[gmx-users] Re: Thanks a million!

[Tsjerk Wassenaar]

Xiangyang Liu,

Although I'm very willing to help you in the sense that I'll try to
give you pointers to find the information you need or to try to track
where a certain mistake is, based on the knowledge I have of these
matters, I hope you do understand that I'm fully reluctant to do the
homework of someone else. Besides, if I fixed the stuff for you, you
wouldn't have learnt a thing and would have to ask someone else again
when facing the exact same problem.

That said, please note that generating a topology is not a trivial
matter. It's not something I can do in five minutes for a "new"
molecule. You have to decide which charges to use, which bonded
parameters (bonds, angles, dihedrals, improper dihedrals) and which
non-bonded parameters (atom types, pair interactions, exclusions).
Prodrg may make this seem easy, but then you still have to do some
work, in particular trying to convert the topology you get to one
which better fits the more modern force fields, such as Gromos 53a6
(ffG53a6*). However, if the properties of the molecule you want to use
in your simulation really matter to you, the parameters you get from
prodrg are likely to be insufficient. In that case, you will have to
do a proper parameterization: determine a charge distribution and
bonded/nonbonded/exclusion parameters which give you a good
correlation against experimentally determined parameters, such as the
free energy of solvation, etc.

I hope I have made myself clear and that this is of any use to you,

Tsjerk

On 11/2/06, Xiangyang Liu <shangdongliu at gmail.com> wrote:
> Hello Tsjerkw:
>
> Thank you very much for your attention.
>
> I am newer in gromacs but i have installed the gromacs in the linux system
> on my pc for several monthes. I have read carefully the topology  chapter of
> gromacs mannual. But the problem was that even the mannual said something i
> still didn't understand them very well because i am have no experience on
> dealing with such kind of problem.
>
> Now I want to do an investigate of the interaction of my molecular with
> membrane with molecular dynamics siumlation by gromacs. The structure of my
> molecular was clear and first i use the Hyperchem to build the pdb file, but
> when i convert the pdb file to top file using gromacs, there was also
> problems appeared. I can see the sturcture C=O convert to C-O when i use the
> hyperchem to save the structure of my molecular as pdb(ent) file. And then
> use spdbview to save the ent file as pdb file and delete the spdbv
> accessories at the end of the pdb file. When I use the gromacs to produce
> top file, There's always problem. I don't know why. Then i use the PRODRG to
> produce the gro file and the top file of my molecular, now there's also some
> problem like you know.
>
> I think you have lots of experience on dealing with such problem. Would you
> like to produce the top file and pdb file of my molecule ?
>
> Thank you very much!
>
> The accessory was the problem i met during my trying process.
>
> Best wishes,
>
> yours sincerely,
>
> Xiangyang Liu
>
>


-- 
Tsjerk A. Wassenaar, Ph.D.
post-doc
NMR, Utrecht University,
Padualaan 8,
3584 CH Utrecht, the Netherlands
P: +31-30-2539931
F: +31-30-2537623