Self-guided molecular dynamics simulation based on concerted movement.

Abstract

Self-guided molecular/Langevin dynamics (SGMD/SGLD) simulation methods were developed to enhance conformational sampling through promoting low frequency motion of molecular systems and have been successfully applied in many simulation studies. The low frequency properties are derived effectively from local averages over the most recent trajectories. Using low frequency momentums and low frequency forces, SGMD/SGLD achieve accelerated diffusion and enhanced energy barrier crossing. In addition to the local frequency motion, we find concerted movement is important for rare event like protein folding and assembly. This work presents a new self-guided simulation method that accelerates conformational search through enhancing concerted movement. We abbreviate this method as SGMDc or SGLDc. Concerted motion is derived from spatial averages over chemical bonded atoms and/or nearby atoms. Combined with the local average scheme, one can derive concerted low frequency properties and use them as guiding forces to achieve accelerated conformational evolvement. Using peptide folding and pore formation in membrane, we demonstrate that SGMDc/SGLDc can accelerate rare events significantly.