Machine Learning Models to Predict Cytochrome P450 2B6 Inhibitors and Substrates

Cytochrome P450 2B6 (CYP2B6) is responsible for the metabolismof & SIM;7% of marketed drugs. The in vitro drug interaction studiesguidance for industry issued by the FDA stipulates that drug sponsorsneed to evaluate whether the investigated drugs interact with themajor drug-metabolizing P450s including CYP2B6. Therefore, there hasbeen greater attention to the development of predictive models forCYP2B6 inhibitors and substrates. In this study, conventional machinelearning and deep learning models were developed to predict CYP2B6inhibitors and substrates. Our results showed that the best CYP2B6inhibitor model yielded the AUC values of 0.95 and 0.75 with the 10-foldcross-validation and the test set, respectively, and the best CYP2B6substrate model produced the AUC values of 0.93 and 0.90 with the10-fold cross-validation and the test set, respectively. The generalizationability of the CYP2B6 inhibitor and substrate models was assessedby using the external validation sets. Several significant substructuralfragments relevant to CYP2B6 inhibitors and substrates were detectedvia frequency substructure analysis and information gain. In addition,the applicability domain of the models was defined by employing anonparametric method based on the probability density distribution.We anticipate that our results would be useful for the predictionof potential CYP2B6 inhibitors and substrates in the early stage ofdrug discovery.

Automated summary

In this study, conventional machine learning and deep learning models were developed to predict CYP2B6 inhibitors and substrates. The best CYP2B6 inhibitor model achieved AUC values of 0.95 and 0.75 with 10-fold cross-validation and the test set, respectively. The best CYP2B6 substrate model produced AUC values of 0.93 and 0.90 with 10-fold cross-validation and the test set, respectively. The models were validated using external validation sets and several significant substructural fragments relevant to CYP2B6 inhibitors and substrates were identified.