Re-import exported particles yield bad reconstructions

Import and Export
1

Hi all,

My intention was to exclude particles based on their alignment angles (tilt specifically). During which, I found that re-import exported particles cannot yield same level of reconstructions.
Check attached.

What I did was as follow:

  1. Press export button in J187 output particles column
  2. cd that path
  3. csparc2star.py J187_particles_exported.cs J211_input.star
  4. In cryosparc, import J211_input.star
  5. In cryosparc, clone J187, replace Input particle stack with the new imported one
  6. Run J211
    This yield a smeared out map.

I found that if I multiply all the values of _rlnOriginXAngst _rlnOriginYAngst by the pixel size. And import, reconstruct again, this can bring the resolution back to similar level (J227). But still not as good as J187.

What else might be missing to bring the simple reconstruction the same as J187?

I am using cryosparc 4.6.0, pyem 0.62.

Thanks very much!
Ruizhi
Screenshot 2024-10-14 152408
Screenshot 2024-10-14 152342
Screenshot 2024-10-14 152323

2

Did you do any aberration correction at all (e.g. beam tilt, mag aniso)? CryoSPARC and RELION handle the numbers differently and they’re usually lost in the conversion.

3

I don’t think so. I only did a NU refinement (with defocuse refinement and global ctf turned off). Then I did a simple reconstruction to get a clean FSC curve. The higher order ctf refinement was not done yet.