[gmx-users] amber99 with berger's lipids

[Mark Abraham]

On Tue, Apr 9, 2013 at 9:59 AM, James Starlight <jmsstarlight at gmail.com> wrote:
> By the way during simulation of the membrane-protein systems in the
> Amber99sb ff (with berger lipids) I've noticed decreased of my system in
> the Z-direction ( I've observed the same also during simulation of such
> systems in the Charm full atomic ff). In both cases the observed effect was
> seen in both GPU and non-gpu regimes with different cutoffs.
>
> It's intresting that in the gromos-united atom ff ( with vdw 1.2 cutoff
> with berger lipids) I didnt observe such decreasing in the Z - dimension of
> my bilayer.  The only difference was in the water models (spc in the gromos
> and tip3p in the amber or charm). Might it be relevant ? Should I switch to
> the spc water with the amber ?

On principle, that is a bad idea. The force fields are parametrized to
work with a particular water model, and you should only use another
one after careful testing that there is a suitable benefit that
outweighs any drawback. They are not plug and play.

You have observed two different systems show different behaviour
contracting in the Z direction. Why do you think one is abnormal and
one is not?

Mark