Structural dissimilarity sampling with dynamically self-guiding selection.

Area of Science:

• Computational Biology
• Biophysics
• Molecular Dynamics

Background:

• Protein conformational transitions are crucial for biological function.
• Traditional molecular dynamics (MD) simulations often struggle to capture rare events like large conformational changes.
• Structural Dissimilarity Sampling (SDS) offers an enhanced approach to conformational sampling.

Purpose of the Study:

• To develop an efficient, dynamically self-guiding measure to accelerate protein structural transitions.
• To extend the original SDS method for improved conformational sampling.
• To validate the efficacy of the extended SDS in reproducing protein state transitions.

Main Methods:

• Proposed a novel, dynamically self-guiding selection measure based on the inner product (IP) between reactant states and simulation snapshots.
• Implemented an extended SDS protocol utilizing IP ranking to select dissimilar initial structures for short-time MD simulations.
• Applied the extended SDS to study the conformational transition of maltodextrin binding protein (MBP) from open to closed states.

Main Results:

• The extended SDS successfully reproduced the open-to-closed state transition of MBP within submicrosecond simulation times.
• Conventional long-time MD simulations failed to capture this specific protein transition.
• The extended SDS demonstrated superior efficiency compared to ordinary SDS and other sampling techniques.

Conclusions:

• The proposed dynamically self-guiding SDS extension significantly accelerates the sampling of protein conformational transitions.
• This enhanced method provides a powerful tool for studying rare events in protein dynamics.
• The extended SDS shows considerable potential for practical applications in computational structural biology.