On Wed, Sep 16, 2015 at 9:16 AM, Ruth Helena Tichauer <rhtichau.laas.fr>
wrote:
> Hello David,
>
> Thank you a lot for the answer. And yes it’s “three” and not “tree”, sorry
> for the repeated spelling mistake.
>
> I’ll keep minimising the hole system then, I’m not aware of how small the
> GMAX and RMS values should be to consider that a system has been
> sufficiently minimised.
>
> I watched both trajectories one from a QM/MM calculation and one from a MM
> calculation as you advised me to do, both molecular dynamics simulations
> starting from the same minimised structure.
>
> In the QM/MM one, the bonds of the three water molecules become extremely
> long, as well as some other bonds (that I didn’t manage to identify yet..)
> leading to a blow up of the structure. The water molecules appear as
> triangles, as if there was a bond between the hydrogen atoms (!!).
>
There *is*, if you're using one of the TIPnP water models. It's a rigid
water model, so the 3 bonds correspond to the 3 constrained distances.
However, all MM bonds are removed from the QM region, so this is not a
cause for concern (the bonds still exist in the prmtop file, which is why
VMD draws them).
>
> In the trajectory from the MM calculation, one of the three water
> molecules slips away from the catalytic site and the rest of the system
> (protein+ligand+Mg ion+two other water molecules) remain quite stable. I’m
> surprised to see the water molecule diffuse away from the rest of the
> cluster as when comparing several minimised structures it seems that the
> assembly is stable and the positions of the water molecules don’t change
> significantly..
>
Why? Water exchanges readily in most processes like this (simple entropy,
after all)... Keep in mind that minimization and MD are different -- the
former simply reduces to a local minimum, whereas the latter allows
particles to move.
The "variable solvent" QM/MM approach was designed to deal with solvent
exchanging by updating the "QM" waters as they exchange with MM waters.
HTH,
Jason
--
Jason M. Swails
BioMaPS,
Rutgers University
Postdoctoral Researcher
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Wed Sep 16 2015 - 08:30:07 PDT
