The Derivation of a Matched Molecular Pairs Based ADME/Tox Knowledge Base for Compound Optimization

Area of Science:

• Computational Chemistry
• Medicinal Chemistry
• Drug Discovery

Background:

• Matched Molecular Pairs (MMP) analysis is crucial for Structure Activity and Property (SAR/SPR) analysis.
• Summarizing MMPs into chemical transforms aids predictive modeling, especially for Absorption, Distribution, Metabolism, and Elimination (ADME) properties.
• A comprehensive knowledge database of these transforms can significantly enhance multidimensional optimization in drug design.

Purpose of the Study:

• To detail a workflow for deriving a knowledge database of chemical transforms from MMP analysis.
• To make the MMP fragmentation algorithm and transform derivation methods freely available via the LillyMol software package.
• To apply the method to ADME/Toxicity (ADME/Tox) data and demonstrate its utility in automated compound design and suggestion.

Main Methods:

• Developed an MMP fragmentation algorithm and statistical summarization for transform derivation.
• Applied the transform database to ADME/Tox assay datasets, identifying discrepancies with traditional medicinal chemistry transforms.
• Created an internal software interface for automated compound design suggestions based on the matched pairs database.

Main Results:

• Highlighted instances where MMP data contradicted established medicinal chemistry transforms for ADME/Tox modulation.
• Demonstrated the utility of the transform database in an automated compound design scenario.
• Compared the knowledge database against larger MMP datasets, revealing limited coverage of all possible transforms but better coverage for common medicinal chemistry strategies.

Conclusions:

• The developed workflow and LillyMol software provide a powerful tool for generating a chemical transform knowledge database.
• The approach aids in moving beyond idea generation towards high-quality prediction of novel ADME/Tox modulating transforms.
• Freely available software and transforms support medicinal chemists in optimizing compound properties.