[3dem] Relion: particles not aligning after Bayesian polishing

[Morici, Martino]

Hi everyone

I am processing a dataset in Relion 4 that is giving me some issues.
After CTF refining, the particles were not aligning anymore upon 3D refinement. Therefore I tried 3D-refining after each CTF step and I figured out the problem was the 4th order aberrations paramenter: when set to No, refinement was then working correctly (I didn´t use trefoil estimation).
However, after I do bayesian polishing, the 3D refinement gives the same problem, i.e. it is not able to align the particles (the job results in a random bowl of density, empty inside.
In order to overcome this, I tried several approaches: I tried polishing with and without training on the particles, I regrouped them, I changed box size, I changed workstation and I also did the job on the whole particle and on a focused area (the particle is a ribosome). I also tried to make the refinement process easier for Relion, keeping the reference not so much low-pass filtered (at this stage it is difficult to bias the result), adding maxsig 3000 as additional argument, using a mask and skipping padding and gridding.
Nothing of this worked and I don´t know how to proceed.

Has anyone ever had this kind of problem and know how to solve it?
Thank you for your time
- Martino

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.ncmir.ucsd.edu/pipermail/3dem/attachments/20230630/d92bda34/attachment.html>