(589e) Large Scale Free Energy and Potential of Mean Force Calculations through Distributed Computing

Recent developments in computational processing power and in theoretical methods are making simulations come increasingly closer to predicting ligand binding affinities and computing accurate, high dimensional potentials of mean force.

This level of precision makes it possible to separate errors caused by poor sampling from the inadequacies of molecular mechanics force fields. These improvements are beginning to change the process of designing drugs with high affinities and specific action from a trial-and-error art to a nanoscale engineering process. However, current computational methods are inefficient and converge relatively slowly, requiring computational resources that far exceed what is currently cost-effective in the pharmaceutical industry. Improvements in methods and increases in available computational power will be necessary to make such calculations feasible in industrial practice.

We present extended ensemble ligand binding simulations for the FKBP-12 system using large scale distributed computing to reach precisions near 1 kcal/mol. Simulations were performed for both AMBER/GAFF and OPLS-AA, two different common force fields used to model biomolecules, in order to understand how much differences in force field can affect results of free energies of binding. Additionally, extended ensemble methods can significantly increase the speed at which simulations sample macromolecular conformations by including Monte Carlo transitions between nonphysical states, and provide some hope for improving efficiency. We will also present data on the comparison between extended ensemble and more traditional free energy calculations.

We also recently presented a minimum variance method to calculate free energies and ensemble averages from multiple equilibrium simulations conducted at different thermodynamic states [1]. This method, based on statistical techniques for correcting biases in sampling, is useful for analyzing both simulations of phase changes and the results of single molecule experiments. This estimator gives the same results as multiple histogram methods [2,3] in the limit of vanishingly small histogram bins, but does not actually require histograms, eliminating bias due to binning.

To solve the problem of efficiently computing averages and free energies with large numbers of states, we derive a modification of the multistate minimum variance method that uses only states with sufficient mutual phase space overlap. For many cases of interest, this drastically decreases the number of states that must be considered with negligible loss of precision.

We demonstrate how these methods can be used to calculate 3-dimensional potentials of mean force from data collected from distributed computing simulations of the ribosome very efficiently.

[1] M. R. Shirts and J. D. Chodera, J. Chem. Phys.129, 124105 (2008)

[2] A. M. Ferrenberg and R. H. Swendsen, Phys. Rev. Lett. 63:1195-1198 (1989)

[3] S. Kumar, D. Bouzida, R. H. Swendsen, P. A. Kollman and J. M. J. Rosenberg, J. Comput. Chem., 13:1011-1021 (1992)