant access to full article PDF
Buy now

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

Fig. 1: Deep learning model for reaction prediction.
Fig. 2: Overview of the RXNGraphormer architecture.
Fig. 3: Distance analysis of reaction embeddings from USPTO-50k dataset.
Fig. 4: Regression performance evaluation of RXNGraphormer across four benchmark datasets.
Fig. 5: External validation of RXNGraphormer on literature-derived datasets.

Similar content being viewed by others

Data availability

The preprocessed dataset comprising 13 million chemical reactions used for model pre-training is available via GitHub at https://github.com/licheng-xu-echo/RXNGraphormer. A permanently archived version is accessible through Zenodo at https://doi.org/10.5281/zenodo.15770470 (ref. 47) and via figshare at https://doi.org/10.6084/m9.figshare.28356077 (ref. 48). Specific datasets, including the Buchwald–Hartwig reaction dataset, Suzuki–Miyaura reaction dataset, the C–H functionalization dataset, the asymmetric thiol addition dataset, the USPTO-derived series and three external validation datasets are available via GitHub at https://github.com/licheng-xu-echo/RXNGraphormer.

Code availability

Codes for data preprocessing, fictitious product molecule generation, delta-mol graph generation, model training, performance evaluation, as well as scripts for analysing pre-trained and fine-tuned models, are available via GitHub at https://github.com/licheng-xu-echo/RXNGraphormer. A permanently archived version is accessible through Zenodo at https://doi.org/10.5281/zenodo.15770470 (ref. 47).

References

  1. Żurański, A. M., Martinez Alvarado, J. I., Shields, B. J. & Doyle, A. G. Predicting reaction yields via supervised learning. Acc. Chem. Res. 54, 1856–1865 (2021).

    Article  Google Scholar 

  2. Zahrt, A. F., Athavale, S. V. & Denmark, S. E. Quantitative structure–selectivity relationships in enantioselective catalysis: past, present, and future. Chem. Rev. 120, 1620–1689 (2020).

    Article  Google Scholar 

  3. Corey, E. J., Long, A. K. & Rubenstein, S. D. Computer-assisted analysis in organic synthesis. Science 228, 408–418 (1985).

    Article  Google Scholar 

  4. Todd, M. H. Computer-aided organic synthesis. Chem. Soc. Rev. 34, 247 (2005).

    Article  Google Scholar 

  5. Cheong, P. H.-Y., Legault, C. Y., Um, J. M., Çelebi-Ölçüm, N. & Houk, K. N. Quantum Mechanical investigations of organocatalysis: mechanisms, reactivities, and selectivities. Chem. Rev. 111, 5042–5137 (2011).

    Article  Google Scholar 

  6. Klucznik, T. et al. Efficient syntheses of diverse, medicinally relevant targets planned by computer and executed in the laboratory. Chem 4, 522–532 (2018).

    Article  Google Scholar 

  7. Crawford, J. M., Kingston, C., Toste, F. D. & Sigman, M. S. Data science meets physical organic chemistry. Acc. Chem. Res. 54, 3136–3148 (2021).

    Article  Google Scholar 

  8. Rinehart, N. I., Zahrt, A. F., Henle, J. J. & Denmark, S. E. Dreams, false starts, dead ends, and redemption: a chronicle of the evolution of a chemoinformatic workflow for the optimization of enantioselective catalysts. Acc. Chem. Res. 54, 2041–2054 (2021).

    Article  Google Scholar 

  9. Rinehart, N. I. et al. A machine-learning tool to predict substrate-adaptive conditions for Pd-catalyzed C–N couplings. Science 381, 965–972 (2023).

    Article  Google Scholar 

  10. Zahrt, A. F. et al. Prediction of higher-selectivity catalysts by computer-driven workflow and machine learning. Science 363, eaau5631 (2019).

    Article  Google Scholar 

  11. Reid, J. P. & Sigman, M. S. Holistic prediction of enantioselectivity in asymmetric catalysis. Nature 571, 343–348 (2019).

