Outcomes of the EMDataResource cryo-EM Ligand Modeling Challenge

Area of Science:

• Structural Biology
• Biochemistry
• Biophysics

Background:

• Cryo-electron microscopy (cryo-EM) is advancing to near-atomic resolution, enabling detailed structural analysis.
• Accurate modeling of ligands within protein and nucleic acid complexes is crucial for understanding molecular interactions.
• Assessing the quality of these complex models requires robust evaluation metrics.

Purpose of the Study:

• To assess the reliability and reproducibility of ligand modeling in cryo-EM maps at near-atomic resolution.
• To compare different modeling approaches and identify best practices.
• To develop recommendations for evaluating cryo-EM structures of liganded macromolecules.

Main Methods:

• The EMDataResource Ligand Model Challenge used three published cryo-EM maps at 1.9–2.5 Å resolution.
• Sixty-one ligand models were submitted by 17 research groups.
• Model quality was assessed using visual inspection, map quality, model-to-map fit, geometry, energetics, and contact scores.

Main Results:

• A single scoring metric was insufficient for evaluating macromolecule-ligand model quality.
• A composite score integrating various metrics provided a more comprehensive assessment.
• Variability in model quality was observed across different targets and modeling strategies.

Conclusions:

• Best practices for assessing cryo-EM structures of liganded macromolecules at near-atomic resolution are recommended.
• The challenge highlighted the need for standardized evaluation protocols.
• Future efforts should focus on refining scoring systems for complex structural models.