From in silico target prediction to multi-target drug design: current databases, methods and applications

Area of Science:

• Computational chemistry
• Cheminformatics
• Pharmacology

Background:

• The rapid expansion of bioactivity databases necessitates advanced computational methods for target identification.
• Predicting protein targets of small molecules is crucial for drug discovery, polypharmacology design, and mode-of-action analysis.

Purpose of the Study:

• To review existing databases and computational methods for ligand-based and structure-based target prediction.
• To discuss successful applications of in silico target identification, including the authors' own work.
• To explore the prospective use of databases for designing ligands with desired polypharmacology.

Main Methods:

• Survey of ligand-based bioactivity databases.
• Outline of computational target prediction methods (ligand-based and structure-based).
• Discussion of case studies and author's experience with phenotypic antibacterial screens and high-throughput screening data analysis.

Main Results:

• Databases for ligand-based target prediction have grown significantly.
• Various computational methods exist for target prediction, leveraging bioactivity data or protein structures.
• In silico target identification has led to successful applications in drug discovery and mode-of-action studies.

Conclusions:

• Computational target prediction is a powerful approach for both retrospective analysis and prospective drug design.
• Databases and predictive tools are essential for understanding small molecule interactions and designing novel therapeutics.
• Future applications include designing ligands for specific polypharmacological profiles.