Grid-based methods for biochemical ab initio quantum chemical applications

Abstract

A initio quantum chemical methods are seeing increased application in a large variety of real-world problems including biomedical applications ranging from drug design to the understanding of environmental mutagens. The vast majority of these quantum chemical methods are {open_quotes}spectral{close_quotes}, that is they describe the charge distribution around the nuclear framework in terms of a fixed analytic basis set. Despite the additional complexity they bring, methods involving grid representations of the electron or solvent charge can provide more efficient schemes for evaluating spectral operators, inexpensive methods for calculating electron correlation, and methods for treating the electrostatic energy of salvation in polar solvents. The advantage of mixed or {open_quotes}pseudospectral{close_quotes} methods is that they allow individual non-linear operators in the partial differential equations, such as coulomb operators, to be calculated in the most appropriate regime. Moreover, these molecular grids can be used to integrate empirical functionals of the electron density. These so-called density functional methods (DFT) are an extremely promising alternative to conventional post-Hartree Fock quantum chemical methods. The introduction of a grid at the molecular solvent-accessible surface allows a very sophisticated treatment of a polarizable continuum solvent model (PCM). Where most PCM approaches use a truncated expansion of the solute`s electric multipole expansion, e.g. net charge (Born model) or dipole moment (Onsager model), such a grid-based boundary-element method (BEM) yields a nearly exact treatment of the solute`s electric field. This report describes the use of both DFT and BEM methods in several biomedical chemical applications.