MarkovFit: Structure Fitting for Protein Complexes in Electron Microscopy Maps Using Markov Random Field.

Area of Science:

• Structural Biology
• Computational Biology
• Biophysics

Background:

• Cryo-electron microscopy (cryo-EM) is increasingly used to determine protein complex structures.
• Fitting known individual protein structures into cryo-EM density maps is crucial for accurate structural modeling, especially at lower resolutions.
• Existing methods face challenges in accurately modeling protein complexes within these maps.

Purpose of the Study:

• To design and validate a novel computational method for fitting atomic protein structures into cryo-EM density maps.
• To improve the accuracy of protein complex modeling, particularly for medium to low-resolution cryo-EM data.
• To enable probabilistic evaluation of fitted structural models.

Main Methods:

• Developed MarkovFit, a method utilizing Markov random fields for structural fitting.
• The method probabilistically evaluates fitted atomic models within cryo-EM density maps.
• Tested on simulated and experimentally determined cryo-EM maps across various resolutions.

Main Results:

• MarkovFit demonstrated superior accuracy compared to existing methods on multiple datasets.
• Performance was validated on 31 simulated cryo-EM maps (resolution ~10 Å).
• Effectiveness confirmed on 9 experimental cryo-EM maps (<4 Å) and 28 experimental maps (6-20 Å).

Conclusions:

• MarkovFit provides a robust and accurate approach for fitting atomic protein structures into cryo-EM density maps.
• The method is particularly effective for medium to low-resolution maps, advancing structural modeling capabilities.
• Probabilistic evaluation enhances the reliability of the fitted protein complex models.