Integrating concept of pharmacophore with graph neural networks for chemical property prediction and interpretation
Abstract Recently, graph neural networks (GNNs) have revolutionized the field of chemical property prediction and achieved state-of-the-art results on benchmark data sets. Compared with the traditional descriptor- and fingerprint-based QSAR models, GNNs can learn task related representations, which...
| Main Authors: | , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
BMC
2022-08-01
|
| Series: | Journal of Cheminformatics |
| Subjects: | |
| Online Access: | https://doi.org/10.1186/s13321-022-00634-3 |
| _version_ | 1811344077021511680 |
|---|---|
| author | Yue Kong Xiaoman Zhao Ruizi Liu Zhenwu Yang Hongyan Yin Bowen Zhao Jinling Wang Bingjie Qin Aixia Yan |
| author_facet | Yue Kong Xiaoman Zhao Ruizi Liu Zhenwu Yang Hongyan Yin Bowen Zhao Jinling Wang Bingjie Qin Aixia Yan |
| author_sort | Yue Kong |
| collection | DOAJ |
| description | Abstract Recently, graph neural networks (GNNs) have revolutionized the field of chemical property prediction and achieved state-of-the-art results on benchmark data sets. Compared with the traditional descriptor- and fingerprint-based QSAR models, GNNs can learn task related representations, which completely gets rid of the rules defined by experts. However, due to the lack of useful prior knowledge, the prediction performance and interpretability of the GNNs may be affected. In this study, we introduced a new GNN model called RG-MPNN for chemical property prediction that integrated pharmacophore information hierarchically into message-passing neural network (MPNN) architecture, specifically, in the way of pharmacophore-based reduced-graph (RG) pooling. RG-MPNN absorbed not only the information of atoms and bonds from the atom-level message-passing phase, but also the information of pharmacophores from the RG-level message-passing phase. Our experimental results on eleven benchmark and ten kinase data sets showed that our model consistently matched or outperformed other existing GNN models. Furthermore, we demonstrated that applying pharmacophore-based RG pooling to MPNN architecture can generally help GNN models improve the predictive power. The cluster analysis of RG-MPNN representations and the importance analysis of pharmacophore nodes will help chemists gain insights for hit discovery and lead optimization. Graphical Abstract |
| first_indexed | 2024-04-13T19:41:29Z |
| format | Article |
| id | doaj.art-a0dd876f3d714592ba937c065335206e |
| institution | Directory Open Access Journal |
| issn | 1758-2946 |
| language | English |
| last_indexed | 2024-04-13T19:41:29Z |
| publishDate | 2022-08-01 |
| publisher | BMC |
| record_format | Article |
| series | Journal of Cheminformatics |
| spelling | doaj.art-a0dd876f3d714592ba937c065335206e2022-12-22T02:32:54ZengBMCJournal of Cheminformatics1758-29462022-08-0114111410.1186/s13321-022-00634-3Integrating concept of pharmacophore with graph neural networks for chemical property prediction and interpretationYue Kong0Xiaoman Zhao1Ruizi Liu2Zhenwu Yang3Hongyan Yin4Bowen Zhao5Jinling Wang6Bingjie Qin7Aixia Yan8State Key Laboratory of Chemical Resource Engineering, Department of Pharmaceutical Engineering, Beijing University of Chemical TechnologyState Key Laboratory of Chemical Resource Engineering, Department of Pharmaceutical Engineering, Beijing University of Chemical TechnologyState Key Laboratory of Chemical Resource Engineering, Department of Pharmaceutical Engineering, Beijing University of Chemical TechnologyState Key Laboratory of Chemical Resource Engineering, Department of Pharmaceutical Engineering, Beijing University of Chemical TechnologyState Key Laboratory of Chemical Resource Engineering, Department of Pharmaceutical Engineering, Beijing University of Chemical TechnologyHyper-Dimension Insight Pharmaceuticals Ltd. Room 511Hyper-Dimension Insight Pharmaceuticals Ltd. Room 511Hyper-Dimension Insight Pharmaceuticals Ltd. Room 511State Key Laboratory of Chemical