[gmx-users] [Fwd: DMF]

[Christoph Freudenberger]

Hi jiri,

-------- Original Message --------
Subject: DMF
Date: 11 Jun 2003 16:54:18 -0400
From: jiri vondrasek <jirka at uochb.cas.cz>
Organization: IOCB AS CR
To: christoph.freudenberger at chemie.uni-ulm.de

 > Hello Christopher
 > I am quite unexperienced user of GROMACS and unfortunately I have to
 > create my own solvent box for dimethylformamide. I hev noticed that you
 > have already prepared DMSO and acetonitrile solvent boxes. Would you
 > please be so kind and briefly tell me how to do it? Including
 > equilibration and and creation of input file?
1. Stop whining about having to do this jobs. Implementing solvent topologies
    is an excellent opportunity to learn what you are accually doing while
    performing MD simulations with gromacs ;-)
2. Search the literature for parameters for the solvent of interest.
    You need at least Lennard Jones parameters, charges and bond lengths and
    angles. If you want/need a flexible model you need the bonded parameters
    as well. Try to find as many papers as possible and select a few
    that reproduce best the properties you are interested in.
3. Think about the topology of your molecule. DMF is planar, right?
    Then it is not possible to keep it planar just by using constraints.
    It might be possible to use a masses/dummies approach similar to
    the my acetonitrile model. With a set of three masses it would be
    possible to "settle" the whole thing, which would be very nice for
    the performance of the simulations. The easy way is to use bond
    and angle constraints in addtion with an improper dihedral to keep
    the atoms in plane.
4. Setup the itp-file. You'll find detailed information about that
    in the manual and you can use the itp's avaivable from the gmx
    topology site as a template. Make sure to define new atomtypes for
    your model and "hardcode" the constraints into the itp, rather than
    use constraints=all-angles in the mdp-file.
5. Create a .gro or .pdb file of a single solvent molecule (using the
    PRODRG-server for instance). Create the solvent box from this
    file usind genbox, by putting some couple of hundreds of DMF's into a box
    that should be roughly 10% larger than you'd expect it to be from
    the experimental denisty.
6. Do EM and pressure coupled MD to equilibrate the system. You'll find
    detailed information about that in the manual and in the gmx
    tutorial.
7. Once the box size and the energy have converged, perform a 1-2 ns simulation
    to check the behaviour of your model. Check the density, DHvap and
    the diffusion coefficient, see if the rdf's look good and so on.
    compare your results to the exp. and MD literature values.


 > Sorry for wasting your time
You can take me to Uffleco for a couple of beers when
I get to Praha once in while ;-)

If you have further questions, I'll try to help, as long as you keep
the disscusion to the gmx-users list.

best regards
-- 
Christoph Freudenberger
Abt. Organische Chemie I AK Siehl - Uni Ulm -Tel: ++49-731-502-2785