Expanding ADMET Predictor®’s Chemical Space: Enhanced bRo5 and Chameleon Molecule Predictions for HTPK

Authors: Jones J, Bachorz RA, Chupakhin V, Lawless M, Miller D, Fraczkiewicz R
Conference: PhysChem Forum
Keywords: admet predictor, ai drug discovery, bRo5 compounds, chameleon molecule predictions, HTPK, in vivo
Software: ADMET Predictor®
Division: Cheminformatics

Abstract

ADMET Predictor has been enhanced to accurately predict properties of beyond Rule-of-Five (bRo5) molecules, including macrocycles and PROTACs. We introduce new descriptors for molecular chameleonicity and three specialized models: EPSA1 (Experimental Polar Surface Area), ChromLogD2, and ChameLogK3 (Chromatographic Chameleonicity), which serve as advanced molecular features for core ADMET predictions. These new descriptors improve predictive capabilities for various HTPK-input models, such as liver microsome/hepatocyte clearance, fraction unbound in plasma, and blood-to-plasma ratio. This subsequently enhances overall HTPK modeling accuracy. Validation on novel molecules demonstrates significant performance improvements for bRo5 compounds. Case studies reveal substantial improvements in predicting key in vivo endpoints for challenging chemical space, supporting modern drug discovery.

By Jeremy Jones, Rafal A. Bachorz, Vladimir Chupakchin, Michael Lawless, David Miller, and Robert Fraczkiewicz

PhysChem Forum (PCF), October 13-14, 2025, Basel, Switzerland