Predicting molecular properties is essential for advancing for advancing drug discovery and design. Recently, Graph Neural Networks (GNNs) have gained prominence due to their ability to capture the complex structural and relational information inherent in molecular graphs. Despite their effectiveness, the “black-box” nature of GNNs remains a significant obstacle to their widespread adoption in chemistry, as it hinders interpretability and trust. In this context, several explanation methods based on factual reasoning have emerged. These methods aim to interpret the predictions made by GNNs by analyzing the key features contributing to the prediction. However, these approaches fail to answer critical questions: “How to ensure that the structure-property mapping learned by GNNs is consistent with established domain knowledge”. In this paper, we propose MMGCF, a novel counterfactual explanation framework designed specifically for the prediction of GNN-based molecular properties. MMGCF constructs a hierarchical tree structure on molecular motifs, enabling the systematic generation of counterfactuals through motif perturbations. This framework identifies causally significant motifs and elucidates their impact on model predictions, offering insights into the relationship between structural modifications and predicted properties. Our method demonstrates its effectiveness through comprehensive quantitative and qualitative evaluations of four real-world molecular datasets.
References
[1]
Gao, H., Wang, Z. and Ji, S. (2018) Large-Scale Learnable Graph Convolutional Networks. Proceedings of the 24th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining, London, 19-23 August 2018, 1416-1424.
https://doi.org/10.1145/3219819.3219947
[2]
Liu, Q., Tan, H.S., Zhang, Y.M. and Wang, G.Y. (2022) Dynamic Heterogeneous Network Representation Method Based on Meta-Path. Acta Electronica Sinica, 50, 1830-1839. https://doi.org/10.12263/DZXB.20211288
[3]
Wasielewski, M.R., Forbes, M.D.E., Frank, N.L., Kowalski, K., Scholes, G.D., Yuen-Zhou, J., et al. (2020) Exploiting Chemistry and Molecular Systems for Quantum Information Science. Nature Reviews Chemistry, 4, 490-504.
https://doi.org/10.1038/s41570-020-0200-5
[4]
Wu, W., Zhu, J., Yao, Y. and Lan, Y. (2024) Can Molecular Quantum Computing Bridge Quantum Biology and Cognitive Science? Intelligent Computing, 3, Article ID: 0072. https://doi.org/10.34133/icomputing.0072
[5]
Fang, X., Liu, L., Lei, J., He, D., Zhang, S., Zhou, J., et al. (2022) Geometry-enhanced Molecular Representation Learning for Property Prediction. Nature Machine Intelligence, 4, 127-134. https://doi.org/10.1038/s42256-021-00438-4
[6]
Gong, X., Liu, M., Sun, H., Li, M. and Liu, Q. (2022) HS-DTI: Drug-Target Interaction Prediction Based on Hierarchical Networks and Multi-Order Sequence Effect. 2022 IEEE International Conference on Bioinformatics and Biomedicine (BIBM), Las Vegas, 6-8 December 2022, 322-327.
https://doi.org/10.1109/bibm55620.2022.9994908
[7]
Bongini, P., Bianchini, M. and Scarselli, F. (2021) Molecular Generative Graph Neural Networks for Drug Discovery. Neurocomputing, 450, 242-252.
https://doi.org/10.1016/j.neucom.2021.04.039
[8]
Jiménez-Luna, J., Grisoni, F., Weskamp, N. and Schneider, G. (2021) Artificial Intelligence in Drug Discovery: Recent Advances and Future Perspectives. Expert Opinion on Drug Discovery, 16, 949-959. https://doi.org/10.1080/17460441.2021.1909567
[9]
Wu, Z., Pan, S., Chen, F., Long, G., Zhang, C. and Yu, P.S. (2021) A Comprehensive Survey on Graph Neural Networks. IEEE Transactions on Neural Networks and Learning Systems, 32, 4-24. https://doi.org/10.1109/tnnls.2020.2978386
[10]
Rao, J., Zheng, S., Lu, Y. and Yang, Y. (2022) Quantitative Evaluation of Explainable Graph Neural Networks for Molecular Property Prediction. Patterns, 3, Article ID: 100628. https://doi.org/10.1016/j.patter.2022.100628
[11]
Luo, D., Cheng, W., Xu, D., Yu, W., Zong, B., et al. (2020) Parameterized Explainer for Graph Neural Network. Advances in Neural Information Processing Systems, 33, 19620-19631.
[12]
Huang, Q., Yamada, M., Tian, Y., Singh, D. and Chang, Y. (2023) GraphLIME: Local Interpretable Model Explanations for Graph Neural Networks. IEEE Transactions on Knowledge and Data Engineering, 35, 6968-6972.
https://doi.org/10.1109/tkde.2022.3187455
[13]
Yuan, H., Yu, H., Wang, J., Li, K. and Ji, S. (2021) On Explainability of Graph Neural Networks via Subgraph Explorations. ReScience, 9, Article 41.
