[gmx-users] Ice box

[Mark Abraham]

Fernando Mattio wrote:
> Dear gromacs users,
> I would like to know if it is possible with gromacs to create a box not 
> with water and protein, but only with ice. I couldn´t generate a 
> topology of the ice in the prodrg sever, and the command pdb2gmx didn´t 
> work... So I am wondering if it is possible...

Find an X-ray diffraction paper that determined the crystal structure, 
and use that to build an algorithm to generate an ice structure. That 
would create "ideal" ice, but inevitably real ice has imperfections 
because the cooling process was not infinitely slow and the crystal 
growth was thus seeded.

My guess is that it is close to vanishingly unlikely that a 
general-purpose forcefield for a multi-center water molecule would 
actually model the behaviour of crystalline water accurately, much less 
the density expansion below 277K. Thus getting a box of water and 
cooling it to 273K slowly isn't likely to work.

pdb2gmx will only have a chance of working if you can generate an ice 
coordinate file some other way.

Mark Abraham