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

[Mark Abraham]

On 4/04/2011 6:11 PM, battistia at libero.it wrote:
>
> Dear all,
>
>
> I got some question about the implicit solvent.
>
> 1)  In gromacs, is it possible simulate a protein in a layer  of 
> explicit water, and put this system (protein + SOL) into a big box and 
> make the MD simulation with implicit solvent?
>
> How I have to set the md.mdp parameter  (; IMPLICIT SOLVENT ALGORITHM) 
> in this case?
>
> Do you know some tutorial about this method?
>

It can't work as simply as that, because the waters on the edge will fly 
off into the implicit solvent region. People have tried various things - 
check out the literature.

> 2) I'd like to put into my box a definied number of explicit number of 
> molecule of water eg. 100, so
>
> I used
>
> genbox -cp conf.gro -cs -maxsol 100  -p topol.top -o out.gro
>
>
>
> but the water molecule is not in the random position in the box, but 
> is in clustered conformation.
>
> Is it possible tell to genbox to put in a random way, in all space the 
> defined number of molecule?
>

What do you actually want - a uniform gas of a given density?

> 3) For that system (100 explicit solvent molecule + implicit solvent) 
> I generated the topol.tpr using the following set up into the mdp file:
>
>
> ; IMPLICIT SOLVENT ALGORITHM
> implicit_solvent         = GBSA
> ;GENERALIZED BORN ELECTROSTATICS;
> ;Algorithm for calculating Born radii
> gb_algorithm             = Still
>
>
> grompp do not give problem at all,  but mdrun give problem:
>
> segmantation fault or
>
> Norm of force     =            nan
>
> I think that the problem is the use of explicit water molecule and the 
> implcit water together.
>

Maybe. We haven't got enough information to know.

Mark