Print

Print


Dear Pu,

What you see after postprocessing is not an increase in resolution but an effect of map SSNR weighting and sharpening. High-resolution features become more evident due to the sharpening, but the map is still the same resolution as in the end of your auto-refinement (4.326 A cutoff).
Now, as Sjors said, the fact that resolution stays the same after postprocessing probably means you need to adjust your masking parameters.
And regarding the size of your dataset, with different subsets resulting in the same resolution, you might want to check the B-factor applied during postprocessing in each case. Even if resolution doesn't change, you would prefer the map that requires less sharpening.

Hope this helps,






--
Ricardo Diogo Righetto

2018-05-09 17:38 GMT+02:00 Sjors Scheres <[log in to unmask]>:
Dear Pu,

As long as you've refined in C9, the postprocessing will take the C9
symmetry into account. Often it is the mask used for post-processing
that causes problems. Is the red line in the generated pdf logfile (the
FSC between the phase-randomised maps) around zero beyond 4-4.5A
resolution? If not, that needs fixing.

HTH,

Sjors



On 05/09/2018 04:03 PM, Pu Qian wrote:
> Dear All,
>
> I'm working on a particle which have C9 symmetry but only has 100 kDa
> MW. RELION works well before 3D classification. However, Auto3D_refine
> always gives a resolution at 4.326A exactly whatever dataset size I
> changed from 160k to 50k. More strangely, when I run PostProcess, the
> resolution reported is 4.326A as well, exactly same as before.
> However, when I viewed 3D models before and after PostProcess,
> resolution was improved obviously. After post process, I can see side
> chains clearly, and even pigment molecules in the protein. Resolution
> of the map is definitely higher than before post process, around
> 3.5-3.7A roughly as I estimated. I viewed FSC, in high resolution
> region, it is quite noise. Is this because RELION does not use C9
> symmetry for FSC calculation?   
>
> Any advice is greatly appreciated. 
>
> Best regards
>
> Dr. Pu Qian
> The University of Sheffield
>
>

--
Sjors Scheres
MRC Laboratory of Molecular Biology
Francis Crick Avenue, Cambridge Biomedical Campus
Cambridge CB2 0QH, U.K.
tel: +44 (0)1223 267061
http://www2.mrc-lmb.cam.ac.uk/groups/scheres