Free energy calculations: an efficient adaptive biasing potential method.

Area of Science:

• Computational Chemistry
• Statistical Mechanics

Background:

• Free energy calculations are crucial for understanding molecular systems.
• Adaptive biasing potential (ABP) methods offer efficient sampling but can suffer from long equilibration times.

Purpose of the Study:

• To develop a more efficient sampling and free energy calculation technique within the adaptive biasing potential (ABP) framework.
• To reduce the equilibration time of adaptive bias potentials.

Main Methods:

• Developed a method using mollified density of states to approximate free energy and bias potential.
• The bias potential's gradient has a simple analytic expression, allowing updates from single observations.
• Errors from mollification can be removed via deconvolution.

Main Results:

• The mollified density of states method significantly reduces equilibration time.
• Achieved a factor of 10 increase in efficiency compared to basic ABP implementations for the alanine dipeptide test case.
• Demonstrated efficiency comparable to well-tempered metadynamics, with post-processing deconvolution offering an advantage.

Conclusions:

• The new method is simple to apply to multi-dimensional free energy and mean force computations.
• It avoids complex procedures like second derivative calculations or matrix manipulations.
• This approach offers a computationally efficient and accurate alternative for free energy calculations.