CONSTANT-TEMPERATURE FREE-ENERGY SURFACES FOR PHYSICAL AND CHEMICAL PROCESSES

A comparison between two methods for combining data from molecular simulations to generate free energy surfaces for physical and chemical processes is presented. The conventional ''splicing'' method is compared with a recently introduced histogram-based technique for the calculation of the free energy surface for folding a short peptide in vacuo. We present an extension of the weighted histogram analysis method, described recently by Kumar et al. (J. Comput. Chem. 1992, 13, 1101-102 1), which is specialized to the case of constant-temperature simulations and generalized to multiple dimensions. Our data indicate that both methods produce identical results, at the qualitative level, for the folding free energy surface. However, the constant-temperature weighted histogram analysis equations we present here yield a smoother surface, incorporate all of the data from the simulations, and are generalized to multiple dimensions. We propose that these equations will provide a general framework for the calculation of free energy surfaces in multiple dimensions.