[3dem] anisotropic magnification distortion correction with mag_distortion_estimate/correct

Fri Aug 14 16:14:22 PDT 2020

Hi Yuejiao,

I'm not sure what you mean by many datasets have a distortion above 1.5%?
You mean data taken on the same microscope?  The drifting can complicate
the analysis, but looking at your images, I'm pretty sure it isn't causing
an issue for you, and you just don't have much distortion in them.

You can use mag_distortion_estimate to work with water rings, as it
asks for a resolution range to use for the analysis.  Simply change your
resolution to match the ring that you would like.  You will have to provide
a stack of images which provide strong water / ice rings if you want to do
this.

Thanks,

Tim

On Fri, 14 Aug 2020 at 15:43, Xian, Yuejiao <yxian at miners.utep.edu> wrote:

> Hello, Tim,
>
> Thank you for your quick feedback. I did compare the rings and find that
> the change is very minor as you pointed out. However, I find it hard to
> believe that the distortion is so minor because many datasets were shown
> had distortion above 1.5%. I am wondering if the drifting would affect the
> estimation, as you can see that the ring is much brighter in the horizontal
> direction?
>
> I am also wondering if the mag_distortion_estimate can work with the water
> rings Marin mentioned in his email? The water ring appears at  3.6Å, 2.2Å
> and 1.8Å. If I was right, the mag_distortion_estimate only consider the
> area encapsulating the gold rings (2.4Å-2.0Å), then I should be able to
> perform estimation on the 2.2Å water ring?
>
> Many thanks again for your time.
>
> Yuejiao
>
> ------------------------------
> *From:* Tim Grant <timothy.r.grant at gmail.com>
> *Sent:* Friday, August 14, 2020 12:53 PM
> *To:* Xian, Yuejiao <yxian at miners.utep.edu>
> *Cc:* 3dem at ncmir.ucsd.edu <3dem at ncmir.ucsd.edu>; Yuejiao Xian <
> xianyuejiao at gmail.com>
> *Subject:* Re: [3dem] anisotropic magnification distortion correction
> with mag_distortion_estimate/correct
>
> Hi Yuejiao,
>
> 0.15% is not a significant result, as it changes the rings so little, so
> the angle changing is likely to be a random effect. Both results are
> essentially telling you that there is very little to no magnification
> distortion.  A good way to confirm this is to overlay the rotationally
> averaged ring with the original ring and flick between them.  If there is
> appreciable distortion you should see a directional difference in the rings
> (it will look like they are being squashed in one direction).  In your
> case, as far as I can tell your rings line up well.
>
> Thanks,
>
> Tim
>
> On Fri, 14 Aug 2020 at 12:50, Xian, Yuejiao <yxian at miners.utep.edu> wrote:
>
> Hello, all.
>
> I am trying to estimate the anisotropic magnification distortion from a
> cryo-EM dataset, however, I am not sure if I got the accurate estimation. I
> appreciate any comments or suggestions.  Here is what I did:
>
> I stacked about 70 gold grating images into one mrc file, which I used as
> the input of the mag_distortion_estimate program developed by Timothy Grant
> and Nikolaus Grigorieff. Here is the output parameter:
>
> =====================================
>
> The following distortion parameters were found :-
>
>
>
> Distortion Angle     = 112.6
>
> Major Scale          = 1.001
>
> Minor Scale          = 0.999
>
>
>
>  Stretch only parameters would be as follows :-
>
>
>
> Distortion Angle     = 112.6
>
> Major Scale          = 1.002
>
> Minor Scale          = 1.000
>
> Corrected Pixel Size = 0.669
>
>
>
> The Total Distortion = .15%
>
> =========================================
>  The outputted ring images are attached. The ring on the left is the
> combined ring before any distortion correction, the one in the middle is
> the rotational averaged ring, the one in the right is the ring after
> distortion correction.
>
> To verify if the parameters are accurate,  I then took the three
> parameters ( highlighted in bold) and the same stack file as an input for
> mag_distortion_correct to correct the distortion. With the corrected stack
> file, I then estimated the distortion again, and it showed the images still
> have 0.15% distortion, with the following parameters:
>
> =========================================
>
> Distortion Angle     = 21.0
>
> Major Scale          = 1.001
>
> Minor Scale          = 0.999.
>
>
>
>  Stretch only parameters would be as follows :-
>
>
>
> Distortion Angle     = 21.0
>
> Major Scale          = 1.002
>
> Minor Scale          = 1.000
>
> Corrected Pixel Size = 0.669
>
>
>
> The Total Distortion = .15%
>
> =========================================
>
> Note that the values of both major and minor scale remind the same as
> before its corrected, but the distortion angle changed dramatically. Why
> could be the reason for this? If you had experience with the
> mag_distortion_estimate/correct, would you please let me know if there is
> anything that I should specifically pay attention to in order to get the
> accurate estimation?
>
> Many thanks in advance.
>
> Yuejiao Xian
> Ph.D candidate,
> Department of Chemistry and Biochemistry,
> University of Texas at El Paso
> _______________________________________________
> 3dem mailing list
> 3dem at ncmir.ucsd.edu
> https://mail.ncmir.ucsd.edu/mailman/listinfo/3dem
> <https://urldefense.com/v3/__https://nam01.safelinks.protection.outlook.com/?url=https*3A*2F*2Fmail.ncmir.ucsd.edu*2Fmailman*2Flistinfo*2F3dem&data=01*7C01*7Cyxian*40miners.utep.edu*7Cbf71f12181d44e5c337508d8408bba21*7C857c21d21a1643a490cfd57f3fab9d2f*7C1&sdata=uWDSgIaT8bl1AZpaXkB1NKEMkH7Ec5tWIvJT*2Bs0MRR0*3D&reserved=0__;JSUlJSUlJSUlJSUlJSU!!Mih3wA!V65gI6CXXx3urLfR3daXPK_-Mo8oHKXcL887PpAcOnTpNXIugMI0PXqESxPTnsF9zA$ >
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.ncmir.ucsd.edu/pipermail/3dem/attachments/20200814/98e25356/attachment.html>