Adaptively Iterative Multiscale Switching Simulation Strategy and Applications to Protein Folding and Structure...

Area of Science:

• Computational biology
• Biophysics
• Structural biology

Background:

• Protein structure prediction is crucial for understanding protein function.
• Current molecular dynamics simulations face challenges with computational cost (all-atom models) and accuracy (coarse-grained models), especially for proteins lacking templates.
• Predicting structures for proteins without detectable templates remains a significant challenge.

Purpose of the Study:

• To develop a novel, universal multiscale simulation strategy for protein structure prediction.
• To overcome the limitations of existing all-atom and coarse-grained simulation models.
• To improve the efficiency and accuracy of predicting protein structures, particularly for novel proteins.

Main Methods:

• Proposed a universal multiscale simulation strategy named Adaptive Integrated Multiscale Simulation (AIMS).
• AIMS iteratively switches between multiple resolutions to balance accuracy and computational efficiency.
• Employed coarse-grained (CG)-guided enhanced sampling to ensure all-atom (AA) accuracy in final results.

Main Results:

• Successfully performed four ab initio and four data-assisted protein structure predictions using AIMS.
• The AIMS strategy achieved computational speeds approximately 40 times faster than conventional AA simulations.
• The prediction results provided insights into the folding of metastable protein states, yielding an ensemble rather than a single structure.

Conclusions:

• AIMS offers a significant advancement in protein structure prediction, particularly for proteins without templates.
• The multiscale approach effectively combines the accuracy of AA models with the efficiency of CG models.
• AIMS provides a powerful tool for exploring protein folding dynamics and understanding protein function more rapidly.