This sounds like the old problem that CHARMM had with 1:4 electrostatics
defeating angle terms and allowing a proton with no Lennard-Jones
parameters to collapse into an oxygen. If you can attach your topology and
coordinates this would be helpful, but before you do that try running your
minimization printing the energy at each step (ntpr = 1). If the
electrostatics are shooting through the roof that would corroborate my
guess. There are a number of solutions I could think of, but I'd have to
know more about the system first. If the problem is what I have guessed,
then you will need new parameters for the system at some level, but again
I'd have to know more about what is there.
If this is simply a problem related to the hydrogen:oxygen bond stretching
to permit the collapse, then you're probably OK if you just set the bond
stiffness to be very high during energy minimization and then run dynamics
with SHAKE. If the bonded parameters are all normal looking and this still
happens, though (which is pretty conceivable to me, as the phosphate is
going to have large charges on the oxygens as well as the hydroxyl protons
which will create a very strong electrostatic attraction), then you will
need some more rigorous parameter development.
Dave
On Sun, Nov 20, 2016 at 9:45 PM, jinfeng liu <jinfengliu1119.gmail.com>
wrote:
> Dear Amber developers,
>
> Recently I am doing molecular dynamics simulation for a molecule
> which contains a fully protonated phosphate group. The Antechamber is used
> to generate the parameters for the molecule.The atomic charges are fitted
> using RESP method in HF/6-31G* level. However, when I used these parameters
> to minimize the molecule, the hydroxylic hydrogen atom always had a bad
> contact with its neighboring oxygen atom which double bonded to the
> phosphorus atom. And this lead to the failure of MD simulation. I tried the
> BCC charges, but this problem was still existed. The initial structure of
> the molecule is 100% OK. So, can anyone tell me how to solve this problem ?
> Thanks in advance.
>
> Best,
> Jinfeng
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber
>
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Sun Nov 20 2016 - 19:30:04 PST
