Computer prediction of protein targets of drugs, drug leads, and natural products

Ligand-protein docking has been developed and used in facilitating new drug discovery. In this approach, single or multiple small molecules are attempted to dock into a receptor site so as to find putative ligands. A number of studies have shown that docking algorithms are capable of finding ligands and binding conformations at a receptor site close to experimentally determined structures. These algorithms are expected to be equally applicable to identification of multiple proteins a small molecule can bind or weakly bind to.

We introduce a ligand-protein inverse docking approach for finding putative protein targets of a small molecule by computer-automated docking search of a protein cavity database. This database is developed from protein structures in the Protein Data Bank (PDB). Docking is conducted by a procedure involving multiple conformer shape-matching alignment of the ligand to the cavity followed by molecular-mechanics torsion optimization and energy minimization on both the ligand and the binding region of the receptor. Scoring is conducted by molecular mechanics energy evaluation and, while applicable, further comparison with binding energetic profile of other ligands that bind to the same receptor site in at least one PDB entry.

Testing Results on several drugs show that 28% to 50% of computer-identified putative protein targets have been confirmed or implicated by experiments.

Application of this approach can potentially facilitate the prediction of unknown and secondary therapeutic target proteins and those related to side effect and toxicity of a drug or drug lead.
 

For more detailed information about general protein target prediction, please check the attached presentation in html format or PDF format.

For more detailed information about toxicity and side effect target prediction, please check the attached presentation in PDF format.