Computational approaches for protein complex modeling for intermediate resolution cryo-EM maps

Area of Science:

• Structural Biology
• Biophysics
• Computational Biology

Background:

• Cryo-electron microscopy (cryo-EM) is crucial for determining macromolecular structures.
• Interpreting intermediate (5-10Å) to low-resolution (over 10Å) cryo-EM density maps for structure modeling remains a significant challenge.
• Accurate structure modeling tools are essential for advancing cryo-EM data interpretation.

Purpose of the Study:

• To address the challenges in cryo-EM map interpretation.
• To develop and present computational tools for structure modeling and validation.
• To introduce the EMSuite web server, a freely accessible platform for cryo-EM structure analysis.

Main Methods:

• Development of the EMSuite server, providing 14 computational tools.
• Focus on three methods for intermediate resolution maps: DiffModeler, DMcloud, and VESPER.
• DiffModeler: Enhances backbone positions using a diffusion model and AlphaFold for complex modeling.
• DMcloud: Performs local structure fitting to refine models against density maps.
• VESPER: Implements global structure-map and map-map fitting for general use.

Main Results:

• The EMSuite server offers a comprehensive suite of tools for cryo-EM structure modeling.
• DiffModeler enables the construction of complete protein complexes.
• DMcloud refines models by identifying and removing incompatible regions with the density map.
• VESPER provides robust global fitting capabilities for structure-map and map-map comparisons.

Conclusions:

• The EMSuite server significantly enhances the process of structure modeling from intermediate to low-resolution cryo-EM maps.
• The presented tools, DiffModeler, DMcloud, and VESPER, offer effective solutions for various aspects of cryo-EM structure determination.
• EMSuite provides a valuable, accessible resource for the structural biology community to improve macromolecular structure analysis.