Combining coarse-grained nonbonded and atomistic bonded interactions for protein modeling

Area of Science:

• Computational Biology
• Biophysics
• Molecular Modeling

Background:

• Accurate protein simulations are crucial for understanding biological processes.
• Existing hybrid models often treat spatially distinct protein regions separately.
• A unified approach is needed for efficient and accurate protein complex analysis.

Purpose of the Study:

• To develop a novel hybrid coarse-grained (CG) and atomistic (AT) model for protein simulations.
• To enable rapid searching and refinement of peptide-protein complexes.
• To improve the accuracy of protein structure representation.

Main Methods:

• Simultaneously employing AT (united atom) and CG models for protein representation.
• Describing protein main chain interactions using a united atom force field for realistic secondary structures.
• Calculating nonbonded interactions with a CG model and knowledge-based potential for side chains.

Main Results:

• Unrestrained molecular dynamics simulations yielded protein trajectories consistent with experimental structures.
• Refinement of docked peptide-protein complexes led to improved structural accuracy.
• The model shows potential for rapid refinement of docked protein-protein complexes, pending further parameter optimization.

Conclusions:

• The developed hybrid CG/AT model offers a robust method for protein simulations.
• This approach effectively refines peptide-protein complex structures.
• Further development may extend its utility to protein-protein complex refinement.