String method in collective variables: minimum free energy paths and isocommittor surfaces.

Area of Science:

• Computational Chemistry
• Statistical Mechanics
• Biophysics

Background:

• Determining minimum free energy paths is crucial for understanding chemical and biological processes.
• Traditional methods for mapping free energy landscapes become computationally expensive as the number of collective variables increases.
• Accurate identification of transition states and mechanisms is essential for molecular dynamics simulations.

Purpose of the Study:

• To develop a novel computational technique for efficiently calculating minimum free energy paths.
• To overcome the limitations of existing methods that scale poorly with the number of collective variables.
• To enable the determination of reaction mechanisms and transition state regions in complex systems.

Main Methods:

• A hybrid approach combining the string method with a local sampling technique.
• Computation of mean force and conditional expectations along a defined path (string).
• Application to alanine dipeptide isomerization using two and four collective variables (dihedral angles).

Main Results:

• The proposed technique efficiently computes minimum free energy paths.
• It successfully determines the committor function and transition state region when sufficient collective variables are used.
• The method demonstrated its capability in capturing the isomerization mechanism of alanine dipeptide with four dihedral angles, but not with two.

Conclusions:

• The new computational technique offers an efficient alternative for calculating minimum free energy paths, especially in systems with numerous collective variables.
• It accurately captures transition mechanisms, providing insights into reaction pathways.
• The choice of collective variables is critical for accurately describing the transition mechanism, as illustrated by the alanine dipeptide example.