Prediction of hERG K+ channel blockage using deep neural networks

• DOI: 10.1111/cbdd.13600
• Publication Date: 2019-08-08T17:58:56Z
• Authors: Yanmin Zhang; Junnan Zhao; Yuchen Wang; Yuanrong Fan; Lu Zhu; Yan Yang; Xingye Chen; Tao Lu; Yadong Chen; Haichun Liu
• Keywords: hERG

Abstract Human ether‐a‐go‐go‐related gene (hERG) K+ channel blockage may cause severe cardiac side‐effects and has become a serious issue in safety evaluation of drug candidates. Therefore, improving the ability to avoid undesirable hERG activity early stage discovery is significant importance. The purpose this study was build predictive models by deep neural networks. For each combination sampling methods descriptors, networks with different architectures were implemented classification models. optimal model M15 three hidden layers, undersampling method, 2D descriptors yielded prediction accuracy 0.78 F1 score 0.75 on test set as well 0.77 0.34 external validation set, outperforming other 35 including 9 random forest Particularly, achieved highest second when compared five from four groups using machine learning algorithms same set. It can be believed that powerful capability toxicity, which great benefit for developing novel