Stochastic path approach to compute atomically detailed trajectories: Application to the folding of C peptide

A novel method to compute long-time molecular dynamics trajectories is empl oyed to study the folding kinetics of C peptide. The computational method m akes it possible to use a time step larger by orders of magnitude compared to widely used molecular dynamics integrators. Rather than solving the traj ectory in small time steps, the whole trajectory is optimized. The algorith m filters high-frequency modes that are modeled as Gaussian noise. The assu mption of "Gaussian noise" is tested numerically in two cases and found to be adequate. In all, 31 trajectories of C peptide that folds into a helix i n explicit solvent (TIP3P water molecules) are computed. The time step is 5 00 ps. The folding pathways and the early formation of structure are discus sed. Comparisons to a 2-ns trajectory calculated with the usual molecular d ynamics approach and to available experimental data are made.