Simulations and active learning enable efficient identification of an experimentally-validated broad coronavirus...

Area of Science:

• Computational Chemistry
• Drug Discovery
• Molecular Biology

Background:

• Traditional drug screening is costly and time-consuming.
• Virtual screening methods face challenges in balancing computational efficiency with experimental accuracy.
• Identifying effective drug candidates from large compound libraries remains a significant hurdle in pharmaceutical research.

Purpose of the Study:

• To develop an integrated computational framework for efficient drug screening.
• To significantly reduce the number of compounds requiring experimental validation.
• To identify novel inhibitors for targets like TMPRSS2.

Main Methods:

• Integration of molecular dynamics (MD) simulations with active learning algorithms.
• Development of a target-specific scoring function to evaluate inhibition potential.
• Generation of a comprehensive receptor ensemble using extensive MD simulations.

Main Results:

• The framework drastically reduced experimental testing candidates to fewer than 10.
• Computational costs were reduced approximately 29-fold.
• Discovered BMS-262084 as a potent TMPRSS2 inhibitor (IC50 = 1.82 nM).
• Confirmed BMS-262084 efficacy against SARS-CoV-2 variants in cell-based assays.

Conclusions:

• The combined MD simulations and active learning approach offers a highly efficient drug screening strategy.
• BMS-262084 demonstrates significant potential as a therapeutic agent for viral infections involving TMPRSS2.
• This framework accelerates the identification of promising drug candidates for various diseases.