Could graph neural networks learn better molecular representation for drug discovery? A comparison study of descriptor-based and graph-based models

Journal of Cheminformatics

Jian Wu¹  Zhenxing Wu²  Chao Shen²  Zhe Wang²  Tingjun Hou³  Dejun Jiang⁴  Guangyong Chen⁵  Chang-Yu Hsieh⁶  Ben Liao⁶  Dongsheng Cao⁷ 

[1]College of Computer Science and Technology, Zhejiang University, Hangzhou, China

[2]Innovation Institute for Artificial Intelligence in Medicine of Zhejiang University, College of Pharmaceutical Sciences, Zhejiang University, 310058, Hangzhou, Zhejiang, China

[3]Innovation Institute for Artificial Intelligence in Medicine of Zhejiang University, College of Pharmaceutical Sciences, Zhejiang University, 310058, Hangzhou, Zhejiang, China

[4]State Key Lab of CAD & CG, Zhejiang University, 310058, Hangzhou, Zhejiang, China

[5]Innovation Institute for Artificial Intelligence in Medicine of Zhejiang University, College of Pharmaceutical Sciences, Zhejiang University, 310058, Hangzhou, Zhejiang, China

[6]State Key Lab of CAD & CG, Zhejiang University, 310058, Hangzhou, Zhejiang, China

[7]College of Computer Science and Technology, Zhejiang University, Hangzhou, China

[8]Shenzhen Institutes of Advanced Technology, 518055, Shenzhen, Guangdong, China

[9]Tencent Quantum Laboratory Tencent, 518057, Shenzhen, Guangdong, China

[10]Xiangya School of Pharmaceutical Sciences, Central South University, 410004, Changsha, Hunan, China

关键词: Graph neural networks;    Extreme gradient boosting;    Ensemble learning;    Deep learning;    ADME/T prediction;   

DOI  :  10.1186/s13321-020-00479-8

来源: Springer

【 摘 要 】

Graph neural networks (GNN) has been considered as an attractive modelling method for molecular property prediction, and numerous studies have shown that GNN could yield more promising results than traditional descriptor-based methods. In this study, based on 11 public datasets covering various property endpoints, the predictive capacity and computational efficiency of the prediction models developed by eight machine learning (ML) algorithms, including four descriptor-based models (SVM, XGBoost, RF and DNN) and four graph-based models (GCN, GAT, MPNN and Attentive FP), were extensively tested and compared. The results demonstrate that on average the descriptor-based models outperform the graph-based models in terms of prediction accuracy and computational efficiency. SVM generally achieves the best predictions for the regression tasks. Both RF and XGBoost can achieve reliable predictions for the classification tasks, and some of the graph-based models, such as Attentive FP and GCN, can yield outstanding performance for a fraction of larger or multi-task datasets. In terms of computational cost, XGBoost and RF are the two most efficient algorithms and only need a few seconds to train a model even for a large dataset. The model interpretations by the SHAP method can effectively explore the established domain knowledge for the descriptor-based models. Finally, we explored use of these models for virtual screening (VS) towards HIV and demonstrated that different ML algorithms offer diverse VS profiles. All in all, we believe that the off-the-shelf descriptor-based models still can be directly employed to accurately predict various chemical endpoints with excellent computability and interpretability.

【 授权许可】

CC BY