Principal component analysis of nonequilibrium molecular dynamics simulations

Area of Science:

• Computational Chemistry and Molecular Dynamics
• Statistical Mechanics
• Biophysics

Background:

• Principal Component Analysis (PCA) is standard for analyzing equilibrium molecular dynamics (MD) simulations to build free energy landscapes.
• Extending PCA to nonequilibrium simulations presents challenges in defining statistical averages and relating equilibrium to nonequilibrium energy landscapes.

Purpose of the Study:

• To develop a well-defined method for performing PCA on nonequilibrium molecular dynamics data.
• To establish a direct relation between equilibrium and nonequilibrium data using a weighting function.
• To demonstrate how nonequilibrium energy landscapes can reveal molecular reaction mechanisms.

Main Methods:

• Utilized targeted MD simulations with a moving distance constraint along a biasing coordinate (s).
• Introduced a weighting function P(s) to relate equilibrium and nonequilibrium data.
• Applied PCA to backbone dihedral angles of decaalanine unfolding simulations.

Main Results:

• Developed a method to perform PCA on nonequilibrium MD data, enabling the construction of energy landscapes.
• Demonstrated that the nonequilibrium energy landscape can reveal molecular reaction mechanisms.
• Successfully discriminated multiple unfolding pathways of decaalanine using PCA on dihedral angles.

Conclusions:

• The proposed PCA approach provides a robust way to analyze nonequilibrium MD simulations.
• Nonequilibrium energy landscapes offer insights into molecular mechanisms and reaction pathways.
• Effective application of this method critically depends on the appropriate choice of the biasing coordinate.