The last time I tried to do something similar, I found HOLLOW ( to be very useful. There's a nice tutorial for showing a binding site, rather than an interior cavity, here: .



On Fri, May 20, 2011 at 7:11 PM, Ben Roberts <> wrote:
Hi list,

I'm trying to use PyMOL to construct an image of the total contiguous contact
surface that starts at a particular place on the surface of a protein, without
all the visual distraction and clutter that comes from other regions of the
protein surface (not to mention internal cavities).

I started by rendering up the whole protein as "Contact surface" with
"Cavities and pockets only" shown. Unfortunately, this displays a large number
of bubbles, some of which, despite being far away in space from the region of
interest, manage to effectively obscure bits and pieces of it.

So my next thought was to try removing the extraneous bits, however, I can't
seem to select atoms in surface view; I have to display everything as lines or
similar. Removing things that way gets very unwieldy, not least because of all
the atoms that don't contribute anything to the surface but appear in front
of, behind, or otherwise near to the atoms that do so.

A third option I tried was to build the surface up, starting from the known
points of interest, but this too is painstaking because of the need to guess
where I should next expand the selection to in order to fill in the gaps.

A remaining method of getting only the cavity of interest would, I guess, be
to create a selection consisting only of those atoms that contribute something
to the currently rendered surface (in this case, cavities and voids). I could
then render those as spheres or something and hide unwanted contributors. Is
there a way of specifying such a selection? Alternatively, is there some other
smart way to achieve what I want (and by smart, I mean a way that isn't time-
consuming and unusually error-prone)? I've already tried Googling, to no
avail; I got, mostly, instructions about how to display surfaces in general.

I should mention that I'm using a recent version of MacPyMol, on Mac OS X

Any tips are welcome. Thanks!

Best regards,
For greater security, I support S/MIME encryption.

What Every C/C++ and Fortran developer Should Know!
Read this article and learn how Intel has extended the reach of its
next-generation tools to help Windows* and Linux* C/C++ and Fortran
developers boost performance applications - including clusters.
PyMOL-users mailing list (
Info Page:

Michael Lerner, Ph.D.
IRTA Postdoctoral Fellow
Laboratory of Computational Biology NIH/NHLBI
5635 Fishers Lane, Room T909, MSC 9314
Rockville, MD 20852 (UPS/FedEx/Reality)
Bethesda MD 20892-9314 (USPS)