RBMC - too few particles per micrograph

Motion Correction
1

I have a Glacios dataset that goes to 3.3A and wanted to see if RBMC would improve the resolution, however, I am running into this error:

[CPU:  697.5 MB]
Resolution cutoffs: alignment 4.746 A, cross-validation 3.356 A
[CPU:  743.4 MB]
Removed 1964 movies with fewer than 2 particles.
[CPU:  708.1 MB]
Removed 114803 particles (in micrographs with fewer than 2 total particles).
[CPU:  419.3 MB]
Traceback (most recent call last):
  File "cryosparc_master/cryosparc_compute/run.py", line 95, in cryosparc_master.cryosparc_compute.run.main
  File "cryosparc_master/cryosparc_compute/jobs/motioncorrection/run_reference_motion.py", line 265, in cryosparc_master.cryosparc_compute.jobs.motioncorrection.run_reference_motion.run_reference_motion_correction
AssertionError: All movies must have the same number of frames

It seems my particles are spread out too much over the micrographs (2000 movies, 110k particles). Is this a general consideration?

2

this doesn’t quite make sense… 1964 movies with a maximum of ~4000 particles doesn’t add up to removing 114k particles - something else is going on.

Also there is a secondary error here - it is complaining that some of the movies have a different number of frames, which will always be fatal - you can run import without skip header check to verify?

1 Like
3

Hi @martinpacesa, one possible reason that there are 114k particles discarded when there were only 1964 movies with fewer than 2 particles is because there may have been input particles that aren’t mapped to one of the input movies at all. For example, suppose you import two sets of movies, preprocess them and extract particles from each, do a refinement using both sets of particles, and then connect both sets of particles to a RBMC job, but only one set of movies. In that case, you’d have a large number of particles that weren’t picked on any of the input movies so RBMC cannot process them and they get discarded.

I agree, however, that the wording of the message is a bit misleading, it should probably be reworded to indicate that this is a possible cause.

4

Thank you for the help, you were right, particles were rejected because they were not assigned properly to the micrographs after reimporting. Reassignment job fixed the problem :slight_smile:

1 Like
5

Hi, I need help as well. I have been encountering a similar problem, where RBMC will throw out particles and movies with assumed incorrect frame rate. It discarded 11 movies but ~5k particles from 11 movies, which is kind of impossible given that the average is 4 per micrograph. I saw this thread and tried what Martin did i.e. reassign particles, but it still throws out the particle. Can you please advise what to do? Also, is there a way to still add those micrographs where it says are less than 2 particles, i.e., with 1 particle >.<?

image
image1008Ă—1125 131 KB

image
image1296Ă—162 10.2 KB

6

@ZahirSid Please can you post a view (or views) of the job tree that includes

  1. NU-Refine job J640
  2. upstream particle picking and movie import jobs

and illustrates any potential branches that need to be considered for reassignment of particles to micrographs.

7

Thanks a lot for getting back @wtempel. I think we are now already discussing the same issue in this thread: Reference based motion correction error: All movies must have the same number of frames - #51 by ZahirSid, therefore I haven’t replied any further here.