Hi Colin,
You can "work with magnification anisotropy corrected images without
issues up until
particle polishing". After polishing, however, you have two opposing effects:
(1) You have better dose weighting and trajectories, which improves the
resolution
(2) You lose magnification anisotropy correction, which degrades the
resolution
Whether the final resolution becomes better or worse depends on which
effect wins.
For example, if anisotropy is very large, Bayesian Polish will compare an
undistorted
reference with distorted movies, and resulting trajectories will be very
bad. Then
the effect #1 is small and #2 is large and probably you end up loosing
resolution.
If you have huge anisotropy and really want to apply both magnification
anisotropy correction
and Bayesian Polishing, you have to write anisotropy corrected but
non-motion corrected
movies and work on them. This requires lots of disk space, but should work
in theory (untested, though).
Best regards,
Takanori Nakane
> Thanks for the heads up! Does this mean I should process only the
> non-magnification anisotropy corrected images or that I can process with
> them, but when I perform particle polishing I wont be able to apply the
> correction? In other words, should I redo the alignment without
> magnification anisotropy correction and work with that data, or can I work
> with magnification anisotropy corrected images without issues up until
> particle polishing? Thanks for all the help!
>
>
> Colin Deniston
> UCSD Graduate Student
> Chemistry and Biochemistry Department
> ________________________________
> From: Takanori Nakane <[log in to unmask]>
> Sent: Monday, September 10, 2018 11:55:49 AM
> To: Colin K Deniston; [log in to unmask]
> Subject: Re: [ccpem] Relion 3.0 MotionCor2 Commands
>
> Hi Colin,
>
> You should put '-Mag x y z' to the "Other MOTIONCOR2 arguments" box
> in the "Motion" tab, not the "Additional arguments" box in the
> "Running" tab.
>
> The command line to MotionCor2 and its output are written in
> (subdirectories of) the job directory as <micrograph_name>.com and
> <micrograph_name>.out, respectively.
>
> Note that RELION 3.0 does NOT support anisotropic magnification
> correction. This means that the shiny particles after Bayesian
> Polishing revert to an uncorrected state. We understand this is a
> problem, but the refinement and correction of anisotropic
> magnification will be in 3.1, not 3.0.
>
> Best regards,
>
> Takanori Nakane
>
> On 2018/09/10 18:12, Colin K Deniston wrote:
>> Hello all,
>>
>>
>> I'm a graduate student running MotionCor2 through the new Relion 3.0 GUI
>> and I wish to run with an additional MotionCor2 input for magnification
>> anisotropy ("-Mag x y z"). I can not tell if Relion is actually passing
>> the command through and if MotionCor2 is using the input values. I have
>> tried adding "-Mag x y z" and "--other_motioncor2_args -Mag x y z" in
>> the additional arguments box in the Relion GUI, however I can not tell
>> from any of the files in the output whether Relion successfully passed
>> the command along. Relion has also output "The following warnings were
>> encountered upon command-line parsing: WARNING: Option -Mag is not a
>> valid RELION argument" both times. Any help in correcting how I input
>> the command and/or where I can look to see that the command was actually
>> passed through to MotionCor2 would be greatly appreciated. Thanks!
>>
>>
>> Colin Deniston
>> UCSD Graduate Student
>> Chemistry and Biochemistry Department
>>
>> ########################################################################
>>
>> To unsubscribe from the CCPEM list, click the following link:
>> https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=CCPEM&A=1
>>
>
>
########################################################################
To unsubscribe from the CCPEM list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=CCPEM&A=1
