A Mode Evolution Metric to Extract Reaction Coordinates for Biomolecular Conformational Transitions.

Area of Science:

• Computational Chemistry
• Biophysics
• Molecular Dynamics

Background:

• Extracting collective variables (CVs) for biomolecular conformational transitions is challenging due to complex energy landscapes.
• Current methods like dimensionality reduction and machine learning (ML) lack clear sampling conditions for reliable CV generation.
• Understanding minimum sampling requirements for accurate CVs describing energy landscapes is crucial.

Purpose of the Study:

• To develop a novel metric, the Mode Evolution Metric (MeM), for extracting reliable CVs from molecular dynamics (MD) and Monte Carlo (MC) trajectories.
• To establish conditions for generating accurate CVs that describe biomolecular conformational transitions and energy landscapes.
• To enable the identification of new states and local transitions near reference minima.

Main Methods:

• Development of the Mode Evolution Metric (MeM) for extracting CVs.
• Mathematical formulation of MeM applicable to statistical dimensionality reduction and ML approaches.
• Application of MeM to MC trajectories of model potentials and MD trajectories of solvated alanine dipeptide.

Main Results:

• MeM successfully extracts CVs that pinpoint new states and local transitions from nonequilibrated MD/MC trajectories.
• Principal components for identifying new states emerge before local equilibration between states.
• MeM facilitates the construction of accurate energy landscape slices linking transitions.

Conclusions:

• MeM provides a robust method for extracting collective variables (CVs) crucial for understanding biomolecular conformational changes.
• The metric accelerates the discovery of new minima and the characterization of transitions between biomolecular states.
• MeM aids in accurate thermodynamic estimations and the description of transitions on complex energy landscapes.