Skip to main content
 

In silico Model Development for Predicting Human CYP3A4 Metabolic Stability Using Different Descriptor Generating Software and Machine Learning Techniques

Thursday, September 15, 2016 — Poster Session II

12:00 p.m. – 1:30 p.m.
FAES Terrace
NCATS
RSCHSUPP-4

Authors

  • P Shah
  • D-T Nguyen
  • A Zakharov
  • N Katori
  • A Jadhav
  • X Xu

Abstract

Hepatic intrinsic clearance (CLint) is an important parameter for a drug candidate as it influences oral bioavailability. To facilitate drug design and accelerate translational research, development of in silico tools for reliable prediction of human clearance will be very useful. The major enzymatic system responsible for metabolism of xenobiotics is the cytochrome P450 (CYP450) family. CYP3A4 is responsible for metabolism of ~50% of known xenobiotics in humans. Our objective is to measure CLint for large sets of compounds with individual major CYP isozymes beginning with CYP3A4, from which in silico prediction tools can be developed. In vitro half-life (t1/2) method was used to determine CLint. Data was generated for 4000 compounds using our automated high-throughput metabolic stability assay. 2300 compounds were used as training set for QSAR model building and 1700 compounds were used as prospective test set for model validation. To construct the QSAR models we calculated different types of descriptors using different software. Several different machine learning were applied to generate the best in silico model. The compounds tested were classified as metabolically stable or unstable using t1/2 of 30 min as the cut off. We find that the QSAR model based on ADMET predictor and QNA based descriptors yielded the best results i.e. balanced accuracy >0.7. The best model was incorporated in the publicly available NCATS web services (http://tripod.nih.gov/adme). The model together with other in silico ADME tools will aid structure optimization and lead selection and accelerate translational research in drug discovery and development.

Category: Research Support Services