Kipf, T.N. and Welling, M. (2016) Semi-Supervised Classification with Graph Convolutional Networks. arXiv: 1609.02907.
[16]
Velickovic, P., Cucurull, G., Casanova, A., et al. (2017) Graph Attention Networks.
https://arxiv.org/abs/1710.10903
[17]
Hamilton, W., Ying, Z. and Leskovec, J. (2017) Inductive Representation Learning on Large Graphs. 31st Conference on Neural Information Processing Systems (NIPS 2017), Long Beach, 4-7 December 2017, 30 p.
[18]
Xu, K., Hu, W., Leskovec, J. and Jegelka, S. (2018) How Powerful Are Graph Neural Networks? arXiv: 1810.00826.
[19]
Schlichtkrull, M., Kipf, T.N., Bloem, P., van den Berg, R., Titov, I. and Welling, M. (2018) Modeling Relational Data with Graph Convolutional Networks. In: Gangemi, A., et al., Eds., The Semantic Web, Springer, 593-607.
https://doi.org/10.1007/978-3-319-93417-4_38
[20]
Baldassarre, F. and Azizpour, H. (2019) Explainability Techniques for Graph Convolutional Networks. arXiv: 1905.13686.
[21]
Pope, P.E., Kolouri, S., Rostami, M., Martin, C.E. and Hoffmann, H. (2019) Explainability Methods for Graph Convolutional Neural Networks. 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Long Beach, 15-20 June 2019, 10764-10773. https://doi.org/10.1109/cvpr.2019.01103
[22]
Ying, Z., Bourgeois, D., You, J., Zitnik, M. and Leskovec, J. (2019) GNNExplainer: Generating Explanations for Graph Neural Networks. 33rd Conference on Neural Information Processing Systems (NeurIPS 2019), Vancouver, 8-14 December 2019, 32 p.
[23]
Wang, X., Wu, Y., Zhang, A., et al. (2021) Towards Multi-Grained Explainability for Graph Neural Networks. Advances in Neural Information Processing Systems, 34, 18446-18458.
[24]
Duval, A. and Malliaros, F.D. (2021) Graphsvx: Shapley Value Explanations for Graph Neural Networks. In: Oliver, N., Pérez-Cruz, F., Kramer, S., Read, J. and Lozano, J.A., Eds., Machine Learning and Knowledge Discovery in Databases, Springer, 302-318. https://doi.org/10.1007/978-3-030-86520-7_19
[25]
Schnake, T., Eberle, O., Lederer, J., Nakajima, S., Schutt, K.T., Muller, K., et al. (2022) Higher-order Explanations of Graph Neural Networks via Relevant Walks. IEEE Transactions on Pattern Analysis and Machine Intelligence, 44, 7581-7596.
https://doi.org/10.1109/tpami.2021.3115452
[26]
Yuan, H., Tang, J., Hu, X. and Ji, S. (2020) XGNN: Towards Model-Level Explanations of Graph Neural Networks. Proceedings of the 26th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining, 6-10 July 2020, 430-438.
https://doi.org/10.1145/3394486.3403085
[27]
Wang, X. and Shen, H.W. (2022) Gnninterpreter: A Probabilistic Generative Model-Level Explanation for Graph Neural Networks. arXiv: 2209.07924.
[28]
Kahneman, D. and Miller, D.T. (1986) Norm Theory: Comparing Reality to Its Alternatives. Psychological Review, 93, 136-153.
https://doi.org/10.1037//0033-295x.93.2.136
[29]
Woodward, J. and Hitchcock, C. (2003) Explanatory Generalizations, Part I: A Counterfactual Account. Noûs, 37, 1-24. https://doi.org/10.1111/1468-0068.00426
[30]
Pearl, J. (2009) Causality. 2nd Edition, Cambridge University Press.
https://doi.org/10.1017/cbo9780511803161
[31]
Lucic, A., Ter Hoeve, M.A., Tolomei, G., et al. (2022) CF-GNNExplainer: Counterfactual Explanations for Graph Neural Networks. International Conference on Artificial Intelligence and Statistics, Valencia, 28-30 March 2022, 4499-4511
[32]
Tan, J., Geng, S., Fu, Z., Ge, Y., Xu, S., Li, Y., et al. (2022) Learning and Evaluating Graph Neural Network Explanations Based on Counterfactual and Factual Reasoning. Proceedings of the ACM Web Conference 2022, Lyon, 25-29 April 2022, 1018-1027. https://doi.org/10.1145/3485447.3511948
[33]
Lin, W., Lan, H. and Li, B. (2021) Generative Causal Explanations for Graph Neural Networks. International Conference on Machine Learning, 18-24 July 2021, 6666-6679.
