I have a test dataset collected on apoF and motion correction (cryosparc v4.4) is giving an unusual result.
Majority of the movies have a in-frame motion of 40 and total motion of 160pix. When checking the trajectory, it seems the large in-frame motion arise from the last few frames.
I ran relion motioncorr on the same set of movies and I do not see a problem with the movies (total motion around 10, in frame motion around 2).
I’m happy to share the movie files to help troubleshoot.
This seems very similar to this bug. Try full-frame motion or set the “Last frame” parameter to N-1 (e.g. 39 if you have 40-frame movies).
Hi Daniel,
Thank you so much. Exactly that did the trick!
Best,
Yibo
Good! May I ask which camera and collection software was used for this data? Falcon 3 or 4 with EPU? We have this problem with a Falcon 3EC camera on a Glacios system using EPU, both version 2 and 3. Full-frame motion correction in CryoSPARC and MotionCor2 does not cause this issue. Only Patch Motion.
The falcon3 camera follows a rule of 7 for fraction. It could be the last fraction you have is very dark and in cryosparc patch motioncor, it’s seen as large drift. They remove last frame doing motioncorr.
The movies can be inspected using imageJ or Thermofisher image viewer.
By “rule of 7”, do you refer to the hardware motion correction (called “Align” in EPU)? The manual says that the hardware alignment option groups the elemental frames in sets of 6. I always tweak the exposure time and the number of dose fractions to have exactly the same number of elemental frames in each dose fraction in EPU, but am still affected by the bug.
