REPROVIS-DB: a benchmark system for ligand-based virtual screening derived from reproducible prospective applications

Area of Science:

• Computational chemistry and cheminformatics
• Drug discovery and medicinal chemistry

Background:

• Virtual screening (VS) is crucial for evaluating drug candidates.
• Traditional VS benchmarking methods often lack statistical validation and realism.
• Existing benchmarks may overestimate VS method performance and do not reflect prospective applications.

Purpose of the Study:

• To address limitations in current VS benchmarking practices.
• To create a more realistic and reproducible environment for evaluating VS methods.
• To facilitate the assessment of VS tools under conditions mimicking practical drug discovery.

Main Methods:

• Development of a publicly available compound database named REPROVIS-DB.
• REPROVIS-DB organizes data from successful ligand-based VS applications.
• Includes reference compounds, screening databases, selection criteria, and confirmed hits for reproducibility.

Main Results:

• The REPROVIS-DB currently contains 25 curated compound datasets.
• These datasets enable the reproduction of successful virtual screens using alternative methods.
• Facilitates direct comparison and assessment of different VS approaches.

Conclusions:

• REPROVIS-DB provides a valuable resource for more accurate VS method evaluation.
• This database enhances the reliability of benchmarking by simulating practical VS conditions.
• Supports improved selection and application of VS tools in drug discovery pipelines.