[gmx-users] ApoA1 benchmark or reading CHARMM topologies in gromacs

[Mark Abraham]

On 28/09/2011 6:16 PM, Pedro Gonnet wrote:
> Hi,
>
> Short summary: Is there a gromacs version of the NAMD ApoA1 benchmark
> somewhere and if not, what tools can I use to convert CHARMM
> parameter/topology files to gromacs?

If either exists, I don't know about it.
> Now for the long version: I'm currently trying to "port" the popular
> ApoA1 benchmark used by NAMD and Desmond to gromacs. I have the input
> files for NAMD (they can be downloaded from
> http://www.ks.uiuc.edu/Research/namd/utilities/apoa1/), i.e. a PDB and
> PSF file as well as the corresponding .xplor force-field parameter
> files. I also have the underlying CHARMM topology files for the lipids
> involved.
>
> I've tried generating gromacs input files using pdb2gmx, using the
> CHARMM27 force-field therein. This doesn't work due to an "Atom OXT in
> residue GLN 243 was not found" error which seems to be an occurrence of
> bug #567 (http://redmine.gromacs.org/issues/567), yet it can't be
> circumvented by replacing the type of of the OT2 atom.

A possible work-around is to copy residuetypes.dat to your working 
directory and append "POPC Lipid" to it. I expect this will inhibit the 
mechanism responsible for the unwelcome atom renaming. Please let us 
know if so.

Alternatively, I believe another work-around is to follow the general 
procedure here 
http://www.gromacs.org/Documentation/How-tos/Adding_a_Residue_to_a_Force_Field, 
modifying the various parts of the POPC residue entry in the .rtp file 
to have atom names other than O1 and O2, and the PDB atom names to match.

Mark

> I tried using the OPLS force field instead (after replacing HSD with HIS
> and other minor modifications in the .pdb), but this fails with the same
> error as above. If I don't include chain ids in the pdb, then it fails
> since it does not recognize the POPC lipid residues. I have the CHARMM
> topology files for these, but I have no idea how to convert these to
> gromacs topologies.
>
> Now, since the ApoA1 benchmark is quite popular, I'm assuming somebody
> might already have a gromacs version out there... Is this the case?
>
> If not, Is there some tool to convert CHARMM input, i.e. parameters and
> topologies, to gromacs? I've found and successfully used Mark Abraham's
> Perl script to convert the parameters. Therein he refers to a script by
> Yuguang Mu to convert topologies, which apparently used to be on the
> gromacs site
> (http://www.gromacs.org/@api/deki/files/76/=charmm_to_gromacs.tgz), but
> which seems to have disappeared.
>
> Many thanks and kind regards,
> Pedro Gonnet
>
>