CryoFold: determining protein structures and data-guided ensembles from cryo-EM density maps.

Area of Science:

• Structural biology
• Computational biology
• Biophysics

Background:

• Cryo-electron microscopy (EM) structural determination requires molecular modeling for refinement.
• Current ensemble modeling is computationally intensive and limited to small proteins.

Purpose of the Study:

• Introduce CryoFold, a novel pipeline for protein structure determination.
• Integrate cryo-EM density data with protein folding simulations.

Main Methods:

• Utilize molecular dynamics simulations guided by cryo-EM density data (3-5 Å resolution).
• Incorporate coarse-grained topological knowledge of protein folds.
• Employ a Python GUI for controlled data-guided protein folding.

Main Results:

• Successfully determined protein structures for systems ranging from 72 to 2000 residues.
• Identified ensembles of common low-energy models and rare, low-probability structures.
• Demonstrated broad applicability across membrane and multi-domain protein systems.

Conclusions:

• CryoFold enables direct protein structure determination from sequence and sparse density data.
• The method captures protein equilibrium distributions, including functionally relevant rare conformations.
• Established best practices for data-guided protein folding using CryoFold.