Re: [AMBER] Problem re. non-integer charge in glycoprotein

From: Dr. Anselm Horn <>
Date: Mon, 17 May 2021 12:42:04 +0200

Hi Rajas,

from what you described it seems that the sulfate moiety is the origin
of the non-integer charge; where do its charges come from?

IMHO it would be save to ignore the charge deviation, since it only
accounts for +0.008 a.u., and you probably have a sound literature
reference for it.
However, a practical workaround to obtain an integer charge would be to
modify the atomic charges of the sulfate moiety to account for the

But other may comment here as well.

Best regards,


Friedrich-Alexander-Universität Erlangen-Nürnberg

On 05/14/2021 06:24 PM, RAJAS MALLENAHALLI RAO wrote:
> Hello all,
> I'm using Amber to simulate a glycoprotein containing keratan sulfate
> glycans with sialic acid cap. I built the system using glycoprotein
> builder of, and I replaced the glycosylated Asparagine
> residues with NLN. I'm using ff14SB for protein & Glycam06-j1 for
> sugars. When I loaded the protein on tleap and checked for its charge,
> it gives me a non-integer value:
>> test = loadpdb prot1a.pdb
> Loading PDB file: ./prot1a.pdb
>   total atoms in file: 4343
>   Leap added 3978 missing atoms according to residue templates:
>        3978 H / lone pairs
>> charge test
> Total unperturbed charge: -49.992000
> Total perturbed charge:   -49.992000
> I tried to check where this is coming from, I tried on unglycosylated
> system, it gives integer value; on glycosylated system (w/o any charged
> residues as well as with sialic acid) it gives integer value. The
> problem starts when I add the sulfate groups. Is it advisable to fiddle
> with the NLN residue library to even out the non-integer value? or is it
> safe to ignore it altogether? Many thanks for suggestions/ help in advance.

AMBER mailing list
Received on Mon May 17 2021 - 04:00:02 PDT
Custom Search