[Chimera-users] how to reduce EM map resolution (Chimera 1.6.2)

Tom Goddard goddard at sonic.net
Tue Feb 12 10:21:59 PST 2013


Hi Sergey,

   You can lower the resolution of the EMDB 5168 (6.6 Angstrom) F-actin 
map by Gaussian filtering

     open #0 emdbID:5168
     vop gaussian #0 sdev 5 model #1

This will create a new map by convolution with a Gaussian that has 5 A 
standard deviation.
Then you can subtract this map from your map.

     open #2 mymap.mrc
     vop subtract #2 #1 minRMS true model #3

Here the "minRMS true" option says to scale the second map to minimize 
the residual (sum of
squares over grid points above second map threshold level).  This 
handles the different normalizations
of the two maps.  What standard deviation should you use in the Gaussian 
smoothing?  I suggest trying
different values and see which one gives the smallest residuals. The 
different definitions of "resolution"
used in EM make it hard to predict the exact standard deviation you will 
need to match the two maps.

   I also noticed that 5168 has some kind of cylindrical noise artifacts 
if you go to lower threshold levels.
These appear in the smoothed version of that map as a solid cylindrical 
shell at low threshold level.
Perhaps the level of that artifact is low enough that it won't bother 
you in your difference map.  But if
it does you may need to mask it out with the Chimera "mask" command and 
"shape cylinder" command.

     Tom



-------- Original Message --------
Subject: [Chimera-users] how to reduce EM map resolution (Chimera 1.6.2)
From: Sergey Ryazantsev
To: chimera-users at cgl.ucsf.edu
Date: 2/11/13 8:02 PM
>  Hello
>  I have low-resolution (~15A) model of decorated F-actin. I want to 
> localize the "decoration" element by subtracting normal F-actin model 
> obtained from Cryo-EM (EMD-5168) from my decorated model. The problem 
> is that EMD-5168 has a higher resolution and "vop" command did not 
> work.  Even if it would work, it would be difficult to interpret the 
> result of such subtraction. I think, both maps should be at 
> approximately the same resolution.  I know how to create 
> low-resolution map from PDB, but I could not find any solution how to 
> lower the resolution of the EMD-5168.map
>
> Any help would be greatly appreciated. Many thanks in advance. Sergey
>



More information about the Chimera-users mailing list