WExplore: hierarchical exploration of high-dimensional spaces using the weighted ensemble algorithm.

Area of Science:

• Biophysics
• Computational Biology
• Biochemistry

Background:

• Conventional molecular dynamics simulations struggle with rare events in biomolecular systems.
• Enhanced sampling algorithms are crucial for studying intrinsically disordered systems and protein conformational ensembles.
• Existing methods often require prior knowledge of system dynamics and are limited to two-basin problems.

Purpose of the Study:

• To develop a novel, efficient enhanced sampling algorithm for determining structural ensembles.
• To address limitations of existing methods, particularly for poorly-studied systems and multi-basin problems.
• To enable the study of protein conformational ensembles for drug development.

Main Methods:

• A novel strategy based on the weighted ensemble algorithm with dynamically defined, hierarchically organized sampling regions.
• Utilizes merging and cloning operations to direct an ensemble of replicas through configuration space.
• Employs a sampling hierarchy to manage a large number of regions with a small number of replicas, balanced across length scales.

Main Results:

• Demonstrated the algorithm's effectiveness on two analytically solvable model systems.
• Applied the method to the 10-residue peptide chignolin in explicit solvent.
• Identified novel hydrogen bonds facilitating folding through configuration space network analysis.

Conclusions:

• The novel hierarchical weighted ensemble algorithm efficiently determines structural ensembles for biomolecular systems.
• This method overcomes limitations of traditional approaches, enabling studies of complex systems with minimal prior information.
• The findings provide new insights into peptide folding mechanisms and offer a powerful tool for drug discovery.