Abstract
Carcinogenicity refers to a highly toxic end point of certain chemicals, and has become an important issue in the drug development process. In this study, three novel ensemble classification models, namely Ensemble SVM, Ensemble RF, and Ensemble XGBoost, were developed to predict carcinogenicity of chemicals using seven types of molecular fingerprints and three machine learning methods based on a dataset containing 1003 diverse compounds with rat carcinogenicity. Among these three models, Ensemble XGBoost is found to be the best, giving an average accuracy of 70.1 ± 2.9%, sensitivity of 67.0 ± 5.0%, and specificity of 73.1 ± 4.4% in five-fold cross-validation and an accuracy of 70.0%, sensitivity of 65.2%, and specificity of 76.5% in external validation. In comparison with some recent methods, the ensemble models outperform some machine learning-based approaches and yield equal accuracy and higher specificity but lower sensitivity than rule-based expert systems. It is also found that the ensemble models could be further improved if more data were available. As an application, the ensemble models are employed to discover potential carcinogens in the DrugBank database. The results indicate that the proposed models are helpful in predicting the carcinogenicity of chemicals. A web server called CarcinoPred-EL has been built for these models (http://ccsipb.lnu.edu.cn/toxicity/CarcinoPred-EL/).
Highlights
Evaluating the toxicity of new compounds is an essential part of the drug development process[1, 2]
The ensemble models outperformed their basic classifiers in both overall accuracy and area under the curve (AUC)
The best ensemble model (Ensemble XGBoost) attained an average accuracy of 70.1 ± 2.9%, sensitivity of 67.0 ± 5.0%, specificity of 73.1 ± 4.4%, and AUC of 76.5 ± 2.9% in five-fold cross-validation and an accuracy of 70.0%, sensitivity of 65.2%, specificity of 76.5%, and AUC of 80.3% in external validation
Summary
Evaluating the toxicity of new compounds is an essential part of the drug development process[1, 2]. Fjodorova et al presented a carcinogenic potency classification model for diverse chemicals that achieved an accuracy of 92.2% on the training set and 68.3% on the test set[20] Their model was constructed using 27 molecular descriptors and a counter-propagation artificial neural network (CP ANN) technique based on a dataset containing 422 carcinogenic and 383 non-carcinogenic organic compounds[20]. Zhang et al built a naïve Bayes classification model using five simple molecular descriptors and extended-connectivity fingerprints (ECFPs), and achieved an overall accuracy of 90% with an internal training set and 68% in five-fold cross-validation[16] These models have a wide AD, but their accuracy in forecasting the carcinogenicity of new compounds (the accuracy estimated by cross-validation or external testing) remains unsatisfactory. Many models that achieve higher accuracy are generated through fine-tuning processes and have not been evaluated by an appropriate cross-validation
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
