Accurate Macromolecular Complex Modeling for Cryo-EM with CryoZeta.

Area of Science:

• Structural Biology
• Computational Biology
• Biophysics

Background:

• Cryogenic electron microscopy (cryo-EM) is a powerful technique for determining 3D structures of biological macromolecules.
• Accurate structural model building from cryo-EM data, especially at non-atomic resolutions, remains a significant challenge.
• Existing methods struggle to fully leverage cryo-EM map density for de novo structure prediction.

Purpose of the Study:

• To develop a novel computational tool, CryoZeta, for de novo structure modeling from cryo-EM data.
• To improve the accuracy and fidelity of structural models by integrating cryo-EM map density with sequence information.
• To provide a robust solution for automated, high-resolution modeling of diverse biomolecular assemblies.

Main Methods:

• CryoZeta utilizes a diffusion-based generative deep neural network.
• It integrates experimental cryo-EM map density features with a biomolecular structure prediction pipeline.
• The method jointly leverages amino acid sequence information and cryo-EM density features.

Main Results:

• CryoZeta generates highly accurate de novo structural models consistent with experimental cryo-EM maps.
• The program demonstrates superior performance in atomic accuracy compared to existing cryo-EM modeling methods.
• Performance was validated on benchmark datasets including protein complexes, protein-nucleic acid assemblies, and nucleic acid systems up to 10 Å resolution.

Conclusions:

• Directly incorporating cryo-EM density into structure prediction pipelines significantly enhances model accuracy.
• CryoZeta represents a robust and automated tool for high-fidelity biomolecular structure modeling from cryo-EM maps.
• This approach advances the field of structural biology by enabling more precise determination of macromolecular structures.