Defocus estimation on images with hardly any thon rings
Forums
Dear Niko,
I have been trying to use ctffind3 to estimate the defocus of my cryoEM images. The parameters that i used are as follow:
CS[mm], HT[kV], AmpCnst, XMAG, DStep[um]
4.2, 200, 0.07, 106379.6, 15
Box, ResMin[A], ResMax[A], dFMin[A], dFMax[A], FStep[A], dAst[A]
256, 50, 4, 8000, 50000, 1000, 1000
and I get:
DFMID1 DFMID2 ANGAST CC
24050.82 21719.82 60.00 0.06457 Final Values
The right side of the output image showed hardly any thon rings(see attachment). I am wondering if the estimation is still reliable in this case.
Also when I change the box size to 512, the results I got was very different:
DFMID1 DFMID2 ANGAST CC
21413.42 22958.71 7.97 0.04508 Final Values
Is this normal? Could you please give me some advice? Really appreciate it.
-Chunfu
Hi Chunfu, As a first step,
Hi Chunfu,
As a first step, you could reduce the ResMax parameter, which is much too high in your case, try reducing it to 8, or 10 Angstroms. This should improve the reliability of your estimate (and is in general a good thing to do).
I would be a bit worried about using such an image, to be honest.
Axel
Thank you, Axel. I will try
In reply to Hi Chunfu, As a first step, by Axel
Thank you, Axel. I will try to reduce the ResMax to 10 and estimate again.
As of the image, what do you think could be the reason why it has such bad PSD profile? (forgive me if it is a stupid question) I was gonna upload the image and ask you to help me take a quick look at it. But I could not find where to attach it. Basically, it is a 4096*4096 image collected using Gatan CCD camera. I saw that people here use film and scanner more often. I am wondering if this could be why.
-Chunfu
If you could upload a binned
In reply to Thank you, Axel. I will try by chunfuxu
If you could upload a binned image, that could help us figure it out. Generally, low defocus images collected on CCD cameras which do not have either carbon or a lot of sample on them will not show a lot of Thon rings in CTFFIND3. If you see two zeros clearly though, that should be good enough, as long as the check images look good.
Axel