[gmx-users] [gmx-developers] modifying nonbonded interaction parameters

[Justin Lemkul]

On 4/9/14, 4:55 PM, ibrahim khalil wrote:
> Thanks for your reply Dr. Justin. But I still cannot get my job done
> completely.
>
> Inside my ffnonbond.itp, I added the part -
>
> [ nonbond_params ]
> CJ 	opls_111	1	.2918	1.9148
>
> ( I am using CJ for my carbon atomtype and opls_111 atomtype for the oxygen
> in water)
>
> My question is –
> 1.	With the above file contents, is .2918 my σco and 1.9148 my εco?  (I
> believe I haven’t modified it further somewhere else)

Yes, sigma is 0.2918 and epsilon is 1.914 (per manual section 5.3.2, which deals 
with the format and specifics of the [nonbond_params] directive).

> 2.	If I change the sigma and epsilon value for opls_111 does it have any
> impact on σco or εco?
>

No.  Remember the distinction here.  Values in [atomtypes] are fed into the 
normal combination rules for calculating sigma_ij and epsilon_ij for a pair of 
atoms i-j.  Those values can be overridden by the presence of an i-j combination 
in [nonbond_params].  If no such combination is present, normal combination 
rules prevail.

If you're having difficulty, you'll have to be more specific about what you're 
doing to diagnose and why things aren't working as you expect.  Gromacs only 
does what you tell it, and you need to define an extremely small test system 
(ideally something for which you can solve the energy by hand) to assess whether 
or not what you're defining is being interpreted correctly.

-Justin

-- 
==================================================

Justin A. Lemkul, Ph.D.
Ruth L. Kirschstein NRSA Postdoctoral Fellow

Department of Pharmaceutical Sciences
School of Pharmacy
Health Sciences Facility II, Room 601
University of Maryland, Baltimore
20 Penn St.
Baltimore, MD 21201

jalemkul at outerbanks.umaryland.edu | (410) 706-7441
http://mackerell.umaryland.edu/~jalemkul

==================================================