Resource Engineering, Department of Pharmaceutical Engineering, Beijing University of Chemical TechnologyAbstract Recently, graph neural networks (GNNs) have revolutionized the field of chemical property prediction and achieved state-of-the-art results on benchmark data sets. Compared with the traditional descriptor- and fingerprint-based QSAR models, GNNs can learn task related representations, which completely gets rid of the rules defined by experts. However, due to the lack of useful prior knowledge, the prediction performance and interpretability of the GNNs may be affected. In this study, we introduced a new GNN model called RG-MPNN for chemical property prediction that integrated pharmacophore information hierarchically into message-passing neural network (MPNN) architecture, specifically, in the way of pharmacophore-based reduced-graph (RG) pooling. RG-MPNN absorbed not only the information of atoms and bonds from the atom-level message-passing phase, but also the information of pharmacophores from the RG-level message-passing phase. Our experimental results on eleven benchmark and ten kinase data sets showed that our model consistently matched or outperformed other existing GNN models. Furthermore, we demonstrated that applying pharmacophore-based RG pooling to MPNN architecture can generally help GNN models improve the predictive power. The cluster analysis of RG-MPNN representations and the importance analysis of pharmacophore nodes will help chemists gain insights for hit discovery and lead optimization. Graphical Abstracthttps://doi.org/10.1186/s13321-022-00634-3Graph neural networks (GNNs)PharmacophoreReduced graph (RG)Hierarchical pooling |
| spellingShingle | Yue Kong Xiaoman Zhao Ruizi Liu Zhenwu Yang Hongyan Yin Bowen Zhao Jinling Wang Bingjie Qin Aixia Yan Integrating concept of pharmacophore with graph neural networks for chemical property prediction and interpretation Journal of Cheminformatics Graph neural networks (GNNs) Pharmacophore Reduced graph (RG) Hierarchical pooling |
| title | Integrating concept of pharmacophore with graph neural networks for chemical property prediction and interpretation |
| title_full | Integrating concept of pharmacophore with graph neural networks for chemical property prediction and interpretation |
| title_fullStr | Integrating concept of pharmacophore with graph neural networks for chemical property prediction and interpretation |
| title_full_unstemmed | Integrating concept of pharmacophore with graph neural networks for chemical property prediction and interpretation |
| title_short | Integrating concept of pharmacophore with graph neural networks for chemical property prediction and interpretation |
| title_sort | integrating concept of pharmacophore with graph neural networks for chemical property prediction and interpretation |
| topic | Graph neural networks (GNNs) Pharmacophore Reduced graph (RG) Hierarchical pooling |
| url | https://doi.org/10.1186/s13321-022-00634-3 |
| work_keys_str_mv | AT yuekong integratingconceptofpharmacophorewithgraphneuralnetworksforchemicalpropertypredictionandinterpretation AT xiaomanzhao integratingconceptofpharmacophorewithgraphneuralnetworksforchemicalpropertypredictionandinterpretation AT ruiziliu integratingconceptofpharmacophorewithgraphneuralnetworksforchemicalpropertypredictionandinterpretation AT zhenwuyang integratingconceptofpharmacophorewithgraphneuralnetworksforchemicalpropertypredictionandinterpretation AT hongyanyin integratingconceptofpharmacophorewithgraphneuralnetworksforchemicalpropertypredictionandinterpretation AT bowenzhao integratingconceptofpharmacophorewithgraphneuralnetworksforchemicalpropertypredictionandinterpretation AT jinlingwang integratingconceptofpharmacophorewithgraphneuralnetworksforchemicalpropertypredictionandinterpretation AT bingjieqin integratingconceptofpharmacophorewithgraphneuralnetworksforchemicalpropertypredictionandinterpretation AT aixiayan integratingconceptofpharmacophorewithgraphneuralnetworksforchemicalpropertypredictionandinterpretation |