From: Axel Kohlmeyer (
Date: Sun May 11 2008 - 11:18:17 CDT

On Sun, 11 May 2008, Alexandre A. Vakhrouchev wrote:

AV> Hi all! Thanks for your response.

AV> As John have already said the unusual part is connected with CA atoms
AV> in carbon nanotube residue. I built it by myself giving all atoms
AV> unique names in form A, B,...,Z, AA, AB, ..., AZ and so on to use
AV> them in 'coord' command of psfgen. I didn't split tube into segments,
AV> that is why there is 202176 atoms of CNT with unique names. Here is a
AV> part of PSF file corresponding to CNT:

so you are a victim of your own script...

AV> 26932 NT 0 C VVNB CA 0.000000 12.0110 0
AV> 26933 NT 0 C WVNB CA 0.000000 12.0110 0
AV> 26934 NT 0 C XVNB CA 0.000000 12.0110 0
AV> 26935 NT 0 C YVNB CA 0.000000 12.0110 0
AV> 26936 NT 0 C ZVNB CA 0.000000 12.0110 0
AV> 26937 NT 0 C AWNB CA 0.000000 12.0110 0
AV> 26938 NT 0 C BWNB CA 0.000000 12.0110 0
AV> VMD 1.8.5 works properly with this PSF/PDB files and '[atomselect top
AV> "type CA"] get name' reads atom names. I was surprised that 1.8.6
AV> could not open it.

the data type to store indexes was made smaller to save a _lot_ of
memory when loading large structures. the unique naming that you are
doing is essentially defeating this purpose. there has to be a different
way of doing what you need without using unique names for atoms. you
can as easily refer to them via "index" or "serial" and that doesn't
need any additional storage.

AV> Waiting for your advice.

short of redesigning your script, why don't you try making a selection
with "resname NT and type CA" and then use "set name CNT" on that
selection to make all CNT atom names the same before writing the
.psf file from psfgen?



Axel Kohlmeyer
   Center for Molecular Modeling   --   University of Pennsylvania
Department of Chemistry, 231 S.34th Street, Philadelphia, PA 19104-6323
tel: 1-215-898-1582,  fax: 1-215-573-6233,  office-tel: 1-215-898-5425
If you make something idiot-proof, the universe creates a better idiot.