Dear James,
> I want to perform list of full atomic simulation of protein-ligand
> complexes in Gromacs using Amber99sb force field and ligand parametrized by
> means of antechamber (acpype).
>
> 1- From antechamber tutorials I've found that GAFF is used for such ligand
> parametrization. What cut-offs for electrostatics as well as vdw should I
> use for the modelling of protein-ligand complexes ? (assuming that I've
> used 1.0 cutoffs with such systems without ligands)
'cutoffs of 1.0' seems quite small - what about using the default value?
I am sure you can find discussions in the Amber mailing list archive
about cutoff values...
> 2- How I can assign charges more carefully ?
More carefully than 'what'? Antechamber handles various models of
atomic charges; RESP, ESP?, AM1-BCC, Mulliken?
You might be interested in using the R.E.D. tools and/or R.E.D. Server
to derive RESP or ESP (i.e. molecular electrostatic based) charge
values...
> Could you provide me with
> some example of such charges assignment based on different chemical
> compounds?
The 'building block' approach developed in the R.E.D. tools and/or
R.E.D. Server has been specially designed for deriving charge values
for a set of molecules belonging to a family of molecules with common
and variable parts.
Let's take an example of 10 molecules with the R1 common part and the
S1-S10 variable parts:
R1-S1
R1-S2
...
R1-S10
Using R.E.D. one can derive charge values for these 10 molecules by
defining 11 elementary building blocks:
R1-x y-S1 (x & y are the connecting groups)
R1-x y-S2
...
R1-x y-S10
By defining specific charge constraints for the x and y connecting
groups one can generate the R1 and S1-S10 molecular fragments that are
combined into the 10 wanted/target molecules:
R1 + S1 ---> R1-S1
R1 + S2 ---> R1-S2
...
R1 + S10 ---> R1-S10
In this approach the conformation(s) of the R1-x and y-S1/10 building
blocks is/are fully controlled, the geometry optimization step is
often far shorter and the charges for the R1 group are empirically
defined and are common in the R1-S1/R1-S10 molecules...
regards, Francois
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Tue Apr 09 2013 - 03:30:03 PDT
