Exporting subset of micrographs from CS to Relion

Import and Export
1

For something specific I want to try—manually picking filaments in Relion—I need to take the micrographs where CryoSPARC found at least some filaments and bring them into Relion.

I initially tried using pyem for this, but wasn’t successful in this particular case. So, I wrote a couple of scripts that identify the micrographs containing particles and generate a .star file listing those micrographs with their full paths.

However, when I import this .star file into Relion and try to run CTF estimation from scratch using CTFFIND, the job fails for all micrographs.

For context:

  • The micrographs were motion-corrected in Relion
  • Then imported into CryoSPARC for particle picking and further processing
  • Now I’m trying to bring a subset of those micrographs back into Relion to start from scratch.

I’ve attached an image of the .star file I generated, which points to the motion-corrected images. I realize the pipeline is a bit convoluted, but I’m hoping to troubleshoot this step so I can proceed with manual filament picking in Relion.

Any idea what might be causing CTFFIND to fail here?

Thanks so much,

Screenshot_2025-08-11_15-02-32
Screenshot_2025-08-11_15-02-32839×767 132 KB

J.

2

Are they 16-bit MRC? RELION defaults to that now and CTFFIND doesn’t support it - RELION outputs a _PS.mrc file in 32-bit to get CTFFIND to work.

You can grep -wFf [micsYouWant.txt] [RELIONmicrographs_ctf.star] > cleanedListOfMics.star

To clean up the output from RELION CTFFIND, which should bypass any pain with trying to re-CTF estimate 16-bit MRC files.