    Article  Google Scholar 

  12. Ahneman, D. T., Estrada, J. G., Lin, S., Dreher, S. D. & Doyle, A. G. Predicting reaction performance in C–N cross-coupling using machine learning. Science 360, 186–190 (2018).

    Article  Google Scholar 

  13. Sandfort, F., Strieth-Kalthoff, F., Kühnemund, M., Beecks, C. & Glorius, F. A structure-based platform for predicting chemical reactivity. Chem 6, 1379–1390 (2020).

    Article  Google Scholar 

  14. Schwaller, P., Vaucher, A. C., Laino, T. & Reymond, J.-L. Prediction of chemical reaction yields using deep learning. Mach. Learn. Sci. Technol. 2, 015016 (2021).

    Article  Google Scholar 

  15. Li, B. et al. A deep learning framework for accurate reaction prediction and its application on high-throughput experimentation data. J. Cheminformatics 15, 72 (2023).

    Article  Google Scholar 

  16. Coley, C. W., Green, W. H. & Jensen, K. F. Machine learning in computer-aided synthesis planning. Acc. Chem. Res. 51, 1281–1289 (2018).

    Article  Google Scholar 

  17. Segler, M. H. S., Preuss, M. & Waller, M. P. Planning chemical syntheses with deep neural networks and symbolic AI. Nature 555, 604–610 (2018).

    Article  Google Scholar 

  18. Coley, C. W., Barzilay, R., Jaakkola, T. S., Green, W. H. & Jensen, K. F. Prediction of organic reaction outcomes using machine learning. ACS Cent. Sci. 3, 434–443 (2017).

    Article  Google Scholar 

  19. Tu, Z. & Coley, C. W. Permutation invariant graph-to-sequence model for template-free retrosynthesis and reaction prediction. J. Chem. Inf. Model. 62, 3503–3513 (2022).

    Article  Google Scholar 

  20. Schwaller, P. et al. Molecular transformer: a model for uncertainty-calibrated chemical reaction prediction. ACS Cent. Sci. 5, 1572–1583 (2019).

    Article  Google Scholar 

  21. Lin, K., Xu, Y., Pei, J. & Lai, L. Automatic retrosynthetic route planning using template-free models. Chem. Sci. 11, 3355–3364 (2020).

    Article  Google Scholar 

  22. Zheng, S., Rao, J., Zhang, Z., Xu, J. & Yang, Y. Predicting retrosynthetic reactions using self-corrected transformer neural networks. J. Chem. Inf. Model. 60, 47–55 (2020).

    Article  Google Scholar 

  23. Kim, E., Lee, D., Kwon, Y., Park, M. S. & Choi, Y.-S. Valid, plausible, and diverse retrosynthesis using tied two-way transformers with latent variables. J. Chem. Inf. Model. 61, 123–133 (2021).

    Article  Google Scholar 

  24. Sacha, M. et al. Molecule edit graph attention network: modeling chemical reactions as sequences of graph edits. J. Chem. Inf. Model. 61, 3273–3284 (2021).

    Article  Google Scholar 

  25. Mao, K. et al. Molecular graph enhanced transformer for retrosynthesis prediction. Neurocomputing 457, 193–202 (2021).

    Article  Google Scholar 

  26. Irwin, R., Dimitriadis, S., He, J. & Bjerrum, E. J. Chemformer: a pre-trained transformer for computational chemistry. Mach. Learn. Sci. Technol. 3, 015022 (2022).

    Article  Google Scholar 

  27. Zhu, J. et al. Dual-view molecular pre-training. In Proc. 29th ACM SIGKDD Conference on Knowledge Discovery and Data Mining (eds Singh, A. K. et al.) 3615–3627 (ACM, 2023).

  28. Lu, J. & Zhang, Y. Unified deep learning model for multitask reaction predictions with explanation. J. Chem. Inf. Model. 62, 1376–1387 (2022).

