[gmx-users] charge group assignment with AMBER parameters

Justin A. Lemkul jalemkul at vt.edu
Fri Mar 19 23:32:59 CET 2010

Jacob Spooner wrote:
>>>> "Justin A. Lemkul" <jalemkul at vt.edu> 3/18/2010 9:35 AM >>>
> Jacob Spooner wrote:
>> I am attempting a simulation of a drug molecule using the ffamber/gaff within
>> gromacs.  When it comes to assigning charge groups the manual says that one
>> should group together nearby atoms and  the groups should have a net integer
>> charge.  Since AMBER charge parameters are obtained from QM calculations of
>> the specific molecule they do not form into integer charge groups like with
>> OPLS whose charge parameters are dictated by atom type and stay the constant
>> in different molecules.  I was wondering if somebody who has worked with
>> AMBER in GROMACS could give some guidance with regards to charge group
>> assignment.
> A quick look through some of the .rtp files from the ffamber distributions 
> indicates that, at least for amino acid and nucleic acid residues, each atom is 
> its own charge group.  That probably has something to do with the underlying 
> theory of the force field (i.e., the derivation scheme did not use charge 
> groups), but you can look into the literature there to be sure.
> -Justin
> Thank you for your response.  I had also noticed that each atom is its won residue in the ffamber .rtp file.
>   I was under the impression that it was incorrect to have each atom as its own charge group as it 
> would give strange effects due to the cutoff.  I am not so clear on my understanding of charge groups
> but if I were using PME electrostatics would I obviously produce questionable results by having each atom
> as its own charge group?  Your comments are greatly appreciated

There shouldn't be a problem with having each atom in its own charge group.  The 
artifacts come when you charge groups are very large, such that atoms within the 
group may be within the cutoff, but the center of mass of the charge group 
itself is not within the cutoff, and thus interactions are not properly calculated.

There are several discussions within the list archive about such things; you 
should be able to find them by searching the mailing list with a query like 
"amber charge groups."


> -Jake
>> Thanks
>> Jake Spooner University of the Fraser Valley


Justin A. Lemkul
Ph.D. Candidate
ICTAS Doctoral Scholar
Department of Biochemistry
Virginia Tech
Blacksburg, VA
jalemkul[at]vt.edu | (540) 231-9080


More information about the gromacs.org_gmx-users mailing list