[gmx-users] Problem putting 2 identical peptide chains in a simulation box

[Mark Abraham]

> Hi All,
>
>    I am trying to put 2 identical peptide chains in a simulation box.
> However, gromacs tends to see both separate peptide chains as one.
>
>    The following is the method that I used, please comment and correct me
> if
> I'm wrong.
>
>    First, I translate the peptide to a certain distance. Then I put the
> .gro
> files of the translated peptide in the first .gro file. Later, I change
> the
> number of the residues and atoms of the second peptide accordingly and
> manually.
>
>    With the .gro file of the 2 peptide chains, I convert it to pdb file.

Not necessary. Almost any structure file type (.pdb, .gro, .g96, .tpr) is
accepted anywhere that a structure with atom names is required by GROMACS
utilities. This you will see in pdb2gmx -h if you look carefully towards
the end.

> And
> then, I use the pdb2mgx to convert the 2peptide.pdb to 2peptide.gro to
> generate topology file. The rest of the methods are the conventional
> gromacs
> simulations such as energy minimisation and position restrained.
>
>    However, gromacs see the 2peptide.pdb file as a file for 1 single
> molecule, (not 2 identical peptide chains in a file) and bond is created
> between this 2 chains. I tried to use -nmol to insert the second peptide
> in
> but that doesn't work. I wonder if anyone have done similar work like mine
> before, and willing to share the method with me? Any other suggestion and
> comments are most welcome.

There's an option for pdb2gmx which controls the merging of chains - see
pdb2gmx -h. Note that this might well require the use of .pdb format to
use one of its fields to demarcate the chains. Google for information on
that format.

However the solution Justin suggested is easily the most elegant. You'd
have to do the above if your chains were not identical.

Mark