[gmx-users] Re: charges

[Erik Lindahl]

Marc Baaden wrote:

>>>>Erik Lindahl said:
>>>>        
>>>>
> >> CHELPG is stable and good to very good even for large molecules. 
>
>On a more general note, I would not really agree with this depending
>on what you call a large molecule. We used to separate our molecules
>in smaller fragments for charge derivation, as the electrostatic
>potential might be biased towards a particular conformation and give
>more weight to residues on the surface of the molecule if you choose
>too large a fragment. This is of course only relevant if your molecule
>has freely rotatable groups, not if it is very rigid.
>
>You can also do several conformations of a fragment (eg cis, trans, gauche)
>and check what the effect on the charges is, to see whether you can/
>should average or have to make an informed choice based on the property
>you want to look at.
>

Hi Marc,

Sure - in the case I talked about I actually determined the charge for 
64 different lipid headgroup conformations and averaged....

And Merz-Kollman might very well be just as good; the main reason for 
not using it is that everything else in OPLS/AA is CHELPG.

Cheers,

Erik