A Comprehensive Exploration of Physical and Numerical Parameters in the Poisson–Boltzmann Equation for Applications to Receptor–Ligand Binding

The prediction of ligand-binding affinities to biomolecules, such as proteins and nucleic acids, is a highly sought-after property in computational biophysics. In recent years, the molecular mechanics Poisson–Boltzmann surface area (MM-PBSA) methodology has been widely adopted by the scientific community as an approach to obtain binding free energies. The application of the method to a range of biologically relevant systems can be attributed to its lower computational cost compared to alchemical free energy calculations, and recent developments in the software that make the method fairly easy to use. However, reports have shown that with this increased adoption by the scientific community, the error reported between computational predictions and experimental measurements has also increased. This increase in error may be attributed to incorrect or less rigorous usage of the underlying theories, particularly the PB implicit solvent model. The PB equation can seem trivial, yet in its application to biological systems it has a range of physical and numerical parameters to which its solution is highly sensitive. We comprehensively tested different parameters on three biological systems: Trypanosoma brucei RNA editing ligase–adenosine triphosphate, neuraminidase–oseltamivir, and DNA-netropsin complexes. In this chapter, we highlight the most notable of these parameters and propose practical guidelines for a handful of the most commonly utilized PB solvers.