GOLD

GOLD predicts protein-ligand binding modes and estimates binding affinities to support virtual screening and lead optimization in drug discovery.

Key Features:

• Scoring functions: Implements Goldscore and Chemscore scoring functions for evaluating protein-ligand interactions.
• Goldscore: Provides superior performance in predicting binding energies, particularly effective with larger ligands, with a reported standard deviation of approximately 10.5 kJ/mol.
• Chemscore: Enables faster docking speeds—reported up to three times quicker than Goldscore—making it suitable for rapid virtual screening.
• Consensus docking protocols: Includes Goldscore-CS and Chemscore-GS protocols that combine docking and ranking steps using both scoring functions.
• Goldscore-CS protocol: Performs docking with Goldscore and ranking with Chemscore, achieving reported success rates up to 81% for top-ranked solutions within 2.0 Å of experimental binding modes.
• Chemscore-GS protocol: Performs docking with Chemscore followed by Goldscore for ranking, yielding reported success rates of about 78% for drug-like compounds and 85% for fragment-like compounds.
• Performance metrics: Reported accuracies were evaluated on a "clean" set of 224 protein-ligand complexes, with docking speeds of about 1–2 min per compound for Goldscore.
• Application suitability: Demonstrates ability to identify correct binding modes across diverse test sets and to support virtual screening and lead optimization workflows.

Scientific Applications:

• Binding mode prediction: Predicts active molecule binding modes to support structure-based analysis.
• Virtual screening: Ranks compound libraries using Goldscore and Chemscore to enrich active molecules in screening campaigns.
• Lead optimization: Assists lead optimization by providing binding affinity estimates and pose predictions for structure–activity relationship analysis.
• Database enrichment: Enhances database enrichment strategies through consensus docking and scoring protocols.
• Computational chemistry and structural biology: Applies to molecular modeling studies that require protein-ligand interaction assessment.

Methodology:

Performs ligand docking using Goldscore and Chemscore scoring functions, applies consensus protocols (Goldscore-CS and Chemscore-GS) that combine docking and ranking, and employs genetic optimization for ligand docking.