[gmx-users] ffG43a2x and ffgmx dppc.itp file

[Steffen Wolf]

P wrote:
> Dear All
> I have question for people who use dppc lipids with ffG43a2x.tgz force 
> field
> (GROMOS96 modified to describe lipids). I want to change to newer 
> GROMOS96
> force field (in order to simulate membrane/protein water systems) and 
> I would
> like to make *.itp file for dppc lipid molecule that is in agreement with
> ffG43a2x.tgz from the contribution section of gromacs website.
> ffG43a2x.tgz force field has two new atoms types LCH3 and LCH2 that can be
> used to describe long lipid chains. There are also new LJ parameters for
> those atoms in ffG43a2xnb.itp. What is the procedure for obtaining *itp
> for dppc molecule that can be used with this force field.
> 1) Should I use PRODRGbeta for creating dppc.itp topology and
> manually change atom types describing lipid long chains CH3 to LCH3 
> and then CH2 to LCH2.
> 2) Should I download dppc.itp topology from gromacs benchmark
> and manually change the atom types to those present in ffG43a2x.
> 3) Is there any available software that can be used for
> creation of dppc.itp file.
> 4) .. is there a simpler way?
> Please give me some hints and suggestion,
> I’m not an expert with gromacs, what is the
> right procedure for this.
> Thank you for your help.
> Best Regards.
Hi P,
there's an alternative way: at 
http://moose.bio.ucalgary.ca/index.php?page=Downloads of the group of P. 
Tieleman you find a topology and file set for dppc and lipids in 
general. Therein, dppc is described as a molecule consisting out of 
completely new atom types, but the usability of this set has been proven 
in a large amount of publications. Everything else depends on your own 
wishes: do you want to simulate a Protein in a membrane: In that case 
you'd have to take lipid.itp and exchange all interaction parameters 
between lipid- and protein atom types with the ones given in ffG43a2.itp 
(e.g. a lipid CH2, named LCH2, interacts with a protein CA exactly like 
a protein CH2 with CA). That's a big task and will probably cost you 
half a weeks work, but as the Tieleman parameters have been validated, 
the resulting parameter set should be a rather reliable one.
cheers
Steffen

-- 
Dipl.-Chem. Steffen Wolf
Department of Biophysics
University of Bochum
ND 04/67
44780 Bochum
Germany