A self-attention based message passing neural network for predicting molecular lipophilicity and aqueous solubility
Abstract Efficient and accurate prediction of molecular properties, such as lipophilicity and solubility, is highly desirable for rational compound design in chemical and pharmaceutical industries. To this end, we build and apply a graph-neural-network framework called self-attention-based message-p...
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
BMC
2020-02-01
|
| Series: | Journal of Cheminformatics |
| Subjects: | |
| Online Access: | http://link.springer.com/article/10.1186/s13321-020-0414-z |
| _version_ | 1818031042774171648 |
|---|---|
| author | Bowen Tang Skyler T. Kramer Meijuan Fang Yingkun Qiu Zhen Wu Dong Xu |
| author_facet | Bowen Tang Skyler T. Kramer Meijuan Fang Yingkun Qiu Zhen Wu Dong Xu |
| author_sort | Bowen Tang |
| collection | DOAJ |
| description | Abstract Efficient and accurate prediction of molecular properties, such as lipophilicity and solubility, is highly desirable for rational compound design in chemical and pharmaceutical industries. To this end, we build and apply a graph-neural-network framework called self-attention-based message-passing neural network (SAMPN) to study the relationship between chemical properties and structures in an interpretable way. The main advantages of SAMPN are that it directly uses chemical graphs and breaks the black-box mold of many machine/deep learning methods. Specifically, its attention mechanism indicates the degree to which each atom of the molecule contributes to the property of interest, and these results are easily visualized. Further, SAMPN outperforms random forests and the deep learning framework MPN from Deepchem. In addition, another formulation of SAMPN (Multi-SAMPN) can simultaneously predict multiple chemical properties with higher accuracy and efficiency than other models that predict one specific chemical property. Moreover, SAMPN can generate chemically visible and interpretable results, which can help researchers discover new pharmaceuticals and materials. The source code of the SAMPN prediction pipeline is freely available at Github ( https://github.com/tbwxmu/SAMPN ). |
| first_indexed | 2024-12-10T05:45:11Z |
| format | Article |
| id | doaj.art-2e341dc48d624ee0b01f065af9d41781 |
| institution | Directory Open Access Journal |
| issn | 1758-2946 |
| language | English |
| last_indexed | 2024-12-10T05:45:11Z |
| publishDate | 2020-02-01 |
| publisher | BMC |
| record_format | Article |
| series | Journal of Cheminformatics |
| spelling | doaj.art-2e341dc48d624ee0b01f065af9d417812022-12-22T02:00:10ZengBMCJournal of Cheminformatics1758-29462020-02-011211910.1186/s13321-020-0414-zA self-attention based message passing neural network for predicting molecular lipophilicity and aqueous solubilityBowen Tang0Skyler T. Kramer1Meijuan Fang2Yingkun Qiu3Zhen Wu4Dong Xu5Fujian Provincial Key Laboratory of Innovative Drug Target Research, School of Pharmaceutical Sciences, Xiamen UniversityDepartment of Electrical Engineering and Computer Science, Informatics Institute, and Christopher S. Bond Life Sciences Center, University of MissouriFujian Provincial Key Laboratory of Innovative Drug Target Research, School of Pharmaceutical Sciences, Xiamen UniversityFujian Provincial Key Laboratory of Innovative Drug Target Research, School of Pharmaceutical Sciences, Xiamen UniversityFujian Provincial Key Laboratory of Innovative Drug Target Research, School of Pharmaceutical Sciences, Xiamen UniversityDepartment of Electrical Engineering and Computer Science, Informatics Institute, and Christopher S. Bond Life Sciences Center, University of MissouriAbstract Efficient and accurate prediction of molecular properties, such as lipophilicity and solubility, is highly desirable for rational compound design in chemical and pharmaceutical industries. To this end, we build and apply a graph-neural-network framework called self-attention-based message-passing neural network (SAMPN) to study the relationship between chemical properties and structures in an interpretable way. The main advantages of SAMPN are that it directly uses chemical graphs and breaks the black-box mold of many machine/deep learning methods. Specifically, its attention mechanism indicates the degree to which each atom of the molecule contributes to the property of interest, and these results are easily visualized. Further, SAMPN outperforms random forests and the deep learning framework MPN from Deepchem. In addition, another formulation of SAMPN (Multi-SAMPN) can simultaneously predict multiple chemical properties with higher accuracy and efficiency than other models that predict one specific chemical property. Moreover, SAMPN can generate chemically visible and interpretable results, which can help researchers discover new pharmaceuticals and materials. The source code of the SAMPN prediction pipeline is freely available at Github ( https://github.com/tbwxmu/SAMPN ).http://link.springer.com/article/10.1186/s13321-020-0414-zMessage passing networkAttention mechanismDeep learningLipophilicityAqueous solubility |
| spellingShingle | Bowen Tang Skyler T. Kramer Meijuan Fang Yingkun Qiu Zhen Wu Dong Xu A self-attention based message passing neural network for predicting molecular lipophilicity and aqueous solubility Journal of Cheminformatics Message passing network Attention mechanism Deep learning Lipophilicity Aqueous solubility |
| title | A self-attention based message passing neural network for predicting molecular lipophilicity and aqueous solubility |
| title_full | A self-attention based message passing neural network for predicting molecular lipophilicity and aqueous solubility |
| title_fullStr | A self-attention based message passing neural network for predicting molecular lipophilicity and aqueous solubility |
| title_full_unstemmed | A self-attention based message passing neural network for predicting molecular lipophilicity and aqueous solubility |
| title_short | A self-attention based message passing neural network for predicting molecular lipophilicity and aqueous solubility |
| title_sort | self attention based message passing neural network for predicting molecular lipophilicity and aqueous solubility |
| topic | Message passing network Attention mechanism Deep learning Lipophilicity Aqueous solubility |
| url | http://link.springer.com/article/10.1186/s13321-020-0414-z |
| work_keys_str_mv | AT bowentang aselfattentionbasedmessagepassingneuralnetworkforpredictingmolecularlipophilicityandaqueoussolubility AT skylertkramer aselfattentionbasedmessagepassingneuralnetworkforpredictingmolecularlipophilicityandaqueoussolubility AT meijuanfang aselfattentionbasedmessagepassingneuralnetworkforpredictingmolecularlipophilicityandaqueoussolubility AT yingkunqiu aselfattentionbasedmessagepassingneuralnetworkforpredictingmolecularlipophilicityandaqueoussolubility AT zhenwu aselfattentionbasedmessagepassingneuralnetworkforpredictingmolecularlipophilicityandaqueoussolubility AT dongxu aselfattentionbasedmessagepassingneuralnetworkforpredictingmolecularlipophilicityandaqueoussolubility AT bowentang selfattentionbasedmessagepassingneuralnetworkforpredictingmolecularlipophilicityandaqueoussolubility AT skylertkramer selfattentionbasedmessagepassingneuralnetworkforpredictingmolecularlipophilicityandaqueoussolubility AT meijuanfang selfattentionbasedmessagepassingneuralnetworkforpredictingmolecularlipophilicityandaqueoussolubility AT yingkunqiu selfattentionbasedmessagepassingneuralnetworkforpredictingmolecularlipophilicityandaqueoussolubility AT zhenwu selfattentionbasedmessagepassingneuralnetworkforpredictingmolecularlipophilicityandaqueoussolubility AT dongxu selfattentionbasedmessagepassingneuralnetworkforpredictingmolecularlipophilicityandaqueoussolubility |