[gmx-users] Using CHARMM force fields in Gromacs, pt. 2

[Mark Abraham]

On 20/02/2012 4:56 PM, Jernej Zidar wrote:
> Hi.
>     Several months ago I was asking how to import and use two CHARMM
> forcefields in Gromacs at the same time in MD simulations. The two
> forcefields are CGENFF and CHARMM36.
>
>    I was able to add the polymer monomers and relevant parameters so I
> can now create a pair of top/gro files directly using the pdb2gmx
> utility. I have also added some lipid molecules to the Gromacs version
> of CHARMM36.
>
>     I was able to solvate the CGENFF-generated compound using the command:
> genbox -cp poly-1-1-1-1-1-new.gro -cs spc216.gro -p poly-1-1-1-1-1.top
> -o wpoly-1-1-1-1-1.gro
>
>    But I can't neutralize the system using the following command:
> grompp -f em.mdp -c wpoly-1-1-1-1-1.gro -p wpoly-1-1-1-1-1.top -o ions.tpr
>
>    To solve the problem I First tried including the spc.itp file from
> CHARMM36 but that didn't work because of the missing atomtypes in
> CGENFF's atomtypes.dat file. Oddly, adding the two missing atomtypes
> to ./cgenff/atomtypes.dat didn't help:
> Program grompp, VERSION 4.5.4
> Source code file: /build/buildd/gromacs-4.5.4/src/kernel/toppush.c, line: 1166
>
> Fatal error:
> Atomtype OW not found

grompp is trying to construct a topology, and so needs to be able to 
find a a suitable entry in [ atomtypes ] of ffnonbonded.itp. Later it 
will need to find other related parameters for the new atom types. 
atomtypes.atp serves other purposes (e.g. pdb2gmx).

Mark

>
>    Then I tried to include the forcefield.itp from CHARMM36 + CHARMM36
> parameters for SPC water. Again it didn't help:
> ERROR 1 [file forcefield.itp, line 1]:
>    Incorrect number of atomtypes for dihedral (0 instead of 2 or 4)
>
> ERROR 2 [file forcefield.itp, line 2]:
>    Incorrect number of atomtypes for dihedral (4 instead of 2 or 4)
>
> ERROR 3 [file forcefield.itp, line 3]:
>    Incorrect number of atomtypes for dihedral (4 instead of 2 or 4)
>
> ERROR 4 [file forcefield.itp, line 4]:
>    Incorrect number of atomtypes for dihedral (3 instead of 2 or 4)
>
> ERROR 5 [file forcefield.itp, line 5]:
>    Incorrect number of atomtypes for dihedral (4 instead of 2 or 4)
>
> ERROR 6 [file forcefield.itp, line 6]:
>    Incorrect number of atomtypes for dihedral (0 instead of 2 or 4)
>
> -------------------------------------------------------
> Program grompp, VERSION 4.5.4
> Source code file: /build/buildd/gromacs-4.5.4/src/kernel/topio.c, line: 668
>
> Fatal error:
> Syntax error - File forcefield.itp, line 13
> Last line read:
> ' 1	2	yes	1.0	1.0'
> Found a second defaults directive.
>
>    I get NO errors if I do a minimization of my compund in vacuo only
> using CGENFF.
>
>    Could someone please advice what is the best course of action if one
> wants to use CHARMM36 and CGENFF together in one simulation?
>
> Thanks in advance,
> Jernej Zidar