An open-source drug discovery platform enables ultra-large virtual screens

Area of Science:

• Computational chemistry and drug discovery
• Bioinformatics and cheminformatics
• Molecular modeling and simulation

Background:

• Drug development is lengthy and costly, averaging over 10 years and US$2-3 billion.
• High attrition rates in preclinical and clinical phases, coupled with expensive lab experiments, hinder drug discovery.
• Structure-based virtual screening offers a promising approach to improve hit quality and reduce development timelines.

Purpose of the Study:

• To introduce VirtualFlow, a novel, automated, and scalable open-source platform for preparing and screening ultra-large compound libraries.
• To demonstrate the platform's capability in handling massive datasets and diverse docking programs.
• To showcase the potential of VirtualFlow in identifying high-affinity binders for therapeutic targets.

Main Methods:

• Development of VirtualFlow, a highly automated and versatile open-source platform for virtual screening.
• Preparation of a large, ready-to-dock ligand library exceeding 1.4 billion commercially available molecules.
• Screening of over 1 billion compounds using VirtualFlow and various docking programs.

Main Results:

• Successful preparation and screening of an ultra-large compound library (over 1.4 billion molecules).
• Identification of structurally diverse molecules with submicromolar affinity for KEAP1.
• Discovery of a lead inhibitor, iKeap1, with nanomolar affinity (114 nM Kd) that disrupts the KEAP1-NRF2 interaction.

Conclusions:

• VirtualFlow provides an efficient, accessible, and scalable solution for large-scale structure-based virtual screening.
• The platform enables exploration of vast chemical spaces to identify high-affinity drug candidates.
• VirtualFlow has the potential to significantly accelerate drug discovery and reduce associated costs.