CompScore

CompScore improves structure-based virtual screening by integrating individual components of docking scoring functions into consensus scoring frameworks to maximize enrichment of true ligands.

Key Features:

• Integration of docking scoring components: Incorporates individual components from docking scoring functions into consensus scoring to provide component-level evaluation of ligand–target interactions.
• Genetic algorithm optimization: Employs genetic algorithms to identify optimal combinations of scoring function components that maximize virtual screening enrichment for specific targets.
• Validation across 102 targets: Validated using ligands and decoys for 102 targets commonly used in virtual screening benchmarks.
• Improved enrichment metrics: Demonstrated an average 45% improvement in initial enrichment compared with traditional consensus scoring methods.
• Predictive accuracy with external data: Maintains performance on previously unseen datasets and after redocking with different software.

Scientific Applications:

• Drug discovery and development: Prioritizes potential ligands for experimental validation within structure-based virtual screening workflows.
• Compound prioritization: Enhances selection of candidate molecules by improving initial enrichment of true positives in screening libraries.

Methodology:

Integration of individual docking scoring function components into consensus scoring and optimization of component combinations using genetic algorithms; validation using datasets of ligands and decoys for 102 targets and assessment of predictive accuracy via redocking with different software.