Real-space refinement with DireX: from global fitting to side-chain improvements.

Area of Science:

• Structural Biology
• Biophysics
• Computational Biology

Background:

• Single-particle cryo-electron microscopy (cryo-EM) is crucial for determining large biomolecular structures.
• Achievable resolution in cryo-EM varies widely (3.5-20 Å), posing challenges for atomic model interpretation.
• Refining atomic models from intermediate- to low-resolution cryo-EM density maps is an active research area.

Purpose of the Study:

• To introduce DireX, a novel real-space structure refinement program specifically designed for cryo-EM density maps.
• To address the challenges of interpreting and refining low-resolution cryo-EM data.
• To evaluate the performance of DireX on diverse cryo-EM datasets.

Main Methods:

• DireX employs Deformable Elastic Network (DEN) restraints to minimize overfitting by reducing the effective degrees of freedom during refinement.
• Restrained grouped occupancy refinement is utilized to handle artifacts caused by missing or reduced density in flexible protein regions.
• The program's efficacy was tested using six density maps from four different proteins in the 2010 Cryo-EM Modeling Challenge.

Main Results:

• DireX demonstrated effective structure refinement across a broad resolution range (3-23 Å).
• The program successfully addressed challenges associated with flexible protein regions and low-resolution data.
• Performance evaluation in the Cryo-EM Modeling Challenge showcased typical and reliable refinement outcomes.

Conclusions:

• DireX provides a robust solution for real-space refinement of cryo-EM density maps.
• The program's methods, DEN restraints and grouped occupancy refinement, enhance the accuracy of atomic models derived from low-resolution cryo-EM data.
• DireX is a valuable tool for structural biologists working with cryo-EM data, particularly at intermediate to low resolutions.