LigandBox: A database for 3D structures of chemical compounds

Area of Science:

• Computational chemistry
• Drug discovery
• Bioinformatics

Background:

• Virtual screening using molecular docking requires comprehensive 3D compound structure databases.
• Existing databases may lack sufficient diversity or structural detail for effective screening.

Purpose of the Study:

• To develop and present the LigandBox database, a large-scale repository of 3D chemical compound structures.
• To provide researchers with a readily accessible resource for virtual screening and compound discovery.

Main Methods:

• Compiled four million compounds from commercial suppliers and curated databases (KEGG, PDB).
• Generated 3D conformers with hydrogen atoms and charges using the myPresto simulation package.
• Calculated ADME-Tox properties, including aqueous solubility (LogS) and carcinogenicity.
• Implemented property/chemical ID and chemical structure search services (descriptor and Maximum Common Substructure (MCS) search via kcombu).

Main Results:

• The LigandBox database contains over four million compounds with pre-calculated 3D structures and properties.
• Efficient search functionalities allow rapid identification of similar compounds based on structure or properties.
• The database is ready for direct use in molecular docking and virtual screening workflows.

Conclusions:

• LigandBox provides a valuable, large-scale resource for virtual screening and the discovery of novel chemical compounds.
• The database facilitates research in medical science, chemical biology, and biochemistry by enabling efficient compound identification.
• The inclusion of 3D conformers and ADME-Tox properties enhances its utility for drug discovery pipelines.