Non-Markov-Type Analysis and Diffusion Map Analysis for Molecular Dynamics Trajectory of Chignolin at a High Temperature

Life (Basel). 2022 Aug 3;12(8):1188. doi: 10.3390/life12081188.

Abstract

We apply the non-Markov-type analysis of state-to-state transitions to nearly microsecond molecular dynamics (MD) simulation data at a folding temperature of a small artificial protein, chignolin, and we found that the time scales obtained are consistent with our previous result using the weighted ensemble simulations, which is a general path-sampling method to extract the kinetic properties of molecules. Previously, we also applied diffusion map (DM) analysis, which is one of a manifold of learning techniques, to the same trajectory of chignolin in order to cluster the conformational states and found that DM and relaxation mode analysis give similar results for the eigenvectors. In this paper, we divide the same trajectory into shorter pieces and further apply DM to such short-length trajectories to investigate how the obtained eigenvectors are useful to characterize the conformational change of chignolin.

Keywords: Markov state model; diffusion map; molecular dynamics simulation; non-Markov-type analysis; rare event; weighted ensemble simulation.

Grants and funding

This research was partially supported by the Japan Society for the Promotion of Science (KAKEN 16K00059, 17KT0101, 22K11941 to H.F., 25120011 to H.S., and 20H03230, 22H04756 to A.M.) and AMED-CREST, AMED (JP20gm0810012) to H.F.