Analysis of High-Dimensional Structure-Activity Screening Datasets Using the Optimal Bit String Tree.

Area of Science:

• Computational chemistry
• Cheminformatics
• Drug discovery

Background:

• Quantitative structure-activity relationships (QSARs) are crucial for predicting drug efficacy.
• Existing QSAR methods may lack accuracy and efficiency in identifying complex relationships.
• Novel computational approaches are needed to enhance drug screening and development.

Purpose of the Study:

• To introduce a new classification method, the Optimal Bit String Tree (OBSTree), for QSAR analysis.
• To develop an optimized stochastic searching scheme for selecting descriptor sets.
• To demonstrate the effectiveness of OBSTree in identifying QSAR rules and active compounds.

Main Methods:

• The OBSTree method utilizes a 'chromosome' concept to represent descriptor combinations.
• A stochastic searching scheme, including weighted sampling and simulated annealing, optimizes splitting variables.
• The method was validated through simulation studies and application to monoamine oxidase (MAO) inhibitor screening.

Main Results:

• OBSTree demonstrated high accuracy and effectiveness in identifying QSAR rules.
• The method successfully identified different classes of active compounds, including MAO inhibitors.
• The OBSTree algorithm provides a robust framework for analyzing chemical structure-activity data.

Conclusions:

• OBSTree offers an advantageous approach for accurate and effective QSAR analysis.
• The method facilitates the discovery of novel active compounds through optimized descriptor selection.
• Supplementary materials, including SAS code and datasets, are available for further research.