R: [gmx-users] implicit water and a layer of explicit water molecule

[Mark Abraham]

On 4/04/2011 7:12 PM, battistia at libero.it wrote:
> Dear Mark,
>
> about point 2, yes I need to have
> a uniform distribution of a defined number
> of water molecule (eg. 100 water molecule ) into my box.
> Is it possible with genbox?

Yes, but not by starting with a uniform distribution of a 
condensed-phase density. You need one of the right density to start with.

A better approach is to decide how large a box of what density you want. 
Work out how much volume that gives to each molecule. Take a single 
molecule and put it in a box of that size with editconf. Move the 
molecule a bit off-center. Then use genconf -rot to replicate that box 
into a large one. Then equilibrate that thoroughly to get rid of the 
residual ordering. If, later on, you want a different size, genbox with 
the box you've equilibrated here will be a good approach.

> After, I'll have to make the md simulation for my
> system in implicit solvent
> (I'll have  "protein + 100 molecule SOL + implicit solvent")
>
> So my next problem is to set the parameter into mdp file, for this mixed type of kind of water.

You have more problems than that. The force fields probably don't have 
implicit solvation parameters for water atom types. You'll need to 
source them somehow. And like I told you last time, your solvent 
molecules are not going to stay happily around your solvent like you 
hope. I'm going to stop repeating myself :-)

Mark