A Hybrid Mixture Discriminant Analysis−Random Forest Computational Model for the Prediction of Volume of Distribution of Drugs in Human

• Published in: Journal of medicinal chemistry 2006-04, Vol.49 (7), p.2262-2267
• Main Authors: Lombardo, Franco, Obach, R. Scott, DiCapua, Frank M, Bakken, Gregory A, Lu, Jing, Potter, David M, Gao, Feng, Miller, Michael D, Zhang, Yao
• Format: Article
• Language: English
• Subjects: Algorithms; Biological and medical sciences; Computer Simulation; Drug Design; General pharmacology; Humans; Medical sciences; Models, Biological; Pharmaceutical Preparations - metabolism; Pharmacokinetics; Pharmacokinetics. Pharmacogenetics. Drug-receptor interactions; Pharmacology. Drug treatments; Tissue Distribution
• ISSN: 0022-2623; 1520-4804
• DOI: 10.1021/jm050200r

A computational approach is described that can predict the VDss of new compounds in humans, with an accuracy of within 2-fold of the actual value. A dataset of VD values for 384 drugs in humans was used to train a hybrid mixture discriminant analysis−random forest (MDA-RF) model using 31 computed descriptors. Descriptors included terms describing lipophilicity, ionization, molecular volume, and various molecular fragments. For a test set of 23 proprietary compounds not used in model construction, the geometric mean fold-error (GMFE) was 1.78-fold (±11.4%). The model was also tested using a leave-class out approach wherein subsets of drugs based on therapeutic class were removed from the training set of 384, the model was recast, and the VDss values for each of the subsets were predicted. GMFE values ranged from 1.46 to 2.94-fold, depending on the subset. Finally, for an additional set of 74 compounds, VDss predictions made using the computational model were compared to predictions made using previously described methods dependent on animal pharmacokinetic data. Computational VDss predictions were, on average, 2.13-fold different from the VDss predictions from animal data. The computational model described can predict human VDss with an accuracy comparable to predictions requiring substantially greater effort and can be applied in place of animal experimentation.