A potential of mean force estimator based on nonequilibrium work exponential averages

Area of Science:

• Statistical mechanics
• Computational biophysics
• Physical chemistry

Background:

• Estimating free energy landscapes is crucial for understanding molecular processes.
• Traditional methods often require slow, equilibrium simulations or complex protocols.
• Out-of-equilibrium methods offer faster alternatives but face challenges with accuracy.

Purpose of the Study:

• To develop a robust potential of mean force (PMF) estimator using non-equilibrium work measurements.
• To improve the accuracy of free energy calculations, especially under fast dynamic conditions.
• To integrate Bennett acceptance ratio (BAR) principles for enhanced free energy estimation.

Main Methods:

• Utilizing work exponential averages from forward and backward non-equilibrium processes.
• Estimating the free energy difference between end states using the Bennett acceptance ratio method.
• Applying the developed estimator to both a deterministic peptide folding model and a stochastic Langevin dynamics system.

Main Results:

• The proposed PMF estimator provides accurate free energy profiles, outperforming unidirectional methods.
• The estimator demonstrates robustness even at high pulling velocities of the control parameter.
• Performance was validated against established methods like the Jarzynski equality and other bidirectional estimators.

Conclusions:

• The novel PMF estimator offers a reliable and efficient approach for calculating free energy landscapes from non-equilibrium simulations.
• This method enhances the utility of work measurements in biophysical and chemical studies.
• The integration of BAR principles significantly improves the accuracy of free energy estimations.