Prediction of cytochrome P450-mediated bioactivation using machine learning models and in vitro validation.

Cytochrome P450 (P450)-mediated bioactivation, which can lead to the hepatotoxicity through the formation of reactive metabolites (RMs), has been regarded as the major problem of drug failures. Herein, we purposed to establish machine learning models to predict the bioactivation of P450. On the basis of the literature-derived bioactivation dataset, models for Benzene ring, Nitrogen heterocycle and Sulfur heterocycle were developed with machine learning methods, i.e., Random Forest, Random Subspace, SVM and Naïve Bayes. The models were assessed by metrics like "Precision", "Recall", "F-Measure", "AUC" (Area Under the Curve), etc. Random Forest algorithms illustrated the best predictability, with nice AUC values of 0.949, 0.973 and 0.958 for the test sets of Benzene ring, Nitrogen heterocycle and Sulfur heterocycle models, respectively. 2D descriptors like topological indices, 2D autocorrelations and Burden eigenvalues, etc. contributed most to the models. Furthermore, the models were applied to predict the occurrence of bioactivation of an external verification set. Drugs like selpercatinib, glafenine, encorafenib, etc. were predicted to undergo bioactivation into toxic RMs. In vitro, IC50 shift experiment was performed to assess the potential of bioactivation to validate the prediction. Encorafenib and tirbanibulin were observed of bioactivation potential with shifts of 3-6 folds or so. Overall, this study provided a reliable and robust strategy to predict the P450-mediated bioactivation, which will be helpful to the assessment of adverse drug reactions (ADRs) in clinic and the design of new candidates with lower toxicities.

Computational prediction for the metabolism of human UDP-glucuronosyltransferase 1A1 substrates.

UDP-glucuronosyltransferase (UGT) 1A1, one of the most important isoforms in UGTs superfamily, has attracted increasing concerns for its special role in the clearance and detoxification of endogenous and exogenous substances. To avoid the clinical drug-drug interactions, it is of great importance to have the knowledge of the metabolic profile of UGT1A1 substrates early. Herein, we purposed to establish machine learning models to predict the metabolic propeties of UGT1A1 substrates. On the basis of the literature-derived substrates database of UGT1A1, automatic metabolism prediction models for the aromatic hydroxyl (ArOH) and carboxyl (COOH) groups were developed with eight machine learning methods, among which, three methods, i.e. Random Forest, Random Subspace and J48, illustrated the best performance either for the aromatic hydroxyl and the carboxyl model. The models illustrated good robustness when they were evaluated with functions like "Precision", "Recall", "F-Measure", "AUC", "MCC", etc. Nice accuracy was observed for the aromatic hydroxyl and carboxyl model of these methods, whose AUCs ranged from 0.901 to 0.997. Additionally, the ArOH model was applied to predict the UGT1A1-mediated metabolism of an external set. Two new unknown substrates, cytochrome P450 (CYPs)-mediated metabolites of gefitinib, were predicted and identified, which were validated by in vitro assays. In summary, this study provides a reliable and robust strategy to predict UGT1A1 metabolites, which will be helpful either in rational-optimization of drug metabolism or in avoiding drug-drug interactions in clinic.