BAR-based optimum adaptive steered MD for configurational sampling

Area of Science:

• Computational Chemistry
• Molecular Dynamics
• Statistical Mechanics

Background:

• Adaptive alchemical free energy simulations, including optimum Bennett's acceptance ratio and optimum Crooks' equation (OCE), utilize statistically optimal bidirectional reweighting.
• These methods determine state importance via the time-derivative of variance, offering speedups over the equal time rule in alchemical transformations.
• Traditional steered molecular dynamics (SMD) relies on biasing collective variables (CVs) to transition between conformational states.

Purpose of the Study:

• To extend the time-derivative of variance guided adaptive sampling method to configurational space within steered molecular dynamics (SMD).
• To introduce a novel adaptive SMD (ASMD) approach that minimizes free energy differences variance along the pathway.
• To evaluate the efficiency of ASMD compared to traditional SMD and criterion-based ASMD schemes.

Main Methods:

• Developed an adaptive SMD (ASMD) method by minimizing the variance of free energy differences along a pathway.
• Employed physically meaningful collective variables (CVs) such as dihedrals or distances to guide the steered simulations.
• Tested the ASMD method on dihedral flipping and distance pulling cases, considering both periodic and nonperiodic CVs.

Main Results:

• The proposed ASMD method demonstrates superior performance over traditional equal-time SMD in nonequilibrium stratification.
• ASMD provides more efficient potential of mean force calculations compared to selection criterion-based ASMD schemes.
• Unlike OCE, the ASMD workflow is independent of the periodicity of the collective variable.

Conclusions:

• Adaptive SMD (ASMD) offers a more efficient and robust approach for free energy calculations in molecular dynamics.
• The method's independence from CV periodicity broadens its applicability to various molecular systems.
• ASMD significantly outperforms traditional SMD and offers advantages over existing adaptive sampling techniques.