Structural heterogeneity and precision of implications drawn from cryo-electron microscopy structures: SARS-CoV-2...

Area of Science:

• Structural biology
• Computational biophysics
• Protein engineering

Background:

• Protein structures from cryo-electron microscopy (cryo-EM) are widely used for functional and mutation effect predictions.
• The SARS-CoV-2 S-protein has been extensively studied using various structural models.

Purpose of the Study:

• To assess the sensitivity of functional implications to the specific protein structure used.
• To evaluate the precision of computational methods for predicting mutation effects on protein stability.
• To explore strategies for improving the reliability of these predictions.

Main Methods:

• Calculated stability changes (ΔΔG) for all possible mutations across 23 distinct SARS-CoV-2 S-protein structures.
• Employed three different computational methods for stability change estimation.
• Analyzed 3.69 million ΔΔG values to assess precision and error sources.
• Investigated structure-averaged group comparisons to remedy errors.

Main Results:

• Individual mutation effect predictions exhibit low precision due to method and structural heterogeneity.
• Results for single mutations are largely uninformative.
• Structure-averaged mean effects for groups of substitutions yield statistically significant and reliable results.
• Functionally important natural mutations show a smaller average stability impact than other mutations, irrespective of protein conformation.

Conclusions:

• Structure-based protein modeling for predicting mutation effects faces substantial precision challenges.
• Sensitivity testing is recommended to quantify the impact of structural variations.
• Structure-averaged ensemble estimates for groups of residues offer a viable solution for improving prediction reliability when multiple structures are available.