Quantification of frequent-hitter behavior based on historical high-throughput screening data

Area of Science:

• Drug discovery and development
• Computational chemistry
• Bioinformatics

Background:

• High-throughput screening (HTS) is crucial for identifying potential drug candidates.
• A significant challenge in HTS is the prevalence of 'frequent hitters'—compounds that appear repeatedly across multiple screens, often indicating false positives.
• Investigating these frequent hitters consumes valuable resources and reduces the overall efficiency of the screening process.

Purpose of the Study:

• To develop and implement a systematic retrospective approach for identifying and characterizing frequent hitters from historical high-throughput screening data.
• To improve the accuracy and efficiency of hit identification in drug discovery pipelines.
• To mitigate the resource wastage associated with false-positive results in HTS.

Main Methods:

• Utilized historical high-throughput screening data for analysis.
• Developed a systematic retrospective method to detect anomalous compound behavior.
• Accounted for data uncertainty when limited screening information is available.
• Extended the methodology to various target and technology classes.

Main Results:

• Successfully identified frequent hitters, which are compounds likely to be false positives.
• The developed method allows for the characterization of anomalous hitter behavior.
• Demonstrated the potential for historical data bias to influence compound annotation, providing specific examples.

Conclusions:

• The identification and characterization of frequent hitters enhance the efficiency of drug discovery by enabling focused follow-up on more promising compounds.
• This approach optimizes resource allocation within the hit-deconvolution cascade.
• Increased efficiency in screening delivery is achieved by reducing the investigation of false-positive results.