Predicting cardiotoxicity in drug development: A deep learning approach

Cardiotoxicity is a critical issue in drug development that poses serious health risks, including potentially fatal arrhythmias. The human ether-à-go-go related gene (hERG) potassium channel, as one of the primary targets of cardiotoxicity, has garnered widespread attention. Traditional cardiotoxici...

Full description

Saved in:
Bibliographic Details
Main Authors: Kaifeng Liu, Huizi Cui, Xiangyu Yu, Wannan Li, Weiwei Han
Format: Article
Language:English
Published: Elsevier 2025-08-01
Series:Journal of Pharmaceutical Analysis
Subjects:
Online Access:http://www.sciencedirect.com/science/article/pii/S2095177925000802
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849227043156262912
author Kaifeng Liu
Huizi Cui
Xiangyu Yu
Wannan Li
Weiwei Han
author_facet Kaifeng Liu
Huizi Cui
Xiangyu Yu
Wannan Li
Weiwei Han
author_sort Kaifeng Liu
collection DOAJ
description Cardiotoxicity is a critical issue in drug development that poses serious health risks, including potentially fatal arrhythmias. The human ether-à-go-go related gene (hERG) potassium channel, as one of the primary targets of cardiotoxicity, has garnered widespread attention. Traditional cardiotoxicity testing methods are expensive and time-consuming, making computational virtual screening a suitable alternative. In this study, we employed machine learning techniques utilizing molecular fingerprints and descriptors to predict the cardiotoxicity of compounds, with the aim of improving prediction accuracy and efficiency. We used four types of molecular fingerprints and descriptors combined with machine learning and deep learning algorithms, including Gaussian naive Bayes (NB), random forest (RF), support vector machine (SVM), K-nearest neighbors (KNN), eXtreme gradient boosting (XGBoost), and Transformer models, to build predictive models. Our models demonstrated advanced predictive performance. The best machine learning model, XGBoost Morgan, achieved an accuracy (ACC) value of 0.84, and the deep learning model, Transformer_Morgan, achieved the best ACC value of 0.85, showing a high ability to distinguish between toxic and non-toxic compounds. On an external independent validation set, it achieved the best area under the curve (AUC) value of 0.93, surpassing ADMETlab3.0, Cardpred, and CardioDPi. In addition, we explored the integration of molecular descriptors and fingerprints to enhance model performance and found that ensemble methods, such as voting and stacking, provided slight improvements in model stability. Furthermore, the SHapley Additive exPlanations (SHAP) explanations revealed the relationship between benzene rings, fluorine-containing groups, NH groups, oxygen in ether groups, and cardiotoxicity, highlighting the importance of these features. This study not only improved the predictive accuracy of cardiotoxicity models but also promoted a more reliable and scientifically interpretable method for drug safety assessment. Using computational methods, this study facilitates a more efficient drug development process, reduces costs, and improves the safety of new drug candidates, ultimately benefiting medical and public health.
format Article
id doaj-art-e24ce28dc6834a759d735c727c437949
institution Kabale University
issn 2095-1779
language English
publishDate 2025-08-01
publisher Elsevier
record_format Article
series Journal of Pharmaceutical Analysis
spelling doaj-art-e24ce28dc6834a759d735c727c4379492025-08-24T05:12:04ZengElsevierJournal of Pharmaceutical Analysis2095-17792025-08-0115810126310.1016/j.jpha.2025.101263Predicting cardiotoxicity in drug development: A deep learning approachKaifeng Liu0Huizi Cui1Xiangyu Yu2Wannan Li3Weiwei Han4Key Laboratory for Molecular Enzymology and Engineering of Ministry of Education, Edmond H. Fischer Signal Transduction Laboratory, School of Life Sciences, Jilin University, Changchun, 130012, ChinaKey Laboratory for Molecular Enzymology and Engineering of Ministry of Education, Edmond H. Fischer Signal Transduction Laboratory, School of Life Sciences, Jilin University, Changchun, 130012, ChinaKey Laboratory for Molecular Enzymology and Engineering of Ministry of Education, Edmond H. Fischer Signal Transduction Laboratory, School of Life Sciences, Jilin