Summary

Local assessment is crucial to evaluate the reliability of specific regions in a structure (like binding sites or loops), even if global scores are good. Different research questions require different levels of local accuracy.

Visual inspection in a molecular viewer (like PyMOL, Chimera, Moorhen, and Coot ) is essential for local assessment.

For X-ray and cryo-EM structures, visually assess the fit of the atomic model to the experimental data:

• Pay attention to areas with weak or missing density, as models in these regions are less reliable.
• Identify local modelling issues using visual inspection and detailed information from the PDB Validation Report:
  • Examine specific stereochemical outliers (Ramachandran, sidechain, bond/angle deviations) flagged in the report. Is the unusual geometry supported by density?
  • Locate and assess the impact of atomic clashes, especially in regions of interest.
  • Understand that occupancy and multiple conformations indicate flexibility or partial presence, affecting local model certainty. Lower occupancy means less confidence in that specific position.

For structures with ligands, perform careful local assessment:

• Visually check the ligand’s fit to the density.
• Consult the validation report for ligand-specific RSCC values.

Use the detailed sections and residue-property plots in the PDB Validation Report to pinpoint the location and nature of local issues.

Consider other factors that influence a model’s completeness, flexibility, and relevance:

• Be aware of unmodeled parts of the protein chain, especially if they are critical to your research. NMR structures often provide ensembles that capture flexibility rather than simply omitting regions.
• For large proteins, structures may be determined from individual fragments. You may need to consult sequence data and molecular biology reports to understand the full functional molecule.
• Understand that X-ray structures often lack explicit hydrogen atoms, which can lead to ambiguity in certain side chains (e.g., asparagines, glutamines). NMR structures typically include all hydrogen atoms.
• Ensure the structure represents the biological state relevant to your study (e.g., bound vs. unbound, active vs. inactive).

Finally, when choosing a structure, consider its relevance to your specific research question, the experimental technique, resolution, and the presence of any bound molecules or interactions.