The computational prediction of protein-ligand binding affinities has
become a key step in the successful virtual screening of compounds for drug
development and discovery. However, consistently accurate protein-ligand binding
affinity calculations are challenging in part due to, 1) the large protein/ligand
conformational space that must be sampled/searched, 2) the inconsistent accuracy of
classical molecular mechanics potentials, commonly used to compute binding
affinities, especially when π-stacking, halogen interaction, or metal centers are present,
or when polarization or charge transfer is significant. In this chapter, recent advances in
quantum mechanical methods that facilitate their application to protein-ligand binding
free energy calculations are discussed, with an emphasis on fragmentation methods and
their combination with conformational search algorithms. The accuracy of these new
approaches with respect to the prediction of protein-ligand binding free energy is
evaluated. New tools to improve workflow and speed up calculations are also
discussed.
Keywords: Ab initio quantum mechanics, Drug design, Entropy, Fragmentation methods, Mining minima, Protein-ligand binding, QM-VM2, Solvation.
