REVO: Resampling of ensembles by variation optimization

Area of Science:

• Computational Biology
• Biophysics
• Molecular Dynamics

Background:

• Conventional molecular dynamics (MD) simulations struggle with high-dimensional biomolecular systems and rough energy landscapes, limiting the observation of rare events.
• Enhanced sampling techniques are crucial for calculating transition rates and studying ligand-protein interactions, which often require extensive conformational sampling.
• Existing methods like WExplore face challenges with conformational space regioning, hindering efficient ensemble generation.

Purpose of the Study:

• To develop a novel, regionless enhanced sampling method based on the weighted ensemble framework.
• To improve the exploration of free energy landscapes and the sampling of diverse protein conformations and transition states.
• To enable more consistent measurements of observables in complex biomolecular systems.

Main Methods:

• A new regionless enhanced sampling algorithm is proposed, optimizing "trajectory variation" via cloning and merging operations.
• The method is demonstrated using the N-dimensional random walk and the trypsin-benzamidine unbinding system.
• Conformation space networks were utilized for system analysis.

Main Results:

• The novel method demonstrated efficient exploration of previously inaccessible conformations.
• The residence time of benzamidine in the trypsin-benzamidine system was confirmed.
• A new unbinding pathway for the trypsin-benzamidine system was identified.

Conclusions:

• The proposed regionless enhanced sampling method offers broader sampling and consistent observable measurements.
• Variation optimization in resampling ensembles is a promising general tool for exploring free energy landscapes.
• This approach facilitates the study of rare events and transition rates in complex biomolecular systems.