University, Changchun, 130012, ChinaCorresponding author.; Key Laboratory for Molecular Enzymology and Engineering of Ministry of Education, Edmond H. Fischer Signal Transduction Laboratory, School of Life Sciences, Jilin University, Changchun, 130012, ChinaCorresponding author.; Key Laboratory for Molecular Enzymology and Engineering of Ministry of Education, Edmond H. Fischer Signal Transduction Laboratory, School of Life Sciences, Jilin University, Changchun, 130012, ChinaCardiotoxicity is a critical issue in drug development that poses serious health risks, including potentially fatal arrhythmias. The human ether-à-go-go related gene (hERG) potassium channel, as one of the primary targets of cardiotoxicity, has garnered widespread attention. Traditional cardiotoxicity testing methods are expensive and time-consuming, making computational virtual screening a suitable alternative. In this study, we employed machine learning techniques utilizing molecular fingerprints and descriptors to predict the cardiotoxicity of compounds, with the aim of improving prediction accuracy and efficiency. We used four types of molecular fingerprints and descriptors combined with machine learning and deep learning algorithms, including Gaussian naive Bayes (NB), random forest (RF), support vector machine (SVM), K-nearest neighbors (KNN), eXtreme gradient boosting (XGBoost), and Transformer models, to build predictive models. Our models demonstrated advanced predictive performance. The best machine learning model, XGBoost Morgan, achieved an accuracy (ACC) value of 0.84, and the deep learning model, Transformer_Morgan, achieved the best ACC value of 0.85, showing a high ability to distinguish between toxic and non-toxic compounds. On an external independent validation set, it achieved the best area under the curve (AUC) value of 0.93, surpassing ADMETlab3.0, Cardpred, and CardioDPi. In addition, we explored the integration of molecular descriptors and fingerprints to enhance model performance and found that ensemble methods, such as voting and stacking, provided slight improvements in model stability. Furthermore, the SHapley Additive exPlanations (SHAP) explanations revealed the relationship between benzene rings, fluorine-containing groups, NH groups, oxygen in ether groups, and cardiotoxicity, highlighting the importance of these features. This study not only improved the predictive accuracy of cardiotoxicity models but also promoted a more reliable and scientifically interpretable method for drug safety assessment. Using computational methods, this study facilitates a more efficient drug development process, reduces costs, and improves the safety of new drug candidates, ultimately benefiting medical and public health.http://www.sciencedirect.com/science/article/pii/S2095177925000802CardiotoxicityHuman ether-à-go-go related gene channelDeep learningMolecular fingerprintDrug development
spellingShingle Kaifeng Liu
Huizi Cui
Xiangyu Yu
Wannan Li
Weiwei Han
Predicting cardiotoxicity in drug development: A deep learning approach
Journal of Pharmaceutical Analysis
Cardiotoxicity
Human ether-à-go-go related gene channel
Deep learning
Molecular fingerprint
Drug development
title Predicting cardiotoxicity in drug development: A deep learning approach
title_full Predicting cardiotoxicity in drug development: A deep learning approach
title_fullStr Predicting cardiotoxicity in drug development: A deep learning approach
title_full_unstemmed Predicting cardiotoxicity in drug development: A deep learning approach
title_short Predicting cardiotoxicity in drug development: A deep learning approach
title_sort predicting cardiotoxicity in drug development a deep learning approach
topic Cardiotoxicity
Human ether-à-go-go related gene channel
Deep learning
Molecular fingerprint
Drug development
url http://www.sciencedirect.com/science/article/pii/S2095177925000802
work_keys_str_mv AT kaifengliu predictingcardiotoxicityindrugdevelopmentadeeplearningapproach
AT huizicui predictingcardiotoxicityindrugdevelopmentadeeplearningapproach
AT xiangyuyu predictingcardiotoxicityindrugdevelopmentadeeplearningapproach
AT wannanli predictingcardiotoxicityindrugdevelopmentadeeplearningapproach
AT weiweihan predictingcardiotoxicityindrugdevelopmentadeeplearningapproach