Student or educator? Sign up with your school email and get 6 months of Pro free.
Cited by 191

The DOI matches its Crossref record; 10 of the 10 references it cites were found in a registry or an index.

Chemistry-intuitive explanation of graph neural networks for molecular property prediction with substructure masking

Open access
Wu, Zhenxing, Wang, Jike, Du, Hongyan, Jiang, Dejun, Kang, Yu, Li, Dan, Pan, Peichen, Deng, Yafeng, Cao, Dongsheng, Hsieh, Chang-Yu, Hou, Tingjun·2023·Nature Communications

Abstract

Graph neural networks (GNNs) have been widely used in molecular property prediction, but explaining their black-box predictions is still a challenge. Most existing explanation methods for GNNs in chemistry focus on attributing model predictions to individual nodes, edges or fragments that are not necessarily derived from a chemically meaningful segmentation of molecules. To address this challenge, we propose a method named substructure mask explanation (SME). SME is based on well-established molecular segmentation methods and provides an interpretation that aligns with the understanding of chemists. We apply SME to elucidate how GNNs learn to predict aqueous solubility, genotoxicity, cardiotoxicity and blood–brain barrier permeation for small molecules. SME provides interpretation that is consistent with the understanding of chemists, alerts them to unreliable performance, and guides them in structural optimization for target properties. Hence, we believe that SME empowers chemists to confidently mine structure-activity relationship (SAR) from reliable GNNs through a transparent inspection on how GNNs pick up useful signals when learning from data.

Overview

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.1038/s41467-023-38192-3

Cites

10 of 10 references verified0 citation errors · 0 not checkable

Statistics

Verification

Claims the paper makes

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

The self-check needs the paper’s full text.

