Particles extracted from different micrographs (bug?)

Hi,

I noticed an odd behaviour during particle extraction in cryosparc today, where it appears that the same particle extraction job was using a (random?) mix of micrograph sources to extract from…

For context, I started this project with a CS live session and exported the pre-processed micrographs from the live session to a workspace for downstream picking/classification. I am dealing with very low SNR particles and I noticed that motioncor3 was doing a better job compared to CS Patch Motion in this case - so I imported motioncor3-corrected micrographs (Y-flipped to match CS convention) and linked the export from CS live session as source movies. I then re-extract particles and continued processing without issues (and the quality of 3D volumes with motioncor3 particles did improve).

Now to the odd part - when I exported my final particle stack using csparc2star.py, I noticed that the MicrographName path in the star file was sometimes referring to the the motioncor3 micrographs (“Jxxx/imported/…”) and sometimes to the CS live Patch motion job (“S1/motioncorrected/…”) - apparently at random and even switching between source micrographs for particles from the same image.

I’m wondering now if this is a bug in csparc2star.py mixing up the source paths somehow or if CS Extract actually used different micrograph sources for the particle extraction?

On a similar note - I just checked the log of the Patch CTF Estimation job that I ran on the imported MotionCor3 micrographs - this job used the motion-corrected micrographs from the CS live session instead of the newly imported micrographs which it was intended to run on.

BW
Sven