[Chimera-users] segmentation fault - surface rendering

[Elaine Meng]

Hi Mike,
Rather than a memory issue, it is likely a failure of the surface  
calculation.  Chimera uses MSMS for surface calculations, and we (and  
many bug reporters) have noted that it fails on a significant  
proportion of structures.  The developers have been working very hard  
to write code to replace MSMS, but it is not yet available.

The MSMS problems appear to be numerical instabilities sensitive to  
small changes in atomic radii, coordinates, surface vertex positions,  
and sometimes even the orientation of the molecule.  Thus for each  
structure you can generally experiment and find some change(s) that  
will allow successful surface calculation.  Unfortunately, there is  
no fix that necessarily works for all of the problem structures, and  
no way to tell beforehand which adjustment will be successful for a  
given structure!

If the surface calculation is failing, here are a few things to try:

(1) increase all atomic radii by some small amount, for example with  
the command:
vdwdefine +0.01

http://www.cgl.ucsf.edu/chimera/docs/UsersGuide/midas/vdwdefine.html

(2) change surface calculation parameters in the "New Surfaces"  
preferences (Favorites... Preferences, change to "New Surfaces"  
category).  For example, in one recent case, it was necessary to  
increase the probe radius to 1.5 and set "show disjoint surfaces" to  
false.  Remember to click Save if you want the settings to apply to  
your next use of Chimera.

http://www.cgl.ucsf.edu/chimera/docs/UsersGuide/preferences.html#New% 
20Surfaces

(3) add hydrogens  (Tools... Structure Editing... AddH) or command  
"addh" - I prefer (1) or (2), however, as adding hydrogens will make  
the surface bumpier

I don't know if you are calculating surfaces in some high-throughput  
fashion or concentrating on a few structures.  If the latter and you  
are having trouble finding a tweak that will allow surface  
calculation, I could take a look at the structure for you (let me  
know the pdb ID, if any, or send the structure).

I also don't know why splitting up your structure into multiple files  
might help, unless opening parts of the structure separately changed  
the surface geometry by making the surface enclose the individual  
parts rather than the entire thing  (that can also be controlled with  
the "surfcat" command rather than splitting the file).

Sorry for the difficulties.  We are all looking forward to the new  
surface calculation code!
Elaine
-----
Elaine C. Meng, Ph.D.                          meng at cgl.ucsf.edu
UCSF Computer Graphics Lab and Babbitt Lab
Department of Pharmaceutical Chemistry
University of California, San Francisco
                      http://www.cgl.ucsf.edu/home/meng/index.html


On Mar 13, 2008, at 11:37 AM, Michael DiMattia wrote:

> Hello, I am trying to surface render a few different proteins  
> within the same Chimera session and I keep encountering a  
> segmentation faullt.  I imagine this is due to a memory limit being  
> reached, although I do not know if it is a Chimera parameter that  
> can be increased/overriden.  Instead of having the different  
> macromolecules in the same pdb, I have actually split it up into  
> many pdbs and opened them one at a time, surface rendering one at a  
> time.  This worked for awhile but it seems I can't even get that  
> far now.
>
> Any advice or tips on what I can do?
>
> I realize surface rendering is very computationally intensive but I  
> would love to get this to work!
>
> Thanks
>
> Mike DiMattia
> _______________________________________________
> Chimera-users mailing list
> Chimera-users at cgl.ucsf.edu
> http://www.cgl.ucsf.edu/mailman/listinfo/chimera-users