Fast Atomistic Molecular Dynamics Simulations from Essential Dynamics Samplings.

We present a new method for fast molecular dynamics simulations in cases where the new trajectories can be considered a perturbation or a combination of previously stored ones. The method is designed for the postgenomic scenario, where databases such as MoDEL will store curated equilibrium trajectories of all biomolecules (proteins, nucleic acids, etc.) of human interest. We demonstrate that the approach outlined here can, with accuracy and great computational efficiency, reproduce and extend original trajectories, describe dynamical effects due to perturbations (e.g., protein-ligand and protein-protein interactions and protein mutations) and predict the dynamics of large polymeric systems built up from previously studied fragments. The method can work simultaneously with low- and high-resolution pictures of the macromolecule, allowing the level of detail to be matched to that required for obtaining the information of biological interest.