Early assessments of pharmacokinetic and toxic properties are important stepping stones in drug development. We have developed prediction models that can quickly assess some of the most important properties, such as cytotoxicity, genotoxicity, cardiotoxicity, drug-drug interactions, microsomal stability, and drug-induced liver injury, with high confidence. The accuracies among the 14 prediction models shown in the radar chart are all greater than 75% and on average the models were able to predict 75% of the compounds in their datasets, using 10-fold cross validation. In this talk, I will highlight prediction models that address blood-brain barrier permeability, drug efflux, and mitochondrial toxicity. These models, which are based on our variable nearest neighbor (vNN) algorithm, all use the fundamental principle that “similar molecules have similar biological activity”. With the use of our developed models, experimentalists can quickly prioritize which molecules to study based on their predicted properties. We are currently developing a free web-based modeling platform where users can run our models or easily build and customize their own models.

<i>In silico</i> pharmacology: Predicting pharmacokinetic and toxic properties

254th National Meeting (2017)

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

In silico pharmacology: Predicting pharmacokinetic and toxic properties