Phase plate data processing CTF correction
Added by Anchi Cheng over 7 years ago
Question from a user, using beta version of Appion
I noticed that for my data, when i manually pick 1000 particles and 2D classify them for templates, all the particles bunch up into one low res class. If I ignore the CTF until the first peak, all the particles separate nicely. Did you guys run into similar issues with your phase plate data? I suppose it is due to poor ctf at very low res for phase plate data. Does ignoring ctf until first peak have a negative impact on resolution?
Replies (1)
RE: Phase plate data processing - Added by Anchi Cheng over 7 years ago
In current Appion particle stack making, the ctf correction part does not handle phase shift at all. The best to do is not to ctf correct, nor phase shift. Phase plate data uses low enough defocus, what you did is fine for 2D classification for template generation or stack cleaning, indeed.
This is not because poor ctf at very low res but unexpectly good (high value) contrast in that range from the program point of view.
In Relion/Frealign/CryoSparc, due to maximum likelyhood nature of them, they all do its own ctf correction (not estimation) during refinement and 2D/3D classification. For you to ignore it now is fine. You will be importing the ctf estimation result from Appion as star file or do preparation step for Frealign. Both of those will include your phase shift values.
If you are running Relion/Frealign/CyoSparc stand alone and have this problem, you may not be running them correctly or you may have an older version that does not handle phase shift.