How to optimize shape-based virtual screening: choosing the right query and including chemical information

Area of Science:

• Computational Chemistry
• Drug Discovery
• cheminformatics

Background:

• Shape-based molecular similarity is a widely used virtual screening technique.
• Purely shape-based screening often overlooks crucial chemical and pharmacophoric details.
• The Directory of Useful Decoys (DUD) serves as a benchmark for evaluating screening methods.

Purpose of the Study:

• To systematically assess the impact of query conformation choice on screening performance.
• To evaluate how selecting different active compounds as query structures affects results.
• To determine the benefit of including pharmacophoric information in shape-based screening.

Main Methods:

• Systematic evaluation using the Directory of Useful Decoys (DUD) dataset.
• Analysis of screening performance based on variations in query conformation.
• Assessment of performance changes with different active compound selections as queries.
• Integration and evaluation of pharmacophoric properties into the screening process.

Main Results:

• The choice of query conformations significantly impacts virtual screening outcomes.
• Selecting optimal active compounds for query structures leads to improved performance.
• Incorporating pharmacophoric information substantially enhances screening effectiveness.
• Achieved the best screening performance reported to date using these shape-based tools.

Conclusions:

• Optimizing query conformation and active compound selection is critical for shape-based virtual screening.
• Integrating pharmacophoric properties offers substantial improvements over shape-only methods.
• Developed guidelines for achieving optimal performance in virtual screening campaigns.