Marek,
> I was thinking that for my system might be a new approach:
>
>
> Gnp = Gdisp + Gcavity = Gdisp + SURFTEN * SASA + SURFOFF
>
>
> better that the original one ( Gnp = SURFTEN * SASA + SURFOFF ).
>
>
> Yes, the new approach clearly is better. Any simple correlation analysis
with TI simulated free energies would show this ...
> There is no problem with GAFF atom types as radii are probably
> automatically read from PRMTOP file.
> The problem is just with Glycam CG atom type. I proved it by changing CG
> to CT in prmtop file.
>
> That's correct.
>
> Anyway if I understood well the new approach for calculation of nonpolar
> solvent contribution is actually
> applicable (with sufficient reliability) probably just for standard atom
> types from proteins, nucleic acids force-fields
> ( ff99SB, ff99bsc0 etc. ) because only for such atomtypes are in this
> moment available optimised (Tan & Luo) radii which
> might be probably considerably different from those in PRMTOP file. Am I
> right ?
>
You are right. We haven't got a chance to calibrate or validate the model
for general organic molecules or sugars even if you may have used the Amber
force field types ...
Ray
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Thu Mar 17 2011 - 11:00:03 PDT