[34]
Numeroso, D. and Bacciu, D. (2021) MEG: Generating Molecular Counterfactual Explanations for Deep Graph Networks. 2021 International Joint Conference on Neural Networks (IJCNN), Shenzhen, 18-22 July 2021, 1-8.
https://doi.org/10.1109/ijcnn52387.2021.9534266
[35]
Sun, L., Dou, Y., Yang, C., Zhang, K., Wang, J., Yu, P.S., et al. (2022) Adversarial Attack and Defense on Graph Data: A Survey. IEEE Transactions on Knowledge and Data Engineering, 35, 7693-7711. https://doi.org/10.1109/tkde.2022.3201243
[36]
Weininger, D. (1988) SMILES, a Chemical Language and Information System. 1. Introduction to Methodology and Encoding Rules. Journal of Chemical Information and Computer Sciences, 28, 31-36. https://doi.org/10.1021/ci00057a005
[37]
Bento, A.P., Hersey, A., Félix, E., Landrum, G., Gaulton, A., Atkinson, F., et al. (2020) An Open Source Chemical Structure Curation Pipeline Using RDKit. Journal of Cheminformatics, 12, Article No. 51. https://doi.org/10.1186/s13321-020-00456-1
[38]
Jin, W., Barzilay, R. and Jaakkola, T. (2018) Junction Tree Variational Autoencoder for Molecular Graph Generation. International Conference on Machine Learning, Stockholm, 10-15 July 2018, 2323-2332.
[39]
Zhang, Z., Liu, Q., et al. (2021) Motif-Based Graph Self-Supervised Learning for Molecular Property Prediction. Advances in Neural Information Processing Systems, 34, 15870-15882.
[40]
Degen, J., Wegscheid‐Gerlach, C., Zaliani, A. and Rarey, M. (2008) On the Art of Compiling and Using ‘Drug-Like’ Chemical Fragment Spaces. ChemMedChem, 3, 1503-1507. https://doi.org/10.1002/cmdc.200800178
[41]
Ho, J., Chen, X., Srinivas, A., Duan, Y. and Abbeel, P. (2019) Flow++: Improving Flow-Based Generative Models with Variational Dequantization and Architecture Design. International Conference on Machine Learning, Long Beach, 10-15 June 2019, 2722-2730.
[42]
Shi, C., Xu, M., Zhu, Z., et al. (2020) Graphaf: A Flow-Based Autoregressive Model for Molecular Graph Generation. arXiv: 2001.09382.
[43]
Zhu, Y., Ouyang, Z., Liao, B., Wu, J., Wu, Y., Hsieh, C., et al. (2023) MolHF: A Hierarchical Normalizing Flow for Molecular Graph Generation. Proceedings of the Thirty-Second International Joint Conference on Artificial Intelligence, Macao, 19-25 August 2023, 5002-5010. https://doi.org/10.24963/ijcai.2023/556
[44]
Kazius, J., McGuire, R. and Bursi, R. (2004) Derivation and Validation of Toxicophores for Mutagenicity Prediction. Journal of Medicinal Chemistry, 48, 312-320.
https://doi.org/10.1021/jm040835a
[45]
Wu, Z., Wang, J., Du, H., Jiang, D., Kang, Y., Li, D., et al. (2023) Chemistry-intuitive Explanation of Graph Neural Networks for Molecular Property Prediction with Substructure Masking. Nature Communications, 14, Article No. 2585.
https://doi.org/10.1038/s41467-023-38192-3
[46]
Sakiyama, H., Fukuda, M. and Okuno, T. (2021) Prediction of Blood-Brain Barrier Penetration (BBBP) Based on Molecular Descriptors of the Free-Form and In-Blood-Form Datasets. Molecules, 26, Article 7428.
https://doi.org/10.3390/molecules26247428
[47]
Wu, Z., Ramsundar, B., Feinberg, E.N., Gomes, J., Geniesse, C., Pappu, A.S., et al. (2018) MoleculeNet: A Benchmark for Molecular Machine Learning. Chemical Science, 9, 513-530. https://doi.org/10.1039/c7sc02664a
[48]
Ramsundar, B., Eastman, P., et al. (2019) Deep Learning for the Life Sciences: Applying Deep Learning to Genomics, Microscopy, Drug Discovery, and More. O’Reilly Media, Inc.
[49]
Ma, J., Guo, R., Mishra, S., Zhang, A. and Li, J. (2022) Clear: Generative Counterfactual Explanations on Graphs. Advances in Neural Information Processing Systems, 35, 25895-25907.
[50]
Prado-Romero, M.A., Prenkaj, B., Stilo, G. and Giannotti, F. (2024) A Survey on Graph Counterfactual Explanations: Definitions, Methods, Evaluation, and Research Challenges. ACM Computing Surveys, 56, 1-37. https://doi.org/10.1145/3618105