Citations

Cites

Show all 10 references

Research works printed but not linked: 54

  1. Feinberg, E. N. et al. PotentialNet for molecular property prediction. ACS Cent. Sci. 4, 1520–1530 (2018). · DOI · Paper page
  2. Gupta, M., Lee, H. J., Barden, C. J. & Weaver, D. F. The blood–brain barrier (BBB) score. J. Medicinal Chem. 62, 9824–9836 (2019). · DOI · Paper page
  3. Rankovic, Z. CNS physicochemical property space shaped by a diverse set of molecules with experimentally determined exposure in the mouse brain: miniperspective. J. Medicinal Chem. 60, 5943–5954 (2017). · DOI
  4. Leeson, P. D. & Young, R.J. Molecular property design: does everyone get it?). ACS Publications (2015). · DOI
  5. Polishchuk, P. Interpretation of quantitative structure–activity relationship models: past, present, and future. J. Chem. Inf. Modeling 57, 2618–2639 (2017). · DOI · Paper page
  6. Henderson, R., Clevert, D.-A., & Montanari, F. Improving molecular graph neural network explainability with orthonormalization and induced sparsity. In: International Conference on Machine Learning) PMLR (2021).
  7. Yuan, H., Yu, H., Gui, S. & Ji, S. Explainability in graph neural networks: A taxonomic survey. In: IEEE Transactions on Pattern Analysis and Machine Intelligence (2022). · DOI · Paper page
  8. Ying, Z., Bourgeois, D., You, J., Zitnik, M. & Leskovec, J. Gnnexplainer: generating explanations for graph neural networks. Advances in neural information processing systems 32, (2019).
  9. Vu, M. & Thai, M. T. Pgm-explainer: Probabilistic graphical model explanations for graph neural networks. Adv. neural Inf. Process. Syst. 33, 12225–12235 (2020).
  10. Baldassarre, F. & Azizpour, H. Explainability techniques for graph convolutional networks. arXiv preprint arXiv:190513686 (2019).
  11. Yuan, H., Yu, H., Wang, J., Li, K., & Ji, S. On explainability of graph neural networks via subgraph explorations. In: International Conference on Machine Learning) PMLR (2021). · DOI · Paper page
  12. Pope, P. E., Kolouri, S., Rostami, M., Martin, C. E., & Hoffmann, H. Explainability methods for graph convolutional neural networks. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition) (2019). · DOI · Paper page
  13. Schwarzenberg, R., Hübner, M., Harbecke, D., Alt, C. & Hennig, L. Layerwise Relevance Visualization in Convolutional Text Graph Classifiers. In: Proceedings of the Thirteenth Workshop on Graph-Based Methods for Natural Language Processing (TextGraphs-13)). 58–62 (2019). · DOI
  14. Schnake, T. et al. Higher-order explanations of graph neural networks via relevant walks. In: IEEE transactions on pattern analysis and machine intelligence 44, 7581–7596 (2021). · DOI · Paper page
  15. Huang, Q., Yamada, M., Tian, Y., Singh, D. & Chang, Y. Graphlime: Local interpretable model explanations for graph neural networks. IEEE Transactions on Knowledge and Data Engineering, 1–6 (2022). · DOI · Paper page
  16. Yuan, H., Tang, J., Hu, X. & Ji, S. Xgnn: Towards model-level explanations of graph neural networks. In: Proceedings of the 26th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining) (2020). · DOI · Paper page
  17. Degen, J., Wegscheid-Gerlach, C., Zaliani, A. & Rarey, M. On the Art of Compiling and Using’Drug‐Like’Chemical Fragment Spaces. ChemMedChem: Chem. Enabling Drug Discov. 3, 1503–1507 (2008). · DOI · Paper page
  18. Hu, Y., Stumpfe, D. & Bajorath, Jr. Computational exploration of molecular scaffolds in medicinal chemistry: Miniperspective. J. Medicinal Chem. 59, 4062–4076 (2016). · DOI · Paper page
  19. Bemis, G. W. & Murcko, M. A. The properties of known drugs. 1. Molecular frameworks. J. Medicinal Chem. 39, 2887–2893 (1996). · DOI · Paper page
  20. Nigam, A., Pollice, R., Krenn, M., dos Passos Gomes, G. & Aspuru-Guzik, A. Beyond generative models: superfast traversal, optimization, novelty, exploration and discovery (STONED) algorithm for molecules using SELFIES. Chem. Sci. 12, 7079–7090 (2021). · DOI · Paper page
  21. Gandhi, H. A. & White A. D. Explaining molecular properties with natural language. (2022). · DOI
  22. Jamieson, C., Moir, E. M., Rankovic, Z. & Wishart, G. Medicinal chemistry of hERG optimizations: highlights and hang-ups. J. Medicinal Chem. 49, 5029–5046 (2006). · DOI · Paper page
  23. Garrido, A., Lepailleur, A., Mignani, S. M., Dallemagne, P. & Rochais, C. hERG toxicity assessment: Useful guidelines for drug design. Eur. J. Medicinal Chem. 195, 112290 (2020). · DOI · Paper page
  24. Kalyaanamoorthy, S. & Barakat, K. H. Development of safe drugs: the hERG challenge. Medicinal Res. Rev. 38, 525–555 (2018). · DOI · Paper page
  25. Zhou, S., Wang, J. & Liu, H. Lead compound optimization strategy (5)–reducing the hERG cardiac toxicity in drug development. Yao xue xue bao= Acta Pharmaceutica Sin. 51, 1530–1539 (2016).
  26. Vaz, R. J. et al. Design of bivalent ligands using hydrogen bond linkers: synthesis and evaluation of inhibitors for human β-tryptase. Bioorg. Medicinal Chem. Lett. 14, 6053–6056 (2004). · doi:10.1016/j.bmcl.2004.09.065
  27. Hankosky, E. R. et al. Fluoroethoxy-1, 4-diphenethylpiperidine and piperazine derivatives: Potent and selective inhibitors of [3H] dopamine uptake at the vesicular monoamine transporter-2. Bioorg. Medicinal Chem. Lett. 27, 5467–5472 (2017). · DOI
  28. Shu, M. et al. Antagonists of human CCR5 receptor containing 4-(pyrazolyl) piperidine side chains. Part 3: SAR studies on the benzylpyrazole segment. Bioorg. Medicinal Chem. Lett. 14, 947–952 (2004). · DOI
  29. McCauley, J. A. et al. NR2B-selective N-methyl-D-aspartate antagonists: synthesis and evaluation of 5-substituted benzimidazoles. J. Medicinal Chem. 47, 2089–2096 (2004). · DOI
