I have a dataset where images become significantly astigmatic tooward the end of the collection. The CTFs of Images at the end of are fit much better with a relaxed astigmatism restraint, but this causes problems when fitting earlier images that are not very astigmatic. I would like to use CTF fits from different CTFfind runs for later processing. However when I make an asset from the picked particles, cisTEM seems to only use defocus vaues from the latest find CTFfind run. Is there a way to specify which CTF fits are used for each image? I thought that the check marks next to the images in the "find CTF" tab under the results menu would do this, but it doesn't seem to make a difference.
Hi Evan,
Yes the tick marks should do this, when you create a refinement package - it should use the defocus values that are associated with the result that is currently ticked for that particular image.
If that is not happening, then it is a bug. Have you selected different jobs for different images, and this is reflected in the ticks - but when you create the refinement package -you can see from the particle list that it has not used the correct values?
Thanks!
Tim
Hi Tim,
Thanks for the response. Yes that is exactly the problem. I select fits from different CTFfind jobs using the tick marks and then pick particles and create a particle stack. However, most of the particles in the refinement package use defocus values from the most recent CTFfind job, not the selected job. Some of the particles in the refinement package use defocus values from other CTFfind jobs that are not the most recent, but none of the particles seem to use the CTF fits that I had selected in the results tab.
Interestingly, when I restart cisTEM and look at the results tab the tick mark selections I had made before are not retained. Instead, the incorrect CTFfind jobs that were used to make the refinement package are selected with the tick marks. So it looks like cisTEM is changing which CTF fits are selected.
When I look at the dialog in the cisTEM terminal I do get an error when I make the refinement package:
Gtk-CRITICAL **: IA__gtk_tree_row_reference_new_proxy: assertion 'path->depth > 0' failed
Hi Evan,
Sorry about that. I will have to spend some time looking into it.
Can you do me a favour - can you open the project, change 2 or 3 of the CTF jobs. Then immediately close the project and re-open. If you do that do they stay changed?
Thanks,
Tim
Hi Tim,
When I selected 20 different CTF fits and then immediately closed cisTEM and then reopened the session, most of the selected CTF fits do not change. This time only one of selecte fits was changed to select the most recent CTF fit.
Evan
Hi Evan,
Thanks for testing - I will look into this, and hopefully resolve it for the next release.
In the meantime, you can maybe use the workaround posted by another user :-
https://cistem.org/bug-saving-selected-fits-find-ctf
Essentially, you have to make sure the eye is not next to the result you want to select - then it should be saved.
Thanks!
Tim