Calculating linear and nonlinear multi-ensemble slow collective variables for protein folding.

Area of Science:

• Computational biology
• Biophysics
• Molecular modeling

Background:

• Traditional molecular dynamics (MD) simulations struggle with conformational sampling, limiting free energy calculations and transition path construction.
• Existing methods for adaptive bias potential rely on collective variables (CVs) often built from insufficient single-temperature data, hindering simulation effectiveness.
• The quality of CVs is critical for efficient sampling in biomolecular simulations.

Purpose of the Study:

• To develop a robust method for constructing high-quality linear and nonlinear slow collective variables (CVs).
• To address the limitations of single-ensemble methods in calculating multi-ensemble averages for CV construction.
• To enhance the efficiency and productivity of enhanced sampling simulations in biomolecular modeling.

Main Methods:

• Applied the standard weighted histogram analysis method (WHAM) to compute multi-ensemble averages.
• Utilized pairs of time-lagged features for the construction of both linear and nonlinear slow CVs.
• Validated the method using simulations of a peptide and a miniprotein.

Main Results:

• The proposed method significantly reduces statistical uncertainties in multi-ensemble averages compared to single-ensemble approaches.
• Generated CVs effectively guided a peptide and a miniprotein towards their near-native states within a reduced simulation time.
• Demonstrated improved sampling efficiency and productivity through the application of the new CV construction technique.

Conclusions:

• The developed method provides a more accurate and efficient way to construct collective variables for enhanced sampling simulations.
• This approach overcomes previous limitations, enabling faster and more reliable exploration of biomolecular conformational landscapes.
• The findings pave the way for more productive molecular dynamics studies in biophysics and computational biology.