Nonparametric Determination of the Committor in Multimolecular Systems

Area of Science:

• Computational Chemistry
• Statistical Mechanics
• Machine Learning

Background:

• Analyzing large longitudinal datasets for multimolecular dynamics is challenging.
• Determining free-energy landscapes requires identifying the optimal reaction coordinate, known as the committor.

Purpose of the Study:

• To develop an effective method for determining the committor in systems with anisotropic interactions.
• To model multimolecular aggregation dynamics using the committor.

Main Methods:

• Combined a nonparametric approach with a systematic method for generating permutationally invariant collective variables.
• Utilized a stringent validation test to verify the optimality of the committor.

Main Results:

• Successfully determined the committor for multimolecular aggregation.
• Demonstrated that a diffusive model along the committor accurately reproduces kinetic properties.
• Validated the committor's optimality through rigorous testing.

Conclusions:

• The developed method effectively determines the committor for analyzing complex molecular dynamics.
• This approach is general and applicable to the machine learning community for analyzing longitudinal data.