[gmx-users] Re: restraints on water oxygen atoms

[Justin Lemkul]

On 6/11/13 9:38 AM, JW Gibbs wrote:
> Hi,
>
> As Justin had already suggested, restraints will lead to nasty atomic
> clashes. Looking at the gro file you have provided, it seems that the system
> is very poorly equilibrated.
>
> I had a similar issue. If the forcefield you are using is accurate enough,
> follow the steps as follows:
>
> 1. Take the system and increase the box size in all the directions. Run
> energy minimization. This will give a somewhat better starting point for the
> membrane.
>
> 2. Now play around with the box size, such that the x-y dimensions conform
> to the reported APL value. Eg., suppose for POPC, you have 48 Lipids per
> leaflet and say the experimental APL is about 5.7 nm^2 then the
> x=y=sqrt(48*5.7) nm. Adjust the z direction according to what thickness of
> water you would want to have.
>

Doing this assumes that the force field will reproduce experimental APL and, 
more often than not, introduces stress on the membrane.  The membrane will 
probably adjust, but I think this is going to slow equilibration more than anything.

> 3. Solvate the system (using genbox maybe?). Albeit some water molecules
> will penetrate the lipids. Do not worry about that. If your forcefield is
> accurate enough, water molecules would diffuse to the top after you follow
> steps 4.
>

Again, this is a barrier to equilibration.  I outline one possible solvation 
mechanism in my tutorial, and there are water removal scripts on the Gromacs 
website.  Water diffusion can be very slow and can seriously perturb the lipids.

> 4. Position restraint the membrane and then use semi-isotropic pressure
> coupling but DO NOT position restraint water. Run NPT for maybe 4-5 ns. You
> would see that water has diffused to the top and that you have a very good
> initial condition.
>
> As I said, it all depends on the correctness of the forcefield that you are
> using, but I think any Lipid forcefield will be parameterized such that the
> tails would have hydrophobic characteristics and you would not have to worry
> about water penetrating the bilayer.
>

Not always the case, but generally true enough.

-Justin

-- 
========================================

Justin A. Lemkul, Ph.D.
Research Scientist
Department of Biochemistry
Virginia Tech
Blacksburg, VA
jalemkul[at]vt.edu | (540) 231-9080
http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin

========================================