[Chimera-users] Best density map contour level

Thomas Goddard goddard at cgl.ucsf.edu
Thu Jul 8 11:51:01 PDT 2010

Hi Maria,

   Unfortunately I don't think your procedure for setting the contour 
level will work very well.  The problem is that the isosurface is not 
just the outside surface of the capsid.  Usually the virus genome has 
lower density then the protein capsid so there contour surface may show 
the virus as hollow, or maybe layers of RNA/DNA will appear.  So the 
expected volume is not going to be the volume of a solid ball.

   There are no well accepted standards for how to set the contour level 
when showing EM maps.  On the other hand there probably is no "right" 
answer.  A good level for looking at the protein capsid will be 
different from a good level for looking at the genome packing.  If a 
capsid feature of an icosahedral virus does not appear at all 60-fold 
symmetric locations then symmetry averaged maps will show lower density 
and you will need a lower contour level to see such features.  In 
practice I believe people analyzing map primarily choose the contour 
level by eye that best suits their analysis.


> Hi Tom,
> Thank you very much for your help.
> I'll measure the radius of my 2 d images afterwards I'll use the radius to calculate the volume of my virus and finally
> set the contour level so that this volume is enclosed.
> Thanks a lot,
> Maria
>> Hi Maria,
>>    Chimera shows the volume data as an "isosurface" meaning all points
>> on the surface are at a specific map value.  That map value is called
>> "level" in the volume viewer dialog.  The values in a map (volume data
>> file) determined by electron microscopy are often not normalized in any
>> way -- useful values could be 10000 or 0.1 depending on the map.  For
>> x-ray maps it is common to normalize so that standard deviation over the
>> unit cell is 1.0.  In EM there no unit cell so the standard deviation
>> would depend on how big a box surrounded the virus density -- so this is
>> not as useful a normalization.  There is no standard accepted EM map
>> normalization that I know of.  If you knew the expected volume occupied
>> by the structure you could set a contour level so that the surface
>> encloses that volume.  Here's an explanation of how to do that in Chimera
>> http://www.cgl.ucsf.edu/chimera/tutorials/eman07/chimera-eman-2007.html#part2
>> More often you just visually choose a level that you believe gives a
>> surface matching the envelope of the proteins in the structure.  The map
>> mean, standard deviation and root mean square value can be calculated
>> with menu entry
>>      Tools / Volume Data / Volume Mean, SD, RMS
>>      Tom
>>> Dear Chimera users,
>>> I'm using Chimera to inspect my calculated density maps of viruses. In
>>> order to compare different tree d reconstructions I would like to know
>>> what the 'Level' value in Volume Viewer dialog means. At present I'm using
>>> the level value 1 but I'm not sure if this is correct.
>>> Could you please give me some advice how to define the right contouring
>>> threshold to present my data in order to inspect it?
>>> Thank you for your help,
>>> Maria L.

More information about the Chimera-users mailing list