An efficient computational chemistry approach to generating negative data for drug discovery pipeline validation

Area of Science:

• Computational Chemistry
• Drug Discovery
• Bioinformatics

Background:

• Virtual high-throughput screening (VHTS) pipelines are crucial for drug discovery but often lack rigorous validation.
• Existing benchmarking studies are limited, focusing mainly on docking and using flawed datasets that inflate performance metrics.

Purpose of the Study:

• To present a new method for VHTS pipeline validation and negative data generation.
• To enable rigorous assessment of each step in VHTS pipelines without experimental costs.

Main Methods:

• Generating vast amounts of negative data by randomizing ligands across experimental structures and creating structural isomers of known binders.
• Utilizing these positive and negative data sets to validate VHTS pipelines at every step.
• Ensuring generated data points closely match in key molecular properties.

Main Results:

• The proposed method provides practically unlimited negative data for VHTS validation.
• This approach allows for precise assessment of enrichment at each stage of a VHTS pipeline.
• It facilitates the distinction between genuinely effective and ineffective VHTS tools.

Conclusions:

• Rigorous validation using high-quality, large-scale negative data is essential for VHTS pipelines.
• This method offers a cost-effective way to improve the reliability and accuracy of VHTS.
• Accurate VHTS tools accelerate hit discovery and lead optimization, addressing critical medical needs.