Ranking targets in structure-based virtual screening of three-dimensional protein libraries: methods and problems

Area of Science:

• Computational chemistry
• Drug discovery
• Structural biology

Background:

• Structure-based virtual screening is a key method for identifying drug targets.
• Inverse screening, where a ligand is docked against multiple targets, offers an alternative approach.
• Scoring functions remain a challenge in molecular docking accuracy.

Purpose of the Study:

• To evaluate ranking protocols for inverse virtual screening.
• To assess the effectiveness of GOLD docking and interaction fingerprints (IFP) in identifying known protein targets for specific ligands.
• To determine if a structure-based approach can efficiently select a limited set of targets for experimental validation.

Main Methods:

• Retrospective virtual screening of the sc-PDB library using GOLD docking.
• Docking of four known ligands (biotin, 4-hydroxy-tamoxifen, 6-hydroxy-1,6-dihydropurine ribonucleoside, methotrexate) against multiple protein targets.
• Evaluation of ranking protocols combining GOLD fitness score and topological molecular interaction fingerprint (IFP) comparison.

Main Results:

• A combined approach using GOLD fitness and two IFP scores successfully recovered most known targets for the investigated ligands.
• This method effectively filtered out a significant number of false positive predictions.
• Rare proteins, unsuitable for statistical analysis, were also successfully identified.

Conclusions:

• The fusion of GOLD docking fitness and IFP scores enhances the accuracy of inverse virtual screening.
• This structure-based strategy enables the efficient selection of a small, manageable number of high-confidence targets for experimental testing.
• The findings support the feasibility of purely structure-based approaches in prioritizing drug targets.