[gmx-users] Strangeness in gro file

Mark Abraham Mark.Abraham at anu.edu.au
Tue Sep 13 08:12:49 CEST 2011

On 13/09/2011 3:27 PM, Sweta Iyer wrote:
> Hi all,
> I am trying to use the g_membed tool to insert my protein into a DMPC membrane.
> My protein is a dimer and hence I separated the two monomers by a TER card in the original pdb file before pdb2gmx step for gromacs to identify it as teo separate entities.. I then merged my protein file with that of the membrane and added water and ions to it. Then when I went ahead to make an index file I notice that the numbering of residues is not continuous in my gro file, ie, it starts from residue 1-29 of a monomer, then again 1-29 of the second monomer, 1-128 for the DMPC moleculaes and similarly for the SOL and ions as well. As a result, I cannot exactly specify residue number for my index file to specify the two protein groups as different as the residue numbers overlap. Even when i try to do it, i get an error message saying Atom 1 in multiple T-Coupling groups  Is there a way to re number them so that there is continuity or is there another way around to making the index file? I have not seen this the previously when I have made an index file. I am using gmx version 4.5.4 and the 53a6 forcefield. I am not attaching the gro file here as it is quite lengthy.

Only atom numbering within a [moleculetype] is relative to that atom. 
Index files need atom numbers from the whole system. These are 
constructed from the first [molecules] entry being 1 to n, then the 
second [molecules] entry n+1 to n+n (if that entry is the same 
[moleculetype], etc. The numbers assigned to atoms and residues in the 
coordinate file supplied to grompp are ignored. Only the atom and 
molecule ordering is significant (and must match the .top).

Do a grompp using an .mdp file that does not use any index groups (e.g. 
disable T-coupling), and pass that resulting .tpr to editconf to get a 
coordinate file back out. I expect the atom numbering will now be from 1 
to N where N is the number of atoms in the system. Use that coordinate 
file to work out your index groups (with make_ndx or otherwise). Later, 
you and grompp will be talking the same language. :)


More information about the gromacs.org_gmx-users mailing list