[3dem] inquiries about data processing

[Sigworth, Frederick]

Dear Mohamed,

It’s great to hear that you are doing cryo-EM at KAUST!
Here are some quick answers, which others might want to improve on…

Fred Sigworth

On Mar 26, 2019, at 10:32 PM, Mohamed A Sayed <mohamed.sobhy at kaust.edu.sa<mailto:mohamed.sobhy at kaust.edu.sa>> wrote:


If you would please advise on the following inquiries
Krios using K2 camera:
1- For super resolution images, is it better not to use binning during motion correction step?
Binning by 2 at the motion correction step, e.g. by the Fourier binning provided by MotionCor2, is a good idea for all but the very highest-resolution reconstructions.

2- In this case, does using binning on super resolution images affect the final resolution of the reconstructed 3D model?
Binning produces a very small decrease in SNR at spatial frequencies near Nyquist. I expect that at 3A resolution and 1A binned pixels there should be negligible deterioration compared to 0.5A unbinned pixels.

3- Is there a range of CTF value cutoff above which I should not include data?
See answer to the following question.

4- If my protein is spherical ≈ 17 nM in diameter, what is the suitable particle box size that I should use considering the CTF refinement later requires bigger box? what is recommended mask size?
Defocus causes dispersion of the high frequency components. They are displaced from the actual image by a distance of dz * lambda / d, where dz is the defocus, lambda the electron wavelength, and d is the resolution. So for d=3A and lambda=.02A, a defocus of 1.5um (15000A) -> displacement of 100A. So if your particle diameter is 170A, and you use defocus up to 1.5um your box should be at least 370A in size.  Clearly you have to limit the defocus values to keep a reasonable box size. And, to save time and storage, you can first do a reconstruction to lower resolution with smaller boxes.

5-How can multibody refinement be applied to a protein that is composed of a multimer of 2 subunits? or how to perform processing of a symmetric particle through subunit masking?
First see if you have a problem with heterogeneity in the first placel. For example, do C1 and C2 reconstructions and compare them.

6-How and in which step can processed data from two different folder be combined without reprocessing all the raw images again?
7- Can data acquired at 0.51 Angstrom/pixel and 0.53 Angstrom/pixel be combined and at which step?
You combine particle stacks, i.e. after CTF determination and particle extraction. For Relion the particles star files contain all the information needed for combining images at differing pixel sizes. So use the Relion function to combine the particle star files.


Thank you very much for your help and advice.​

_______________________________________________
3dem mailing list
3dem at ncmir.ucsd.edu<mailto:3dem at ncmir.ucsd.edu>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.ncmir.ucsd.edu/pipermail/3dem/attachments/20190327/62a7245c/attachment.html>