Backmapping from Multiresolution Coarse-Grained Models to Atomic Structures of Large Biomolecules by Restrained...

Area of Science:

• Computational Biology
• Biophysics
• Molecular Modeling

Background:

• Coarse-grained (CG) simulations offer enhanced scale for biomolecular dynamics compared to all-atom (AA) molecular dynamics (MD).
• Backmapping CG models to AA structures is crucial for detailed insights but existing methods struggle with non-residue-level CG models and deviations.

Purpose of the Study:

• To develop a new, robust backmapping method applicable to various CG models, including those with low resolution.
• To accurately reconstruct all-atom (AA) structures from coarse-grained (CG) models by incorporating CG model deviations.

Main Methods:

• A novel backmapping technique employing Bayesian inference and restrained MD simulations.
• Definition of log harmonic energy restraints based on target CG models and Bayesian inference to estimate CG deviations.
• Generation of AA structures from initial AA models (experimental or homology-derived) using MD with defined restraints.

Main Results:

• The method successfully generated accurate AA structures from multiresolution CG models.
• Validation was performed on the epidermal growth factor receptor and nucleosome core particle.
• The approach proved effective for CG models with residue-level and lower resolutions.

Conclusions:

• The developed Bayesian inference and restrained MD method provides accurate backmapping from diverse CG models to AA structures.
• This technique enhances the utility of CG simulations by enabling detailed structural analysis of biomolecules.
• The method's ability to handle low-resolution CG models broadens its applicability in molecular modeling.