A Computational Method for Unveiling the Target Promiscuity of Pharmacologically Active Compounds

Area of Science:

• Computational chemistry
• Pharmacology
• Machine learning

Background:

• Drug discovery traditionally seeks selective ligands, but many drugs interact with multiple macromolecular targets.
• Understanding polypharmacology is crucial for designing future medicines and avoiding off-target effects.

Purpose of the Study:

• To present a straightforward computational method for identifying novel targets of known drugs.
• To validate the method's efficacy in discovering previously unknown targets for specific compounds.

Main Methods:

• Development of a computational model to predict drug-target interactions.
• Application of the model to analyze known compounds, including resveratrol and celecoxib.
• Experimental validation of predicted novel targets.

Main Results:

• The computational method successfully identified previously unknown macromolecular targets for resveratrol and celecoxib.
• Validation experiments confirmed these novel drug-target interactions.
• Demonstrated the utility of machine learning in uncovering polypharmacology.

Conclusions:

• The developed computational method is effective for discovering new drug targets.
• Machine learning facilitates polypharmacology-based drug design and drug repurposing.
• The approach aids in understanding and assigning phenotypic drug effects to specific targets.