Conformational sampling techniques

Abstract

The potential energy hyper-surface of a protein relates the potential energy of the protein to its conformational space. This surface is useful in determining the native conformation of a protein or in examining a statistical-mechanical ensemble of structures (canonical ensemble). In determining the potential energy hyper-surface of a protein three aspects must be considered; reducing the degrees of freedom, a method to determine the energy of each conformation and a method to sample the conformational space. For reducing the degrees of freedom the choice of solvent, coarse graining, constraining degrees of freedom and periodic boundary conditions are discussed. The use of quantum mechanics versus molecular mechanics and the choice of force fields are also discussed, as well as the sampling of the conformational space through deterministic and heuristic approaches. Deterministic methods include knowledge-based statistical methods, rotamer libraries, homology modeling, the build-up method, self-consistent electrostatic field, deformation methods, tree-based elimination and eigenvector following routines. The heuristic methods include Monte Carlo chain growing, energy minimizations, metropolis monte carlo and molecular dynamics. In addition, various methods to enhance the conformational search including the deformation or smoothing of the surface, scaling of system parameters, and multi copy searching are also discussed.