[gmx-users] Re: Ligand_charges

[Mark Abraham]

pkmukher wrote:
> Dr Spoel,
> 
>   Thanks for your suggestions. Actually i would be using the
> GROMOS96 force field(option 0 in pdb2gmx GROMOS96 43a1).I am
> new to the GROMACS package.I wish to simulate a
> ligand-protein system and had been using the drug-enzyme
> tutorial available for this purpose. In the tutorial they
> use the PRODRG2 server to generate the ligand structure
> files. I find that the charges were calculated for a
> structure with only polar hydrogens present. This is then
> used to form the ligand-protein complex and solvation,
> neutralization etc.I was wondering
> 
> 1) which ligand file should i use to generate my gaussian
> inputs. one with all the hydrogens or one with just the
> polar hydrogens.

Well the latter won't get you very far.

> 2) I presume that for the gaussian calculation I have to
> submit an all hydrogen structure. In that case how do I
> refit the charges onto the polar hydrogen only structure
> which i will use for the generataion of the ligand-protein
> complex(as per the tutorial)

Not a simple question. Force field extension is a dark art and very easy 
to do badly. You can search the mailing list for advice people have 
given before on this topic, but it's mostly been me and Tsjerk, and we 
mostly tell people they are crazy to start their MD experience 
parameterizing an extension for a force field. That doesn't seem to stem 
the flow of (under)graduate students being set tasks that are likely to 
be beyond their capacity to do without expert supervision - and if they 
had that they wouldn't need to ask basic questions about parametrization 
on a mailing list. Nonetheless, you must read up on how the rest of the 
force field was developed in order to come up with an even vaguely 
reasonable strategy for extending it.

> 3) Do you think that the current forcefield is good for
> simulation of ligand-protein complexes.

You should find some relevant literature and read it yourself :-)

Mark