Real-space quantum-based refinement for cryo-EM: Q|R#3.

Area of Science:

• Structural Biology
• Computational Chemistry
• Biophysics

Background:

• Electron cryo-microscopy (cryo-EM) is a powerful technique for determining the 3D structure of biomolecules.
• While cryo-EM resolution has improved, refining atomic models from low-resolution maps remains challenging.
• Current methods often yield models lacking precise stereochemical accuracy.

Purpose of the Study:

• To investigate the utility of quantum-chemical calculations for improving atomic model refinement in cryo-EM.
• To assess the impact of ab initio and semi-empirical quantum methods on model quality.
• To introduce new algorithms for real-space quantum refinement.

Main Methods:

• Ab initio quantum-chemical calculations (HF-D3/6-31G) were used to derive geometry restraints.
• Semi-empirical calculations (GFN1-xTB) were employed for larger structures.
• These restraints were incorporated into the real-space refinement process using the qr.refine software.

Main Results:

• Inclusion of quantum-chemical restraints demonstrably improved the refinement of an example cryo-EM structure.
• The method proved robust for larger biomolecular structures (up to 7000 atoms).
• Enhanced stereochemical quality of the refined atomic models was observed.

Conclusions:

• Quantum-chemical calculations provide valuable restraints for refining cryo-EM structures, particularly at lower resolutions.
• This approach addresses a key limitation in obtaining accurate atomic models from cryo-EM data.
• The implemented real-space quantum refinement algorithms offer a practical solution for the structural biology community.