Hello! I’m looking for some advice on interpreting the empirical dose weight heat maps from a reference-based motion correction job run on v4.4.0+231114. If I’m correctly interpreting the figure (see below), both early and late frames appear to be highly weighted at high frequencies. Additionally, there is a sharp step change in weighting around 20 frames for mid to low frequencies. I would have assumed a more gradual down weighting of later frames, especially at high frequencies, so I am wondering what is going on here. Have others observed similar behaviour?
Regarding the steep change in the weighting on your dataset, this is something we have not observed before. We had a few questions related to your data collection and upstream processing:
Was this data collected using a variable dose scheme where frames 21-40 had a different dose compared to frames 1-20?
Are there any irregularities in the upstream processing or patch motion correction job?
Thanks for your follow up questions and sorry for missing the previous topic. Thanks to @olibclarke for guiding me in the right direction.
Movies were collected with a consistent dose over 40 frames with a total dose of 63 e-/A2.
I was wondering whether upstream processing was causing the sharp drop in weighting too and have found the cause. I am dealing with a ligand that we suspect is sensitive to radiation damage and which has consistently weaker density for one region. I was previously using particles from micrographs where only the first 20 frames of data were used during patch motion correction to see if this would help get better density for the whole ligand. I then used these particles for the reference-based motion correction job. I wasn’t expecting this to be an issue; however, when I repeated the reference-based motion correction job, this time using the same batch of particles that had been taken from micrographs processed with all frames in a patch motion correction job, the outputs look much more normal (see below).
This looks a lot better. If you find that the first 20 frames might yield a better result in respect to your ligand, you can perform RBMC using only the first 20 frames. You will need a Patch Motion Correction job where you specify the end frame as 19. Then, perform Patch CTF Estimation to get CTF info, and then using the Reassign Particles to Micrographs job to associate your 40 frame particles to your 20 frame patch motion correction job. Then you should be able to run RBMC on only the first 20 frames.
