From: Peter Freddolino (petefred_at_ks.uiuc.edu)
Date: Wed Oct 03 2007 - 08:19:30 CDT

Hi Lucie,
as noted in the documentation
(http://www.ks.uiuc.edu/Research/vmd/plugins/solvate/), the basic unit
of input for a nonstandard solvent to solvate is an equilibrated box of
the solvent of interest (in your case, a water/methanol mix in the
appropriate ratio). This box should be prepared by generating enough
water and methanol to fill a box that is, say, 30 or 40 angstroms on a
side, and then running an NPT equilibration on it to get a cube of the
solvent of interest. The psf and pdb file of this equilibrated cube are
needed for solvate, along with a selection text that occurs for exactly
one atom in each solvent residue (for a water/methanol box, "oxygen"
would be appropriate for this).

To prepare your pure solvent box, you should be able to generate a water
box with solvate, change the residue names and oxygen names of 30% of
the residues (assuming you mean 30% methanol as an absolute ratio and
not w/vol) to be those found in methanol, write a new pdb, and then run
it through psfgen.

Please let me know if you still have trouble.

Best,
Peter

lucie huynh wrote:
>
> Hi evryboby,
>
> I'd like to build a solvent box for a protein (30% methanol). I tried
> the options dedicated to nonstandard solvent but it seems that I did
> something wrong.
> I have a pdb and a psf file of methanol and its topology file. But my
> pdb file only define one molecule of methanol. Do I have to build a
> pdb file of the rigth number of methanol molecule needed for a solvent
> with 30% of methanol?
> I also did not understand the option "key selection" and box "padding".
> Do the nonstandard solvent option allow the generation of a solvent
> with different types of molecule?
> Or do I have to build one box of each (water and methanol) and then
> superimpose or combine them?
>
> Thanks a lot for any help or suggestion,
> Have a nice day!
>
> Lucie