search

eisfabian / PACEtomo

Parallel cryo electron tomography via beam image shift
GNU General Public License v3.0
27 stars 6 forks source link

Tracking defocus on 1.7 #25

Open hamid13r opened 7 months ago

hamid13r commented 7 months ago

Hi Fabian,

I upgraded to 1.7 to try SPACEtomo. One thing I noticed was that no matter what I chose for tracking defocus it still just follows the range. We use an autofocus on tracking to help with the defocus consistency if that matters.

I tried looking at the code but could not make sense of it.

We have reverted to 1.4 because of ctffinf and ctfplotter crashes for the time being.

Thank you so much, Hamid

eisfabian commented 7 months ago

Hi Hamid,

First of all, you can set both doCtfFind and doCtfPlotter to False and it won't impact how PACEtomo is running unless you use features like geoRefine.

Regarding the trackDefocus, it only applies it if it is smaller than the maxDefocus value. Can you send me your settings file of that run?

hamid13r commented 7 months ago

We need an autofocus step, so I think we need the ctffind to run. Also, the georefine has been helping with our tracking.

I am a little confused here. We have been using a higher defocus to get more contrast for the tracking position. Isn't that how it worked in all the previous versions?

eisfabian commented 7 months ago

Interesting, I hadn't heard from anyone using geoRefine so far! I don't think it is very precise unless you have a lot of targets. Are you collecting a target pattern?

The autofocus step (addAF) does not use CTFfind. It uses the default SerialEM autofocus routine.

Yeah, smaller defocus (negative), larger underfocus, sorry.