[gmx-users] questions related to genpairs=no

[Berk Hess]

>From: David van der Spoel <spoel at xray.bmc.uu.se>
>Reply-To: Discussion list for GROMACS users <gmx-users at gromacs.org>
>To: Discussion list for GROMACS users <gmx-users at gromacs.org>
>Subject: Re: [gmx-users] questions related to genpairs=no
>Date: Tue, 14 Feb 2006 08:54:36 +0100
>
>Wusheng Zhu wrote:
>>Dear All,
>>I am trying to set combination rule =3 and genpairs=no in the [default] 
>>part. I have two questions:
>>(1) when I listed explicit [pairtypes] or [pairs], what parameters should 
>>follow each pairs atomA-atomB ? Are they simply epsilon_AB and sigma_AB  
>>calculated by combination rule, or should one of they be scaled 
>>multiplying with Fudge factors such as epsilon_AB*FudgeLJ and the other 
>>one simply be sigma_AB?
>>(2) Where to define FudgeQQ for 1-4 pairs Coulomb interaction as there are 
>>only two LJ parameters V,W (epsilon, sigma) following the pairs list based 
>>on the manual? So basically, how to put FugeLJ and FugeQQ into the 
>>topology file when genpairs=no?
>
>First of all you can check the result in the tpr file by using
>gmxdump -s topol.tpr | less
>and search for LJ14.
>
>In the pairtypes or pairs section you probably can not give sigma and 
>epsilon but have to compute the C6 and C12 directly. Please check using the 
>above method.

In Gromacs 3.3 you should give sigma and epsilon when using combination rule 
3.

>
>You only need to specify pairtypes if they are different than what would be 
>generated. FudgeLJ is *only* used with genpairs = no, otherwise you have 
>full LJ interactions. FudgeLJ will modify the C12 and C6 params directly in 
>the tpr file. FudgeQQ is used for 1-4 Coulomb interactions if you specify 
>it in the topology. You will find the number back in the parameter section 
>of the tpr file (and md.log file).

No.

FudgeLJ is *only* used with genpairs=yes.
And then only for pairs that are not listed in the pairstypes section,
i.e. those that are really generated from the normal non-bonded parameters.

Berk.