IFPTarget: A Customized Virtual Target Identification Method Based on Protein-Ligand Interaction Fingerprinting...

Area of Science:

• Computational chemistry
• Drug discovery
• Chemical biology

Background:

• Small-molecule target identification is crucial but challenging for drug discovery.
• Current structure-based virtual target identification methods often use universal scoring functions that may lack target-specific binding feature considerations.
• This can lead to inaccurate identification of potential protein targets for a molecule of interest.

Purpose of the Study:

• To develop a customized virtual target identification method (IFPTarget) that enhances accuracy by incorporating target-specific interaction analyses.
• To improve the prioritization of potential protein targets for small molecules.
• To provide a novel in silico tool for drug discovery efforts.

Main Methods:

• IFPTarget utilizes an interaction fingerprinting (IFP) method for detailed, target-specific binding interaction analyses.
• A comprehensive index (Cvalue) is employed for effective target ranking.
• The method was evaluated on a large library of 11,863 protein structures (2842 unique targets).

Main Results:

• The IFP method significantly improved binding pose prediction accuracy.
• Cvalue demonstrated excellent performance in ranking potential targets.
• IFPTarget successfully identified known targets and discovered novel potential targets for various drugs within the top-ranked list.
• The method identified metallo-β-lactamase VIM-2 as a novel target for quercetin, confirmed by enzymatic assays.

Conclusions:

• IFPTarget offers a superior in silico approach for small-molecule target identification compared to existing methods.
• The developed tool enhances the accuracy of predicting drug-target interactions.
• This study paves the way for developing more sophisticated, target-customized computational methods in drug discovery.