[phenixbb] Getting the triangulated solvent excluded surface of a protein (PDB) as a .obj file

[Francois Berenger]

(was posted on CCP4 bb also)

Hello,

What are the good software to do this?

I already know of Pymol and Jmol.

Both can export the result to a .obj file, which is interesting for
later processing.

However, Pymol's algorithm is not exactly what Connolly described,
cf.
http://www.mail-archive.com/pymol-users@lists.sourceforge.net/msg00179.html

Also, Pymol will save the surface with some translation and rotation
which are not present in the PDB read in, which is a real pain.

For Jmol, sometimes the surface is not a closed polyhedra, so
it is kind of useless for my purpose.

I know of MSMS, but I don't think it is that robust (I read some source
code where someone was using MSMS, the source code had many dirty things
in order to try handling the apparently many cases where MSMS was crashing).

Is there something freely usable for research, using beta shapes
internally, for example? That should be a robust approach.

A robust software would be much appreciated (not crashing, whatever
the PDB we give it as input).
Open source and not in Fortran would be heaven. ;)

Thanks a lot for suggestions,
F.