Potential-based dynamical reweighting for Markov state models of protein dynamics

Area of Science:

• Computational biology
• Molecular dynamics simulations
• Biophysics

Background:

• Simulating large cellular components in silico faces challenges in sampling slow molecular dynamics.
• Biased simulations can accelerate sampling but require methods to recover equilibrium properties.

Purpose of the Study:

• To develop and validate a general scheme for extracting equilibrium kinetic properties from biased molecular dynamics trajectories using Markov state models (MSMs).
• To address the limitations of biased simulations in accurately representing molecular system dynamics.

Main Methods:

• Utilizing Markov state models (MSMs) to reweight data from biased potential energy surfaces.
• Validating the reweighting protocol on a simple two-well potential system.
• Testing the method on potential-biased simulations of the Trp-cage miniprotein.

Main Results:

• The reweighting protocol reliably reproduced equilibrium populations and timescales.
• Dynamical processes were accurately recovered from biased simulation data.
• The method's performance was validated against unbiased simulation datasets.

Conclusions:

• The presented scheme effectively harnesses MSMs to extract equilibrium kinetic properties from biased molecular dynamics simulations.
• This approach offers a reliable way to analyze data from biased simulations, overcoming sampling barriers.
• Further applications are suggested for complex molecular systems and biological processes.