Measurement of atomic scattering factors by cryoelectron microscopy

Area of Science:

• Structural biology
• Biophysics
• Computational chemistry

Background:

• Accurate specimen structure determination in cryo-electron microscopy (cryo-EM) relies on precise electrostatic potential modeling.
• Current electron scattering models often neglect the influence of chemical bonds on macromolecular potential.
• This limitation impacts the accuracy of structural analysis in biological macromolecules.

Purpose of the Study:

• To develop a fast, interpretable, and transferable Bayesian method for estimating atomic scattering factors.
• To incorporate the effect of the molecular environment on scattering factors.
• To improve the accuracy of cryo-EM based structure determination.

Main Methods:

• A Bayesian approach was developed to infer atomic scattering factors directly from electrostatic potential maps obtained via cryo-EM single particle analysis.
• The method bypasses computationally intensive theoretical calculations.
• Empirical scattering factors were inferred from high-resolution cryo-EM reconstructions of catalase enzymes and public datasets.

Main Results:

• The inferred empirical scattering factors demonstrated improved agreement with independent cryo-EM reconstructions.
• Variance of unmodeled signal decreased by up to a factor of three in the resolution range of 1/15 to 1/3 Å⁻¹.
• Validation was achieved through comparison with magnetic susceptibility values and application to atomic model refinement.

Conclusions:

• The proposed Bayesian method provides an efficient and accurate way to estimate atomic scattering factors for cryo-EM.
• This approach enhances the reliability of structural models for biological macromolecules.
• The method shows broad applicability and potential for routine use in structural biology.