Calculation and application of activity discriminants in lead optimization

Area of Science:

• Drug Discovery and Medicinal Chemistry
• Computational Chemistry and Cheminformatics

Background:

• Predicting in vitro assay activity is crucial for efficient drug discovery.
• Lead optimization requires accurate structure-activity relationship (SAR) analysis and activity prediction.

Purpose of the Study:

• To present a technique for computing activity discriminants for in vitro assays.
• To apply this technique for predicting the in vitro activities of synthetic targets during lead optimization.
• To develop a method that emulates medicinal chemists' SAR analysis and activity prediction processes.

Main Methods:

• Computing activity discriminants based on 6 common medicinal chemistry descriptors.
• Utilizing visualization tools (e.g., Spotfire) for analyzing query molecule relationships and discriminant relevance.
• Validating the approach with historical compound data from AstraZeneca Wilmington (since 2006).

Main Results:

• Developed interpretable activity discriminants easily understood by medicinal chemists.
• Demonstrated the utility of visualization for assessing compound relationships and discriminant relevance.
• Validated the approach's effectiveness in prioritizing new synthetic targets for synthesis.

Conclusions:

• The presented technique provides a valuable tool for predicting in vitro activities.
• This method supports medicinal chemists in making informed decisions during lead optimization.
• The approach enhances the efficiency of drug discovery by prioritizing synthetic targets effectively.