SimG: an alignment based method for evaluating the similarity of small molecules and binding sites

Area of Science:

• Computational chemistry
• Drug discovery
• Molecular modeling

Background:

• Virtual screening is crucial for identifying potential drug candidates.
• Existing methods often struggle with accurately representing molecular shape and chemical features for similarity assessment.

Purpose of the Study:

• To develop a novel molecular similarity metric, SimG, integrating Gaussian volume overlap and chemical features.
• To assess SimG's effectiveness in both structure-based and ligand-based virtual screening strategies.
• To investigate the impact of binding site properties on virtual screening performance.

Main Methods:

• Devised a similarity metric combining Gaussian volume overlap and chemical feature analysis.
• Employed downhill simplex searching for similarity evaluation.
• Represented molecular and binding site shapes using chemical features and Gaussian volumes.
• Validated the method using actives vs. decoys analysis and comparison with reference methods.
• Conducted retrospective virtual screening on DUD datasets.

Main Results:

• SimG effectively identifies active compounds by comparing molecular and binding site shapes.
• Virtual screening performance is significantly influenced by binding site solvent exposure.
• Structure-based screening on less solvent-exposed sites outperformed ligand-based screening.

Conclusions:

• SimG offers a robust approach for virtual screening, supporting both structure- and ligand-based strategies.
• Binding site characteristics, especially solvent accessibility, are critical factors for successful structure-based virtual screening.
• The developed method shows promise for improving hit identification in drug discovery.