    Article  Google Scholar 

  29. Li, S.-W., Xu, L.-C., Zhang, C., Zhang, S.-Q. & Hong, X. Reaction performance prediction with an extrapolative and interpretable graph model based on chemical knowledge. Nat. Commun. 14, 3569 (2023).

    Article  Google Scholar 

  30. Shi, R., Yu, G., Huo, X. & Yang, Y. Prediction of chemical reaction yields with large-scale multi-view pre-training. J. Cheminformatics 16, 22 (2024).

    Article  Google Scholar 

  31. Xia, J. et al. Mole-BERT: rethinking pre-training graph neural networks for molecules. In Proc. Eleventh International Conference on Learning Representations (eds Yan, L. et al.) https://openreview.net/pdf?id=jevY-DtiZTR (ICLR 2023).

  32. Ying, C. et al. Do transformers really perform badly for graph representation? In 35th Conference on Neural Information Processing Systems (NeurIPS 2021) https://openreview.net/pdf?id=OeWooOxFwDa (NeurIPS, 2021).

  33. Shi, Y. et al. Benchmarking Graphormer on large-scale molecular modeling datasets. Preprint at https://arxiv.org/abs/2203.04810 (2023).

  34. Vaswani, A. et al. Attention is all you need. In 31st Conference on Neural Information Processing Systems (NIPS 2017) https://papers.nips.cc/paper_files/paper/2017/file/3f5ee243547dee91fbd053c1c4a845aa-Paper.pdf (2017).

  35. Perera, D. et al. A platform for automated nanomole-scale reaction screening and micromole-scale synthesis in flow. Science 359, 429–434 (2018).

    Article  Google Scholar 

  36. Li, X., Zhang, S., Xu, L. & Hong, X. Predicting regioselectivity in radical C−H functionalization of heterocycles through machine learning. Angew. Chem. Int. Ed. 59, 13253–13259 (2020).

    Article  Google Scholar 

  37. Schneider, N., Stiefl, N. & Landrum, G. A. What’s what: the (nearly) definitive guide to reaction role assignment. J. Chem. Inf. Model. 56, 2336–2346 (2016).

    Article  Google Scholar 

  38. Dai, H., Li, C., Coley, C., Dai, B. & Song, L. Retrosynthesis prediction with conditional graph logic network. In 33rd Conference on Neural Information Processing Systems (NeurIPS 2019) https://proceedings.neurips.cc/paper_files/paper/2019/file/0d2b2061826a5df3221116a5085a6052-Paper.pdf (NeurIPS, 2019).

  39. Schwaller, P., Gaudin, T., Lányi, D., Bekas, C. & Laino, T. ‘Found in translation’: predicting outcomes of complex organic chemistry reactions using neural sequence-to-sequence models. Chem. Sci. 9, 6091–6098 (2018).

    Article  Google Scholar 

  40. Jin, W., Coley, C., Barzilay, R. & Jaakkola, T. Predicting organic reaction outcomes with Weisfeiler-Lehman network. In 31st Conference on Neural Information Processing Systems (NIPS 2017) https://papers.nips.cc/paper_files/paper/2017/file/ced556cd9f9c0c8315cfbe0744a3baf0-Paper.pdf (2017).

  41. Arjovsky, M., Bottou, L., Gulrajani, I. & Lopez-Paz, D. Invariant risk minimization. Preprint at https://arxiv.org/abs/1907.02893 (2020).

  42. Schwaller, P. et al. Mapping the space of chemical reactions using attention-based neural networks. Nat. Mach. Intell. 3, 144–152 (2021).

    Article  Google Scholar 

  43. Schneider, N., Lowe, D. M., Sayle, R. A. & Landrum, G. A. Development of a novel fingerprint for chemical reactions and its application to large-scale reaction classification and similarity. J. Chem. Inf. Model. 55, 39–53 (2015).

