[Chimera-users] find the region around a volume imported from 3V voxelator

Elaine Meng meng at cgl.ucsf.edu
Thu Sep 9 15:21:34 PDT 2010

Hi Fabian,
I agree 3V is a very useful server for viewing pocket and channel volumes in Chimera.

Unfortunately in this situation, the zoning features in Chimera are based on distances to atoms -- you can do something to surface points based on their distance to atom(s) of interest, but not vice versa.  That is, I can't think of any simple way to select all the atoms that are within "X" distance of a surface.  In your case, the surface is an isosurface of the MRC volume data.  I had some ideas, but they may not be good enough...

Idea 1 (did not turn out to be useful):
Assign local density values from the MRC map to atoms with the "Values at Atom Positions" tool, select only atoms with values above a cutoff with "Select by Attribute".  Unfortunately the map in this case contains only 0 and 1, and none of it overlaps with the atoms, so they are all assigned 0 even if close to the surface.  I thought it might be possible to Gaussian smooth the map, and then use the smoothed-out one in this approach, but it seems nontrivial to figure out the appropriate settings for smoothing .

Idea 2:
The command "ac sd"  (to use keyboard shortcut "sd" in the command line) will show, in the Reply Log, the distance from every selected atom to the surface.  For each selected atom, you'd get something like this:
Distance from #0:18.water at O to surface MSMS main surface of 1a0m.pdb
 d = 2.09, surface point (3.14, -1.28, 7.85), side 1
Then you could save all those lines to some text file and do postprocessing to find the atoms within some cutoff.  Inconveniences with this approach are: you'd need to write your own little script to do the postprocessing, and there are 2 lines per atom rather than a more convenient single line, and if all atoms were selected, the output is large.  Instead of selecting all atoms for the measurement, you might try to narrow it down first, for example by selecting only those near the volume using the mouse, or by first selecting all alpha-carbons and doing the measurement, then from that output figuring out which full residues to select for a second round of measurements.  Perhaps this output could be processed into an attribute assignment file,
which would then allow selecting by value as mentioned above.  Selected atoms can be written out as a list or as a PDB file using the Actions menu or commands (writesel, write).

My only other idea is to take a look at what you get for this structure from Castp (for example, start Chimera, use command: open castp:3dhg) because that method automatically gives the atoms that form each pocket.  However, different methods have different definitions of pockets, and when I tried this there was no pocket that was exactly like what you see from 3V.  Pocket #285 from Castp is the most similar and is the same as about half of the 3V pocket in that channel-004.mrc file.  Castp results are automatically shown in a dialog in Chimera and there are options to select the pocket atoms or pocket+mouth atoms when you choose a row or rows in the dialog.

Sorry I didn't have some nicer solutions for this problem,
Elaine C. Meng, Ph.D. 
UCSF Computer Graphics Lab (Chimera team) and Babbitt Lab
Department of Pharmaceutical Chemistry
University of California, San Francisco

On Sep 9, 2010, at 11:26 AM, Fabian Glaser wrote:

> Daer chimera team,
> I am interested in finding a channel in a toluene hydrogenase, and
> extracting all the residues / atoms surrounding this channel. To do
> that I calculated the channel itself using the 3V voxelator
> (http://3vee.molmovdb.org), which produces a volume object that can be
> imported to chimera. See for example the output page
> http://3vee.molmovdb.org/viewResults.php?jobid=10sep02.ee6, which
> provides the file channel-004.mrc.gz, which needs to be first
> unzipped. When uploaded to chimera the volume is very clear and nice.
> But I want to find the residues/ atoms that form the channel, one the
> pdb itself 3DHG.pdb.gz, although the select and zone features seem to
> not work with the volume object.
> Any idea how to do that?
> Thanks a lot,
> Fabian

More information about the Chimera-users mailing list