[gmx-users] problem with grompp and atomtype

[Stéphane Téletchéa]

Anna Marabotti a écrit :
> Hi all,
> 
> I'm trying to simulate the MD of a protein with a ligand. Since it is the
> first time, I followed exactly (at least, I think so!) suggestions from
> "GROMACS tutorial for drug-enzyme complex" by J. Kerrigan. I started from
> the PDB coordinates of the complex, I extracted the coordinates of the
> ligand and created a .itp file for the ligand using the PRODRG server. Then
> I submitted the protein alone to pdb2gmx, and then I manually edited the
> .pdb and the .top files in order to re-introduce the coordinates of the
> ligand into the protein. Therefore, I opened the .pdb file of the protein
> created by pdb2gmx, and pasted the coordinates of the ligand obtained by
> PRODRG (changing the n. of residue to continue numbering), and I edited the
> .top file by adding #include "lig.itp" in the ff statement, LIG in the [
> molecules ] section etc. Then I created the box with editconf, solvated it
> with genbox and started grompp before neutralizing the -5 charge of the
> protein with genion.
> The grompp output gave me a fatal error:
> 
> Source code file: toputil.c, line 61
> FATAL ERROR: ATOMTYPE 'CH3' NOT FOUND
> 

There is nothing wrong in you procedure. For the moment, proDRG outputs 
atom types for the deprecated gromacs force field (ffgmx). You need to 
alter these atom types in order to get the proper ones (i.e. the ones 
coming from the force field you're using).

In a first *rough* approach, you can compare the atom types presents in
share/gromacs/top/ffgmx.atp (the one you have from proDRG)
to
share/gromacs/top/ffG45a3.atp (for the force filed G45a3 for instance)

Please pay great attention to atom types since changing the atom does 
not provide a straight-forward conversion from the old FF to the new one.

Atom types are not everything, you'll need to check bonds, angles, 
dihedrals, and overall, if your ligand is correctly described with these 
modifications on the itp file.

I'll add a section on the wiki when it's up to be more explicit (kinda 
sort of updating the tutorial).

Note there is a proDRG beta 2.5 mentionning gromacs 96.1 but for now i 
have no idea about the forcefield used, its validation state, and my 
question to Dr. van Aalten about this beta server version has not been 
answered yet (if someone on the list has an explanation, i'm opened to 
explanations).

Cheers,
Stéphane

-- 
Stéphane Téletchéa, PhD.                  http://www.steletch.org
Unité Mathématique Informatique et Génome http://migale.jouy.inra.fr/mig
INRA, Domaine de Vilvert                  Tél : (33) 134 652 891
78352 Jouy-en-Josas cedex, France         Fax : (33) 134 652 901