Membrane transporters mediate many biological effects of chemicals including food safety, pharmacokinetics, drug resistance, and drug-drug interactions. Using public sources, we have assembled and curated the largest, among publicly available, human intestinal transporter database (over 5000 interaction entries for over 3600 small molecules across 10 major transporters). We have developed a compendium of thoroughly validated Quantitative Structure-Activity Relationship (QSAR) classification models to predict substrates and inhibitors of several transporters (MDR1, BCRP, MRP1-4, PEPT1, ASBT).  Models have been developed using several machine learning approaches (random forest, kNN, and SVM) with both Dragon and MOE descriptors. These models afforded high external prediction accuracies of 66-98% as assessed by 5-fold external validation. These models constitute a platform for virtual screening of chemical libraries to predicts unfavorable drug interactions with transporters as well as identify novel drug candidates or environmental agents with the desired transporter profiles.

Cheminformatics analysis of major human intestinal transporters

243rd National Meeting (2012)

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.