[MMTK] Solvation example

Konrad Hinsen hinsen at cnrs-orleans.fr
Tue Oct 14 16:45:34 CEST 2003

On Tuesday 14 October 2003 11:51, Michiel Jan Laurens de Hoon wrote:

> 1) The solvation is done by scaling up the universe, adding water
> molecules at random, and slowly scaling it back to its original size. Is
> there a reason for this, other than that it is easier to avoid
> overlapping water molecules when placing them randomly? I may just be

That's the main reason. Placing molecules at random without overlap at real density is next to impossible.

There are of course many other ways to set up a solvated system, each with its particular advantages and problems. This one is not the fastest one, but it is rather foolproof, it works with any kind of solute and most solvents, and when finished it gives a reasonably equilibrated structure, with little risk of residual crystal structure or other such artifacts.

> missing some basic molecular modeling background here. Does anybody know
> any good textbooks on these kind of things?

I can send some references when I am back home next week.

> 2) If I try to view the universe after running solvation.py, I get the
> error message "Universe version numbers do not agree" (complete
> traceback is printed below). For some reason, if I remove the lines

This is hard to analywe without the script that produced the traceback. In general, the message means that two different quantities referring to a universe were created before and after the addition or removal of objects (molecules or environment objects like thermostats).

This may in fact be a bug in MMTK but I can't say without looking at it in more detail. I also plan to relax the compatibility conditions a bit in the future.

> 3) If I then view the universe, I noticed that the protein usually ends
> up off to one side or in a corner (so one side of the protein does not
> have water but the vacuum around it). This is also the case if I
> increase the size of the box so I have lots of water (269 molecules). I
> was hoping to find the protein somewhere in the bulk of the water. Is
> something going wrong here? What is the function of solvation if the
> protein doesn't see most of the water?

First of all, you need to take into account periodic boundary conditions. If the protein is in a corner, it nevertheless interacts with the water at the "other end" of the box.

There seems to be a bug in the code that prepares coordinates for visualization, so sometimes you see periodic images of molecules instead of the "real" molecules. This doesn't affect calculations in any way, only visualization.


More information about the mmtk mailing list