Cross-Classified Multilevel Modelling of the Effectiveness of Similarity-Based Virtual Screening

Area of Science:

• Computational chemistry
• Cheminformatics
• Drug discovery

Background:

• Chemical similarity searching is vital for identifying bioactive molecules.
• Its effectiveness is influenced by bioactivity, similarity coefficients, fingerprints, and reference structures.
• Previous studies have not fully elucidated the relative importance of these factors.

Purpose of the Study:

• To investigate the relative importance of four key factors influencing chemical similarity search effectiveness.
• To analyze similarity searches within the ChEMBL database using advanced statistical modeling.

Main Methods:

• Utilized cross-classified multilevel modeling to analyze the impact of different factors.
• Evaluated chemical similarity searches against the ChEMBL database.
• Assessed the roles of bioactivity, similarity coefficient, fingerprint type, and reference structure nature.

Main Results:

• The molecular fingerprint type significantly impacts similarity search effectiveness more than the similarity coefficient.
• The nature of the reference structure is also a critical factor in search performance.
• The bioactivity of interest influences the overall success of the search.

Conclusions:

• Molecular fingerprints are a primary determinant of chemical similarity search success.
• Future comparative studies of similarity measures should incorporate a broader range of reference structures.
• Optimizing fingerprint selection is key to enhancing drug discovery screening efforts.