Evaluating the Diversity and Target Addressability of DELs using Scaffold Analysis and Machine Learning.

Area of Science:

• Medicinal Chemistry
• Cheminformatics
• Drug Discovery

Background:

• DNA-encoded libraries (DELs) are crucial for drug discovery, enabling efficient screening of large combinatorial libraries.
• Maximizing DEL success requires optimizing scaffold diversity and target addressability alongside physicochemical properties.
• Existing tools lack a combined computational approach for evaluating both scaffold diversity and target-orientedness.

Purpose of the Study:

• To develop and present a novel cheminformatics tool for evaluating scaffold diversity and target addressability in DELs.
• To provide a computational method that integrates scaffold analysis and machine learning for DEL library assessment.
• To aid medicinal chemists in selecting appropriate DELs for specific drug discovery objectives.

Main Methods:

• Development of a cheminformatics tool utilizing scaffold analysis.
• Integration of machine learning algorithms to assess target-orientedness.
• Case study analysis of two in-house DEL libraries to validate the workflow.

Main Results:

• The tool successfully evaluates both scaffold diversity and target addressability.
• The workflow can differentiate between generalist and focused DEL libraries.
• Demonstrated capability to guide library selection for hit-finding and hit-optimization.

Conclusions:

• The presented tool offers a valuable computational approach for DEL library assessment.
• It enhances the strategic selection of libraries in drug discovery programs.
• The tool is accessible as a free web application and Python script.