    Article  Google Scholar 

  44. Schleinitz, J. et al. Machine learning yield prediction from NiCOlit, a small-size literature data set of nickel catalyzed C–O couplings. J. Am. Chem. Soc. 144, 14722–14730 (2022).

    Article  Google Scholar 

  45. Xu, L. et al. Towards data‐driven design of asymmetric hydrogenation of olefins: database and hierarchical learning. Angew. Chem. Int. Ed. 60, 22804–22811 (2021).

    Article  Google Scholar 

  46. Xu, L.-C. et al. Enantioselectivity prediction of pallada-electrocatalysed C–H activation using transition state knowledge in machine learning. Nat. Synth. 2, 321–330 (2023).

    Article  Google Scholar 

  47. Xu, L.-C., Tang, M.-J., An, J., Cao, F. & Qi, Y. A unified pre-trained deep learning framework for cross-task reaction performance prediction and synthesis planning. Zenodo https://doi.org/10.5281/zenodo.15770470 (2025).

  48. Xu, L.-C., Tang, M.-J., An, J., Cao, F. & Qi, Y. A unified pre-trained deep learning framework for cross-task reaction performance prediction and synthesis planning. figshare https://doi.org/10.6084/m9.figshare.28356077 (2025).

Download references

Acknowledgements

Generous support by the National Natural Science Foundation of China (grant nos. 82394432 and 92249302, Y.Q.) and the Shanghai Municipal Science and Technology Major Project (grant no. 2023SHZDZX02, Y.Q.). The computations in this research were performed using the CFFF platform at Fudan University. The Data Platform Department at the Shanghai Academy of Artificial Intelligence for Science is acknowledged for their contributions in the open-source reaction dataset collection process.

Author information

Authors and Affiliations

  1. Shanghai Academy of Artificial Intelligence for Science, Shanghai, China

    Li-Cheng Xu, Miao-Jiong Tang, Junyi An, Fenglei Cao & Yuan Qi

  2. Center of Chemistry for Frontier Technologies, Department of Chemistry, Zhejiang University, Hangzhou, China

    Miao-Jiong Tang

  3. Artificial Intelligence Innovation and Incubation Institute, Fudan University, Shanghai, China

    Yuan Qi

Authors
  1. Li-Cheng Xu
    View author publications

    Search author on:PubMed Google Scholar

  2. Miao-Jiong Tang
    View author publications

    Search author on:PubMed Google Scholar

  3. Junyi An
    View author publications

    Search author on:PubMed Google Scholar

  4. Fenglei Cao
    View author publications

    Search author on:PubMed Google Scholar

  5. Yuan Qi
    View author publications

    Search author on:PubMed Google Scholar

Contributions

L.-C.X., F.C. and Y.Q. conceived the initial idea for the project. F.C. and Y.Q. supervised the project. L.-C.X. designed the RXNGraphormer model framework and the pre-training strategy, wrote the codes, trained the models and analysed the results. M.-J.T. contributed to implement the algorithm for fictitious reaction generation. J.A. contributed to design the model encoder. All authors took part in discussions. L.-C.X. wrote the paper with input from all the authors.

Corresponding authors

Correspondence to Li-Cheng Xu, Fenglei Cao or Yuan Qi.

Ethics declarations

Competing interests

The authors declare no competing interests.

Peer review

Peer review information

Nature Machine Intelligence thanks Kuangbiao Liao, Jolene Reid 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

Details of model framework, comparison of results with other models and t-SNE embedding analysis results, together with Supplementary Figs. 1–23 and Tables 1–8.

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

Xu, LC., Tang, MJ., An, J. et al. A unified pre-trained deep learning framework for cross-task reaction performance prediction and synthesis planning. Nat Mach Intell 7, 1561–1571 (2025). https://doi.org/10.1038/s42256-025-01098-4

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue date:

  • DOI: https://doi.org/10.1038/s42256-025-01098-4

Search

Advanced search

Quick links

Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing