CurveP Method for Rendering High-Throughput Screening Dose-Response Data into Digital Fingerprints.

Area of Science:

• Computational chemistry
• Bioinformatics
• Drug discovery

Background:

• High-throughput screening (HTS) generates large datasets with potential artifacts.
• Standard data normalization may not fully address signal interferences and incomplete readouts.
• Accurate dose-response curve analysis is crucial for reliable HTS data interpretation.

Purpose of the Study:

• To introduce CurveP, a heuristic algorithm for cleaning dose-response curves in HTS data.
• To enable the generation of reliable sample fingerprints or feature vectors from corrected curves.
• To facilitate downstream computational applications like QSAR modeling and toxicology prediction.

Main Methods:

• CurveP employs a user-tunable, heuristic approach to correct dose-response curves.
• The algorithm identifies and applies a minimal set of corrections to achieve monotonic curves.
• Numeric features are extracted from the cleaned curves to characterize samples.

Main Results:

• CurveP successfully generates monotonic dose-response curves by addressing artifacts.
• The algorithm produces a set of numeric features serving as sample fingerprints.
• These features are suitable for use as independent variables in computational models.

Conclusions:

• CurveP offers an effective solution for HTS data preprocessing, improving data quality.
• The generated features enhance the utility of HTS data for QSAR, toxicology, and bioinformatics.
• This algorithm supports more robust and accurate computational analysis in cheminformatics and drug discovery.