[gmx-users] pdb2gmx capabilities

[Peter I. Hansen]

David van der Spoel wrote:
> On Fri, 2005-06-10 at 11:52 +0200, Peter I. Hansen wrote:
> 
>>Hello users
>>
>>I'm looking into using gromacs for my work, but having some 
>>problems since pdb2gmx only seems to expect peptides.
>>
>>I have been trying to modify the gmxdemo/demo script to run 
>>molecular dynamics on glucose. Since the GLCA unit is not 
>>complete I've tried to hack it together by adding an H and OH 
>>from water, without much luck.
>>
>>I've looked at the prodrg website, which is nice, but not so 
>>usefull if you want to build a oligosaccharide with a certain 
>>conformation.
>>
>>So, my questions are these. Is pdb2gmx only intended for peptide 
>>work? Do any of you know of a way to build gromacs topologies 
>>from pdb/mol2/xyz format files?
> 
> you can use pdb2gmx for any polymer as long as you use the right
> building blocks (rtp and hdb files) and options (-ter to turn off
> protein-centric termini generation).
 >
> How useful the existing GLC* blocks are I don't know, they were used in
> the early 90ties in the Berendsen group.

The GLC* blocks are strangely incomplete. I tried to build a 
GLCA unit topology with PRODRG, and this gave me completely 
different charges.

In a paper from 1996 (journal of computational chemistry, vol. 
17, no. 8, pp. 1068--1084) Ott and Meyer parametrize the gromos 
forecefield for oligosaccharides, but I guess these parameters 
never made it into the program. The data from the paper are not 
accesible.

Do you recommend I try to build a new GLCA unit from PRODRG, or 
try to expand the existing one?

/Peter