The confirmation rate of primary hits: a predictive model

Area of Science:

• Drug Discovery
• Biotechnology
• Computational Chemistry

Background:

• High-throughput screening (HTS) is crucial for identifying bioactive compounds.
• Traditional HTS data analysis relies on activity cutoffs, which can lead to suboptimal selection of true positives and inclusion of false positives.
• Existing methods for analyzing screening data lack optimization for maximizing screening efficiency.

Purpose of the Study:

• To present an alternative approach to HTS data analysis using a calculated probability of being active.
• To optimize the selection of primary positives for follow-up studies to maximize true positives while minimizing false positives.
• To enable the deconvolution of orthogonal mixtures without presetting a deconvolution threshold.

Main Methods:

• Development of a probability-based analysis for HTS data.
• Calculation of a predicted confirmation rate derived from the probability of activity.
• Application of the probability method for deconvoluting orthogonal mixtures in screening assays.

Main Results:

• The probability-based method allows for optimization of the number of primary positives selected for follow-up.
• This approach maximizes the identification of true positives while minimizing false positives compared to traditional cutoff methods.
• Orthogonal mixtures can be deconvoluted effectively without a predefined threshold, recording individual compound results regardless of assay format.

Conclusions:

• The probability of being active offers a more optimized and flexible approach to HTS data analysis.
• This method enhances the efficiency of drug discovery screening campaigns.
• It provides a robust framework for analyzing complex screening data, including mixtures.