An Adaptive Branch-and-Bound Minimization Method Based on Dynamic Programming

Abstract

Computational studies are potentially useful for analyzing various putative mechanisms of protein folding. The general question that can be addressed is whether or not it is possible to obtain native-like conformations of a protein on the basis of a well-defined set of physical and chemical principles. In particular, computation is the only tool available for studying the relations between native protein structures and the conformations that correspond to various free energy minima (Skolnick and Kolinski, 1989; Chan and Dill, 1993). If the “thermodynamic hypothesis” is valid and the free-energy evaluation is of acceptable accuracy, then the native structure can be identified in principle by systematic evaluation of the energy of every possible conformation. However, using a detailed model of protein geometry and a force field of near-atomic resolution, we are at present unable to explore the conformational space even for the smallest proteins.KeywordsDynamic ProgrammingGlobal MinimumOptimal PathConfiguration SpaceConformational StateThese keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.