[gmx-users] Max #atoms in a charge group: 68 > 64

[Mark Abraham]

Samuel Coulbourn Flores 花山 wrote:
> Hi Guys,
> 
> Thanks for past help received on this list.
> 
> I am having a problem which I hope I can get some insight on.  I am 
> trying to simulate Adenylate Kinase binding substrate analog AP5A (from 
> Shapiro et al).  The problem is that this ligand (which is basically two 
> Adenosines chained together with five phosphates) is a little on the 
> large side, with 68 atoms.  I get the following error on trying to run 
> genion:  
> 
> Max #atoms in a charge group: 68 > 64
> 
> Can someone explain  to me how the ligand is different from the protein 
> (which certainly has more than 64 atoms!)?  Is the protein not 
> considered a charge group?  Can I somehow tell gromacs not treat the 
> ligand as something other than a charge group?  Alternatively, can I 
> increase the max # atoms in a charge group?  I'm happy to let it run a 
> little more slowly if that's the only consequence.

Look for "charge groups" in the GROMACS manual. Ideally for MD, a 
molecule will have small groups of atoms whose total partial charge is 
zero, and which can be lumped together in some non-bonded calculations. 
"Small" will normally be a lot less than 64, hence the constraint in 
genion. So probably you should go back to your .top/.itp file and have a 
look at the [atoms] section and how the charge groups are working (see 
table 5.3 of the manual).

Mark