[gmx-users] Re: generation of a box of water

[Jeroen van Bemmelen]

Hi Kat,

> Thank you for the suggestion. The difficulty is that I want to have a box 
> with a higher than average density, and gromacs seems to have problems with 
> this. I have tried the genbox -cs -o water.pdb -box <x> <y> <z> command and 
> have tried -nmol and -maxsol, but i can't get it to give me a dense enough 
> water box. Does anyone have a suggestion for getting around this? I want to 
> have a density of one molecule per 28 Angs^3.

Maybe you should work the other way around. Instead of trying to push 
as many water molecules in your box as needed for a certain density, 
you could try to adjust the box dimensions. Simply generate a box of 
water with genbox at 'normal' density containing enough molecules for 
your purpose, and afterwards resize it with 'editconf -density ...'. 
You may need to minimize and equilibrate your box first before 
resizing, although I don't expect that.

This might still give problems when you want your system to be 
extremely dense, in which case Davids method may be the only way. 
You'll probably have to use extremely high pressures for that though.

Also watch out that your .gro file doesn't specify water molecules as 
'HOH' (as generated by pdb2gmx) when using editconf, because in that 
case (in my experience) it doesn't count the hydrogens in the density 
calculation. For a correct calculation the water has to be of type 
'SOL'.

Hope it helps,
Jeroen