Preprocessing cryo-EM data

COSMIC² offers two different pipelines for preprocessing cryo-EM data: movies to particles (A) or micrographs to particles (B)

 

 

 

 

 

 

 

 

 

 

 

Key features of the pipeline:

1. Particle picking using crYOLO, a deep-learning based neural network particle picker:

  • The general model of deep learning for picking works for particles of all shapes and sizes
  • It effectively avoids ice, empty areas, carbon film
  • Removes need of users to select ‘good’ particles

2. Removal of ‘bad’ micrographs by setting resolution cutoff of micrographs

  • User-defined resolution cutoff to remove micrographs with CTFFIND4 fits worse than a specified resolution (Default=6 Angstroms)

3. Automation!

  • This streamlined pre-processing allows users to input movies or micrographs and then receive particle stacks as an output that can go straight into RELION analysis

Data formatting & upload

In order to upload movies or micrographs, please follow these steps:

  1. Place all files into a single directory for upload to COSMIC²
    • Place all movie files within directory
    • Place star file within directory
  2. Create star file in this new directory and provide file names of movies (must be .tif) or micrographs to be analyzed:
    • For movies in the directory named ‘Movies’ in .tif format: 
      • printf 'data_\nloop_\n_rlnMicrographMovieName\n' >> Movies/movies.star | ls Movies/*.tif >> Movies/movies.star
    • For micrographs in the directory named ‘micrographs’ in .mrc format: 
      • printf 'data_\nloop_\n_rlnMicrographName\n' >> Micrographs/micrographs.star | ls Micrographs/*.mrc >> Micrographs/micrographs.star
  3. On COSMIC2, use Globus to select entire directory (‘Movies’ or ‘Micrographs’) for upload

Considerations:

  • Save upload time by compressing all movies into TIFF format
  • All movies will be deleted after they undergo motion correction

If you use either of these preprocessing pipelines, please cite the following software: 

“RELION: Implementation of a Bayesian approach to cryo-EM structure determination.” S.H.W. Scheres J. Struct. Biol. (2012) 180:519-530. Link

“CTFFIND4: Fast and accurate defocus estimation from electron micrographs.” A. Rohou & N. Grigorieff J. Struct. Biol. (2015) 192:216-221. Link

“SPHIRE-crYOLO: A fast and well-centering automated particle picker for cryo-EM.” T. Wagner, F. Merino, M. Stabrin, T. Moriya, C. Gatsogiannis, S. Raunser bioRxiv (2018) 356584; doi: https://doi.org/10.1101/356584. Link