[gmx-users] Overriding "genpairs=yes" default for one specific molecule? Possible?

[Walter Ash]

First, I am sorry if this has come up on the mailing list before - a 
quick search didn't reveal anything obvious. 

I am in a position where I would like to use a molecular topology in 
conjuction with the OPLS-aa protein forcefield, but on first glance they 
seem to be incompatible.  

My problem is that OPLS uses genpairs=yes and fudgeLJ and fudgeQQ 
parameters of 0.5.  I would like to override this behavior for this one 
specific molecule - e.g. with exclusions.  However, after excluding the 
automatic calculation of 1-4 non-bonded interactions I need to specify a 
set of [pairs] to modify the dihedral functions.  Is this possible to do 
with the OPLS defaults, or will I have to reparameterize my dihedral 
functions for this molecule?

Unfortunately the gromacs manual does not very specifically explain the 
precedence of the different topology statements (e.g., do 'exclusions' 
override 'pairs' or 'genpairs=yes' directive?  Or does genpairs=yes take 
precedence over everything?).  My initial testing shows that the 
topology is unstable if I attempt to first exclude all 1-4 interactions, 
then define pairs, with genpairs=yes.