Beam-induced motion correction for sub-megadalton cryo-EM particles

Sjors H. W. SchereseLife 2014 · DOI

TL;DR

This paper introduces a refined algorithm for correcting beam-induced motion in cryo-electron microscopy (cryo-EM) images. The new method significantly improves the resolution of cryo-EM structures, particularly for smaller protein complexes (under 200 kDa), by better modeling particle movement and accounting for radiation damage. This advancement broadens the applicability of high-resolution cryo-EM to a wider range of biological molecules.

What Problem Does This Paper Address?

Electron cryo-microscopy (cryo-EM) is a powerful technique for determining the 3D structure of biological molecules. However, a major challenge is that the electron beam used for imaging causes the sample to move, blurring the images and limiting the achievable resolution. This motion blur is particularly problematic for smaller particles, where the signal-to-noise ratio in individual movie frames is lower, making it harder to accurately track their movements. Previous methods, while effective for large particles like ribosomes, struggled with smaller specimens. This paper aims to address this limitation by developing a more robust and versatile motion correction algorithm.

Methods

The core of the paper is an improved movie processing algorithm that builds upon previous work. The key innovations include:

  • Particle-based movement correction: Instead of just correcting individual particles, the new algorithm considers multiple particles within the same field of view. It fits straight movement tracks to these particles, effectively averaging out noise and improving the accuracy of the estimated motion. This is achieved by minimizing a weighted least-squares target that incorporates neighboring particles.
  • Dose-dependent radiation damage modeling: The algorithm models how the quality of high-resolution information degrades with increasing electron dose. It estimates "relative B-factors" for each movie frame, which describe how the signal fall-off in that frame compares to the average. These relative B-factors are then used to assign frequency-dependent weights to each frame, down-weighting frames that suffer more from radiation damage at high resolutions.
  • Hybrid approach: For cases with significant initial drift, a hybrid strategy combining a field-of-view-based correction (like that by Li et al.) with the new particle-based method is proposed.

Datasets

The new algorithm was tested on four previously published cryo-EM datasets:

  • Human y-secretase (molecular mass ~170 kDa)
  • Escherichia coli β-galactosidase (molecular mass ~450 kDa)
  • Bovine Complex-I (molecular mass ~1 MDa)
  • Large subunit (LSU) of the yeast mitochondrial ribosome (molecular mass ~1.9 MDa)

These datasets represent a range of particle sizes, with a particular focus on smaller complexes.

Key Insights and Ideas

  • Leveraging collective motion: The idea that multiple particles in proximity can collectively inform the motion correction for each individual particle is a significant advancement. This "wisdom of the crowd" approach is particularly effective for improving signal-to-noise in motion estimation for smaller particles.
  • Quantifying and mitigating radiation damage: Explicitly modeling the dose-dependent fall-off of high-resolution information allows for a more nuanced weighting of movie frames. This is crucial because early frames, while potentially having less accumulated dose, can suffer from initial beam-induced motion, while later frames might have more dose but better initial stability.
  • Broadening cryo-EM's reach: The algorithm's success with particles under 200 kDa is a major breakthrough, opening up high-resolution structural determination to a much larger class of biological macromolecules.

Key Results

  • The new algorithm significantly improves cryo-EM structures for all four tested datasets, enabling the identification of finer biological details.
  • For particles smaller than 500 kDa, the new approach yields substantial resolution improvements, whereas the original method did not.
  • The resolution achieved for y-secretase improved from 4.9 Å to 4.5 Å, for β-galactosidase from 4.3 Å to 4.0 Å, and for Complex-I from 5.9 Å to 4.8 Å. The mitoribosome LSU saw a slight decrease in resolution from 3.23 Å to 3.3 Å, but this difference was deemed negligible in practice.
  • The new method is computationally faster than the original approach, with a speed-up factor of up to 6 for large particles.
  • The algorithm has been implemented in the open-source RELION program as "particle polishing."

Limitations and Future Work

  • Limited to straight tracks: The current algorithm models particle movement as straight tracks. It may not perform well for more complex, non-linear movements (e.g., bent or zig-zag tracks). The authors suggest a hybrid approach for such cases.
  • Out-of-plane rotations: The algorithm does not explicitly consider out-of-plane rotations, which might slightly affect results for very large particles where precise rotational determination is possible.
  • Future work: The authors express hope that this new approach will contribute positively to the structure determination of a wide range of other cryo-EM samples, particularly smaller ones. The implementation in RELION suggests its immediate utility for the broader cryo-EM community.

Summary updated

Abstract

In electron cryo-microscopy (cryo-EM), the electron beam that is used for imaging also causes the sample to move. This motion blurs the images and limits the resolution attainable by single-particle analysis. In a previous Research article (Bai et al., 2013) we showed that correcting for this motion by processing movies from fast direct-electron detectors allowed structure determination to near-atomic resolution from 35,000 ribosome particles. In this Research advance article, we show that an improved movie processing algorithm is applicable to a much wider range of specimens. The new algorithm estimates straight movement tracks by considering multiple particles that are close to each other in the field of view, and models the fall-off of high-resolution information content by radiation damage in a dose-dependent manner. Application of the new algorithm to four data sets illustrates its potential for significantly improving cryo-EM structures, even for particles that are smaller than 200 kDa.
BibTeX
@inproceedings{scheres2014beam,
  author = {Sjors H. W. Scheres},
  title = {{Beam-induced motion correction for sub-megadalton cryo-EM particles}},
  booktitle = {eLife},
  year = {2014},
  doi = {10.7554/elife.03665},
}
Figure