[gmx-users] Non-zero charged system

[Erik Lindahl]

Hi Tividar,

> I want to simulate myoglobin witch has a non-zero charge (-1). This was 
> been detected by grompp.
> My question is:
> Should I run grompp and obtain the .tpr as an output, after that run 
> genion with the .tpr, from grompp, as an input and add 5 Na and 4 Cl and 
> obtain a .gro output file.

I wouldn't worry too much about it, and more important I probably 
wouldn't add ions. Although they would neutralize the system, you run a 
much higher risk of getting artifacts due to the ions not being in 
equilibrium positions unless plan to run the system for a microsecond or 
so.

If you place ions in random position it will simply take "forever" for 
them to get to good equilibrium positions, and even then you need very 
long simulations to get reasonably good sampling of the ion positions.


Still, if you want to include ions you don't need to recreate the entire 
topology, since the protein description is still identical. Just edit 
the main *.top file, include the file "ions.itp" at the top. Close to 
the end there is a system descripition like:

[ molecules ]
; Compound        #mols
Protein             1
SOL               488


Just add your ions there with lines like
Na 
		5
Cl 
		4



And you should be set. The genion program has already added them to the
coordinate file.


Cheers,

Erik