[gmx-users] Intra x inter coulombic and van-der-waals terms

[Mark Abraham]

Jones de Andrade wrote:
> Hi!
> 
> Well, I was trying to calculate the intermolecular total energy of
> some long systems here, when I got locked in a problem.
> 
> g_energy allows me to isolate Coulomb-(SR), Coulomb-14 and
> Coul.-recip. for the coulombic terms, for example. The seems to be
> that, in long chain molecules (by long, I mean more than 1-4 maximum
> distance), the program outputs the 1-n (n>4) intramolecular terms of
> the interactions together with most of the other intermolecular
> coulombic terms in Coulomb-(SR) (exception for the molecules so far
> that are in the Coul.-recip. part, but my molecules are some times
> smaller than any cutoff radius used).
> 
> I would like to know if there is any work-around for this, in order to
> remove the 1-n (n>4) intramolecular terms from the other
> intermolecular total energies I get.

As you discovered below, removing them is not really what you want to do 
- you want to calculate the non-bonded interactions for atoms in your 
long molecules in two parts. Specifically, they are atom-atom 
interactions in the same molecule and atom-atom interaction in other 
molecules or solvent. The way to do this is using "energy groups" which 
you construct using make_ndx - please try make_ndx and read 
documentation for details. Once you've set each molecule up as a 
separate group, g_energy will report pairwise total interactions between 
all pairs of groups, which you can now add in whatever combination make 
sense for what you're trying to do.

Mark

> Just in time: I tried to increase, in the topology file, the nrexcl
> (to 9, the largest number of bonds between two atoms in the molecules
> simulated) parameter. It helped a lot, and started to yeld really
> better values. But, on the other hand, it also yelded very different
> trajectories, probably due to changes in conformation as a consequence
> of the lack of forces derived from those intramolecular interactions
> it excluded.  :(
> 
> Does anybody have any idea of how to deviate from this problem?
> 
> Thanks a lot in advance!
> 
> Jones
> _______________________________________________
> gmx-users mailing list    gmx-users at gromacs.org
> http://www.gromacs.org/mailman/listinfo/gmx-users
> Please don't post (un)subscribe requests to the list. Use the www 
> interface or send it to gmx-users-request at gromacs.org.
> Can't post? Read http://www.gromacs.org/mailing_lists/users.php
>