MDexciteR: Enhanced Sampling Molecular Dynamics by Excited Normal Modes or Principal Components Obtained from...

Area of Science:

• Computational biology
• Biophysics
• Molecular modeling

Background:

• Enhanced sampling methods are crucial for exploring protein conformational dynamics.
• Molecular dynamics with excited normal modes (MDeNM) offers a low-bias approach to study protein motions.
• Existing MDeNM implementations require significant user expertise and lack broad accessibility.

Purpose of the Study:

• To develop and validate an automated, open-source implementation of MDeNM.
• To generalize MDeNM for incorporating experimental data via principal components.
• To assess the utility of coarse-grained normal modes in MDeNM simulations.

Main Methods:

• Development of MDexciteR, an automated MDeNM implementation compatible with GPU-accelerated molecular dynamics engines.
• Generalization of MDeNM to utilize principal components derived from experimental ensembles.
• Evaluation of coarse-grained normal modes from elastic network models for MDeNM.
• Testing the method on globular proteins and a protein-membrane system.

Main Results:

• MDexciteR successfully integrates MDeNM with GPU-enabled simulation packages.
• The generalized MDeNM approach accurately captures conformational changes using experimental data.
• Coarse-grained normal modes provide a computationally efficient alternative without significant loss of accuracy.
• Performance and limitations were characterized across diverse protein systems.

Conclusions:

• MDexciteR enhances the accessibility and applicability of MDeNM for protein dynamics studies.
• The generalized MDeNM approach facilitates the integration of experimental and simulation data.
• Coarse-grained MDeNM offers a viable strategy for large-scale conformational sampling.
• The study provides insights into the strengths and weaknesses of these advanced MDeNM techniques.