Building Protein Atomic Models from Cryo-EM Density Maps and Residue Co-Evolution

Area of Science:

• Structural Biology
• Biophysics
• Computational Biology

Background:

• Electron cryo-microscopy (cryo-EM) provides high-resolution 3D structures of macromolecular complexes.
• De novo atomic model building from cryo-EM maps is challenging due to poor side-chain density and subunit segmentation issues.

Purpose of the Study:

• To develop a computational method for accurate de novo atomic model building from cryo-EM density maps.
• To address limitations in sequence assignment and subunit segmentation for lower-resolution cryo-EM data.

Main Methods:

• A graph-based approach was developed to thread C-alpha traces into cryo-EM density maps.
• Minimum spanning trees and pruning algorithms were used to identify C-alpha atom positions.
• Sequence co-evolutionary information was integrated with experimental maps.

Main Results:

• The method successfully segments EM density maps into individual subunits.
• Amino acid sequences were accurately assigned to backbone traces.
• Atomic models were generated from near-atomic resolution cryo-EM data.

Conclusions:

• The developed graph-based method significantly improves de novo atomic model building from cryo-EM data.
• This approach enhances the utility of cryo-EM for structural determination, especially at lower resolutions.
• Integration with co-evolutionary data aids in accurate segmentation and sequence assignment.