Tutorial on how to build non-Markovian dynamic models from molecular dynamics simulations for studying protein...

Area of Science:

• Computational Biology
• Biophysics
• Molecular Dynamics Simulations

Background:

• Protein conformational changes are vital for biological functions.
• Markov State Models (MSMs) are widely used to model these changes from molecular dynamics (MD) simulations.
• A key limitation of MSMs is the requirement for a sufficiently long lag time to ensure Markovian dynamics, often constrained by simulation length.

Purpose of the Study:

• To introduce Generalized Master Equation (GME)-based approaches for modeling non-Markovian protein dynamics.
• To present two novel GME-based models: quasi-Markov State Model (qMSM) and Integrative Generalized Master Equation (IGME).
• To provide a practical tutorial for applying qMSM and IGME to biomolecular systems.

Main Methods:

• Development of Generalized Master Equation (GME)-based models encoding non-Markovian dynamics with a time-dependent memory kernel.
• Construction and application of the quasi-Markov State Model (qMSM).
• Construction and application of the Integrative Generalized Master Equation (IGME).

Main Results:

• Demonstrated the ability of GME-based approaches to capture non-Markovian dynamics, overcoming MSM lag time limitations.
• Successfully applied qMSM and IGME to model the conformational dynamics of alanine dipeptide and villin headpiece.
• Provided accessible protocols for implementing these advanced non-Markovian models.

Conclusions:

• GME-based models like qMSM and IGME offer a powerful alternative to MSMs for studying complex protein dynamics.
• These methods enhance the accuracy and feasibility of analyzing biomolecular motions from MD simulations.
• The tutorial facilitates broader adoption of non-Markovian modeling techniques by researchers.