Cryo-EM targets in CASP14.

Area of Science:

• Structural biology
• Biophysics
• Computational biology

Background:

• The Critical Assessment of protein Structure Prediction (CASP) competition assesses computational methods for protein structure prediction.
• Cryo-electron microscopy (cryo-EM) is a powerful technique for determining high-resolution structures of biological macromolecules.

Purpose of the Study:

• To evaluate the accuracy of computational models submitted for CASP14 targets against experimental cryo-electron microscopy (cryo-EM) data.
• To assess the impact of a-posteriori refinement on model accuracy and its correlation with experimental cryo-EM density maps.

Main Methods:

• Determination of seven CASP14 target structures using cryo-electron microscopy (cryo-EM) with resolutions ranging from 2.1 to 3.8 Å.
• Evaluation of submitted computational models based on their fit to cryo-EM density maps and experimental reference structures.
• Coordinate-to-density and coordinate-to-coordinate fitting metrics were employed to measure model accuracy.

Main Results:

• A-posteriori refinement of the most accurate models resulted in structures closely resembling the experimental reference structures.
• Some refined regions exhibited improved fit to the cryo-EM density compared to the initial reference structure.
• Regions with poor refineability correlated with high diversity among CASP models and low goodness-of-fit to the cryo-EM density.

Conclusions:

• Computational models can be refined against experimental cryo-EM data to achieve high accuracy.
• The study identifies specific regions in protein structures that pose challenges for both prediction and refinement.
• Understanding these challenging regions can guide future improvements in protein structure prediction and experimental structure determination.