  30. Mittal, A. et al. Artificial intelligence uncovers carcinogenic human metabolites. Nat. Chem. Biol. 18, 1204–1213 (2022). · DOI
  31. Ishikawa, M. & Hashimoto, Y. Improvement in aqueous solubility in small molecule drug discovery programs by disruption of molecular planarity and symmetry. J. Medicinal Chem. 54, 1539–1554 (2011). · DOI · Paper page
  32. Delaney, J. S. ESOL: estimating aqueous solubility directly from molecular structure. J. Chem. Inf. computer Sci. 44, 1000–1005 (2004). · DOI · Paper page
  33. Wang, J. & Hou, T. Recent advances on aqueous solubility prediction. Combinatorial Chem. high. throughput Screen. 14, 328–338 (2011). · DOI
  34. Tang, B. et al. A self-attention based message passing neural network for predicting molecular lipophilicity and aqueous solubility. J. Cheminformatics 12, 1–9 (2020). · DOI · Paper page
  35. Lusci, A., Pollastri, G. & Baldi, P. Deep architectures and deep learning in chemoinformatics: the prediction of aqueous solubility for drug-like molecules. J. Chem. Inf. Modeling 53, 1563–1575 (2013). · DOI · Paper page
  36. Savjani, K. T., Gajjar, A. K. & Savjani, J. K. Drug solubility: importance and enhancement techniques. International Scholarly Research Notices 2012 (2012). · DOI · Paper page
  37. Wu, Z. et al. Mining toxicity information from large amounts of toxicity data. J. Medicinal Chem. 64, 6924–6936 (2021). · DOI
  38. Bakhtyari, N. G., Raitano, G., Benfenati, E., Martin, T. & Young, D. Comparison of in silico models for prediction of mutagenicity. J. Environ. Sci. Health, Part C. 31, 45–66 (2013). · doi:10.1080/10590501.2013.763576
  39. Xu, C. et al. In silico prediction of chemical Ames mutagenicity. J. Chem. Inf. Modeling 52, 2840–2847 (2012). · DOI · Paper page
  40. Hansen, K. et al. Benchmark data set for in silico prediction of Ames mutagenicity. J. Chem. Inf. Modeling 49, 2077–2081 (2009). · DOI · Paper page
  41. Polishchuk, P. G., Kuz’min, V. E., Artemenko, A. G. & Muratov, E. N. Universal approach for structural interpretation of QSAR/QSPR models. Mol. Inform. 32, 843–853 (2013). · DOI · Paper page
  42. Kazius, J., McGuire, R. & Bursi, R. Derivation and validation of toxicophores for mutagenicity prediction. J. medicinal Chem. 48, 312–320 (2005). · DOI · Paper page
  43. Shamovsky, I. et al. Mechanism-based insights into removing the mutagenicity of aromatic amines by small structural alterations. J. Medicinal Chem. 64, 8545–8563 (2021). · doi:10.1021/acs.jmedchem.1c00514
  44. Wang, S., Li, Y., Xu, L., Li, D. & Hou, T. Recent developments in computational prediction of HERG blockage. Curr. Top. Medicinal Chem. 13, 1317–1326 (2013). · DOI
  45. Laverty, H. et al. How can we improve our understanding of cardiovascular safety liabilities to develop safer medicines? Br. J. Pharmacol. 163, 675–693 (2011). · DOI · Paper page
  46. Jing, Y., Easter, A., Peters, D., Kim, N. & Enyedy, I. J. In silico prediction of hERG inhibition. Future Medicinal Chem. 7, 571–586 (2015). · doi:10.4155/fmc.15.18
  47. Braga, R. C. et al. Pred‐hERG: a novel web‐accessible computational tool for predicting cardiac toxicity. Mol. Inform. 34, 698–701 (2015). · DOI · Paper page
  48. Ryu, J. Y., Lee, M. Y., Lee, J. H., Lee, B. H. & Oh, K.-S. DeepHIT: a deep learning framework for prediction of hERG-induced cardiotoxicity. Bioinformatics 36, 3049–3055 (2020). · DOI
  49. Tong, X. et al. Blood–brain barrier penetration prediction enhanced by uncertainty estimation. J. Cheminformatics 14, 1–15 (2022). · DOI
  50. Sakiyama, H., Fukuda, M. & Okuno, T. Prediction of blood-brain barrier penetration (bbbp) based on molecular descriptors of the free-form and in-blood-form datasets. Molecules 26, 7428 (2021). · DOI
  51. Kearnes, S., McCloskey, K., Berndl, M., Pande, V. & Riley, P. Molecular graph convolutions: moving beyond fingerprints. J. Computer-aided Mol. Des. 30, 595–608 (2016). · DOI · Paper page
  52. Veličković, P. et al. Graph Attention Networks. In: International Conference on Learning Representations (2018).
  53. Xiong, Z. et al. Pushing the boundaries of molecular representation for drug discovery with the graph attention mechanism. J. Medicinal Chem. 63, 8749–8760 (2019). · DOI · Paper page
  54. Schlichtkrull, M., Kipf, T. N., Bloem, P., Berg, Rvd., Titov, I., & Welling, M. Modeling relational data with graph convolutional networks. In: European semantic web conference) Springer (2018). · DOI · Paper page

Cited by

See all 191

Metadata

Authors
Wu, Z., Wang, J., Du, H., Jiang, D., Kang, Y., Li, D., Pan, P., Deng, Y., Cao, D., Hsieh, C.-Y., & Hou, T.
Venue
Nature Communications
Published
2023-05-04
Volume
14
Issue
1
ISSN
2041-1723 (ISSN-L)
OpenAlex
W4372330430
Cited by
191
References
10
FWCI
25.14
Type
journal article
Language
en
Open access
gold
Licence
Funders
National Natural Science Foundation of China (22220102001)

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

How to cite

Cite CoChat for the checks on this page, and the paper for its own findings.

This page
CoChat. (2026). Citation check: Chemistry-intuitive explanation of graph neural networks for molecular property prediction with substructure masking. https://app.cochat.ai/papers/10.1038/s41467-023-38192-3
The paper
Wu, Z., Wang, J., Du, H., Jiang, D., Kang, Y., Li, D., Pan, P., Deng, Y., Cao, D., Hsieh, C.-Y., & Hou, T. (2023). Chemistry-intuitive explanation of graph neural networks for molecular property prediction with substructure masking. Nature Communications. https://doi.org/10.1038/s41467-023-38192-3

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.