COFFEE-PRESC: A Fast Prescreening Method Using Compound Retrieval by Pairwise Positional Relationship of...

Area of Science:

• Computational Chemistry
• Drug Discovery
• Bioinformatics

Background:

• Protein-ligand docking is crucial for structure-based virtual screening in early drug discovery.
• Exhaustive screening of billions of compounds is computationally infeasible due to long calculation times per compound.

Purpose of the Study:

• To introduce COFFEE-PRESC (COmpound Filtering by Fragment pair-based Efficient Evaluation for PRESCreening), a novel, rapid prescreening method.
• To enable efficient computational screening of ultralarge compound libraries.

Main Methods:

• COFFEE-PRESC employs a fragment-based approach, docking representative fragments to the target protein.
• It enumerates favorable protein-fragment poses and pairs them to assess spatial relationships.
• Compounds with structures similar to fragment pairs are retrieved via similarity searches, ensuring no spatial collisions.

Main Results:

• COFFEE-PRESC demonstrated a 32-fold increase in speed compared to existing tools like Spresso.
• The method achieved higher accuracy than Spresso in prescreening ultralarge compound libraries.
• The fragment set effectively covers a large and diverse chemical space.

Conclusions:

• COFFEE-PRESC offers a computationally efficient and accurate solution for prescreening ultralarge compound libraries.
• Its speed and accuracy make it a valuable tool for accelerating drug discovery pipelines.
• The developed code is publicly available under an MIT license.