Comparison of methods for sequential screening of large compound sets

Area of Science:

• Drug discovery and cheminformatics
• Computational chemistry and data analysis

Background:

• Sequential screening is an iterative method to enhance hit rates compared to random or non-iterative approaches.
• Understanding factors influencing enrichment rates is crucial for optimizing drug discovery pipelines.

Purpose of the Study:

• To investigate the impact of compound ranking methods, molecular descriptor sets, and initial training set selection on enrichment rates in sequential screening.
• To identify key factors that maximize the recovery of active compounds during iterative screening processes.

Main Methods:

• Evaluated the effects of three factors: compound ranking method, molecular descriptor set, and initial training set selection.
• Compared enrichment rates achieved through different training set selection strategies.
• Analyzed the influence of these factors on the recovery of active compounds.

Main Results:

• The method of selecting the initial training set was identified as the primary factor influencing recovery rates.
• Diverse training sets resulted in substantially lower recovery rates for active compounds compared to other selection methods.
• Sequential screening demonstrated significant improvement in average active compound recovery rates over non-iterative procedures.

Conclusions:

• Initial training set selection critically affects the success of sequential screening.
• Avoiding diverse initial training sets is recommended to improve active compound recovery.
• Sequential screening, when optimized, offers a substantial advantage in identifying active compounds compared to traditional methods.