Related Experiment Video
Updated: Jul 14, 2026
11:06
Human Liver Microphysiological System for Assessing Drug-Induced Liver Toxicity In Vitro
Published on: January 31, 2022
Development of in silico models for human liver microsomal stability
Pil H Lee1, Lourdes Cucurull-Sanchez, Jing Lu
1Computer-Assisted Drug Discovery, Pfizer Global Research and Development, Ann Arbor, MI 48105, USA. pil.h.lee@pfizer.com
Journal of Computer-Aided Molecular Design
|June 30, 2007
Summary
We created predictive models for human liver microsomal (HLM) stability to aid drug discovery. These models accurately forecast metabolic clearance, helping to optimize drug compounds and prevent late-stage failures.
Area of Science:
- Drug Discovery and Development
- Computational Chemistry
- Pharmacokinetics
Background:
- Human liver microsomal (HLM) stability is crucial for predicting a compound's metabolic clearance.
- Accurate in silico models for metabolic clearance are vital in early drug discovery for lead optimization and failure prediction.
Purpose of the Study:
- To develop highly predictive classification models for HLM stability.
- To utilize apparent intrinsic clearance (CL(int, app)) as the endpoint for HLM stability prediction.
Main Methods:
- Employed Random Forest and Bayesian classification algorithms.
- Utilized MOE, E-state descriptors, ADME Keys, and ECFP_6 fingerprints for model development.
Main Results:
- Achieved 80% prediction accuracy on the test set.
- Attained 75% prediction accuracy on the validation set.
- Identified significant descriptors and assessed prediction confidence.
Conclusions:
- Developed robust computational models for predicting HLM stability.
- These models can significantly aid in early-stage drug discovery and lead selection.
