Computer simulations of protein folding: Classical trajectories by optimization of action

The vast range of time scales (from nanoseconds to seconds) during protein folding is a challenge for experiments and computations. To make concrete predictions of folding mechanisms, atomically detailed simulations of protein folding using potentials derived from chemical physics principles, are desired. However, due to their computational complexity, straightforward molecular dynamics simulations of protein folding are impossible using today's technology. An alternative algorithm for atomically detailed simulations is discussed that makes it possible to compute approximate atomically detailed long time trajectories (SDEL-the Stochastic Difference Equation in Length). This algorithm is used to compute folding trajectories of a helical peptide, protein A, and the protein cytochrome c. A brief account of the methodology, the results, and comparison to experiments are provided.