Related Experiment Video
Updated: Apr 6, 2026
High-resolution Single Particle Analysis from Electron Cryo-microscopy Images Using SPHIRE
Published on: May 16, 2017
A Bayesian approach for suppression of limited angular sampling artifacts in single particle 3D reconstruction
Toshio Moriya1, Erman Acar1, R Holland Cheng2
1Department of Signal Processing, Tampere University of Technology, P.O. Box 553, FI-33101 Tampere, Finland; BioMediTech, Tampere University of Technology, P.O. Box 553, FI-33101 Tampere, Finland.
Sequential maximum a posteriori estimate with expectation maximization (sMAP-EM) improves 3D structure accuracy in cryo-electron microscopy by addressing limited angular sampling. This method enhances resolution and fills missing data, facilitating near-atomic protein structure determination.
Area of Science:
- Structural Biology
- Cryo-Electron Microscopy
- Computational Imaging
Background:
- Single particle reconstruction in cryo-EM is often limited by angular sampling artifacts, leading to inaccurate 3D structures.
- Existing methods like 2D class averaging and random-conical tilt reconstruction introduce their own limitations, such as insufficient angular sampling or the 'missing cone' artifact.
Purpose of the Study:
- To address the challenge of limited angular sampling in 3D structure determination using cryo-electron microscopy.
- To develop and evaluate a novel algorithm, sequential maximum a posteriori estimate with expectation maximization (sMAP-EM), for improved 3D reconstruction.
Main Methods:
- Implementation and application of the sMAP-EM algorithm to both simulated and experimental cryo-electron microscope datasets.
- Comparison of sMAP-EM with the direct Fourier method, evaluating reconstruction error and resolution.
- Analysis of noise level and sampling sparseness effects on sMAP-EM reconstructions and development of quantitative measurements for missing cone information.
Main Results:
- sMAP-EM consistently outperformed the direct Fourier method in terms of reconstruction accuracy and resolution across various sampling strategies, noise levels, and sampling sparseness.
- Frequency domain analysis confirmed sMAP-EM's ability to incorporate meaningful information into unmeasured angular space without prior object knowledge.
- Quantitative assessments demonstrated effective filling of the 'missing cone' artifact in conical tilt sampling simulations.
Conclusions:
- The sMAP-EM algorithm effectively overcomes the limitations of limited angular sampling in single particle reconstruction.
- sMAP-EM demonstrates significant potential for advancing the determination of 3D protein structures at near-atomic resolution.
- The developed quantitative measurements provide valuable tools for assessing reconstruction quality in cryo-EM.
