Uni-GBSA: an open-source and web-based automatic workflow to perform MM/GB(PB)SA calculations for virtual screening

Area of Science:

• Computational Chemistry
• Drug Discovery
• Molecular Modeling

Background:

• Accurate binding free energy calculations are crucial for drug discovery.
• Molecular mechanics/Generalized-Born (Poisson-Boltzmann) surface area (MM/GB(PB)SA) offers a balance of accuracy and efficiency.
• Existing MM/GB(PB)SA tools present limitations and high user entry barriers.

Purpose of the Study:

• To introduce Uni-GBSA, a user-friendly, automated workflow for MM/GB(PB)SA calculations.
• To provide an efficient tool for virtual screening and binding free energy estimation.
• To overcome the limitations and high entry barriers of current MM/GB(PB)SA software.

Main Methods:

• Uni-GBSA automates topology preparation, structure optimization, and binding free energy calculations.
• It incorporates a batch mode for parallel evaluation of thousands of molecules against protein targets.
• Default parameters were optimized using the PDBBind-2011 refined dataset.

Main Results:

• Uni-GBSA demonstrated a satisfactory correlation with experimental binding affinities.
• The workflow outperformed AutoDock Vina in molecular enrichment during virtual screening.
• Systematic testing ensured reliable default parameter settings.

Conclusions:

• Uni-GBSA provides an accessible and efficient solution for MM/GB(PB)SA calculations in drug discovery.
• The open-source package and web server enhance user-friendliness and accessibility.
• Uni-GBSA facilitates large-scale virtual screening and accelerates the identification of lead compounds.