PrepFlow: A Toolkit for Chemical Library Preparation and Management for Virtual Screening

Area of Science:

• Computational Chemistry
• Cheminformatics
• Drug Discovery

Background:

• The increasing volume of chemical data necessitates automated solutions for efficient virtual screening.
• Existing methods for chemical library preparation can be time-consuming and resource-intensive.

Purpose of the Study:

• To introduce PrepFlow, a novel toolkit for the automated preparation and management of chemical libraries.
• To demonstrate the speed, robustness, and efficiency of PrepFlow in generating 3D molecular structures.

Main Methods:

• PrepFlow processes chemical compounds from SMILES or 2D formats into 3D structures.
• The toolkit is designed for efficient High-Performance Computing (HPC) resource utilization.
• An archiving strategy is implemented to optimize computational time, storage, and human effort.

Main Results:

• PrepFlow achieves a preparation time of 6.6 seconds per ligand on a desktop computer.
• A chemical library of 2 million compounds (ChEMBL size) was prepared in approximately 3 days using HPC parallelization.
• The toolkit ensures speed, robustness, and efficient resource management.

Conclusions:

• PrepFlow significantly enhances the efficiency of chemical library preparation for drug discovery.
• The toolkit's performance on large datasets demonstrates its scalability and utility.
• PrepFlow is freely available, promoting wider adoption in computational chemistry research.