Macromolecular density matrices and electron densities with adjustable nuclear geometries

Based on the additive fuzzy electron density fragmentation principle introduced earlier within the ab initio Hartree-Fock quantum chemical computational framework, two new methods are introduced for the construction of geometry-adjustable, ab initio quality macromolecular electron densities. Both methods are designed for the computation of ab initio quality electron densities and other properties for macromolecules of arbitrary reference nuclear geometry, as well as for the rapid computation of approximate electron densities and other molecular properties for nuclear geometries slightly distorted with respect to the reference geometry. This latter feature is expected to improve the description of some of the vibrational and dynamic properties of macromolecules. The first of the two techniques, the Adjustable Local Density Assembler, or ALDA method, generates geometry-adjusted macromolecular electron densities directly, using Mulliken-Mezey fragment density matrices, basis set information, and nuclear coordinates. The method requires an ALDA fragment electron density matrix database. The second technique, the Adjustable Density Matrix Assembler or ADMA method, is introduced for the generation of ab initio quality approximate density matrices for macromolecules. The method assembles Mulliken-Mezey fragment density matrices designed to fulfill a macromolecular compatibility condition. The ADMA method generates macromolecular density matrices without requiring the computation of a macromolecular wavefunction. The ADMA method allows one to apply most of the density matrix techniques of conventional quantum chemistry to macromolecules such as proteins.