Ligand-based virtual screening approach using a new scoring function.

Area of Science:

• Computational Chemistry
• Cheminformatics
• Drug Discovery

Background:

• Virtual screening is crucial for identifying potential drug candidates.
• Existing ligand-based virtual screening methods have limitations in accuracy and robustness.
• Development of improved scoring functions is essential for enhancing screening performance.

Purpose of the Study:

• To develop a novel ligand-based virtual screening approach.
• To introduce an effective shape-overlapping procedure and a robust scoring function (HWZ score).
• To evaluate the performance of the HWZ score-based virtual screening against established methods.

Main Methods:

• Implemented a new ligand-based virtual screening strategy.
• Utilized an effective shape-overlapping procedure.
• Employed a robust scoring function, termed the HWZ score, for compound evaluation.
• Tested the approach against 40 protein targets from the Database of Useful Decoys (DUD).
• Assessed performance using standard metrics: Area Under the Receiver Operator Characteristic Curve (AUC), Enrichment Factor (EF), and Hit Rate (HR).

Main Results:

• The HWZ score-based virtual screening demonstrated improved overall performance compared to other examined approaches.
• Achieved an average AUC of 0.84 ± 0.02 across 40 protein targets.
• Obtained average Hit Rates (HR) of 46.3% ± 6.7% (top 1%) and 59.2% ± 4.7% (top 10%).
• The method showed reduced sensitivity to the specific choice of target protein.

Conclusions:

• The developed HWZ score-based virtual screening approach offers a more effective and robust method for ligand-based screening.
• This advancement holds significant potential for accelerating drug discovery by improving the identification of active compounds.
• The approach's robustness suggests broad applicability across diverse protein targets.