[gmx-users] 1-4 interactions in a small molecule are missing

[Lee-Ping Wang]

Hi everyone,

 

I've been trying to understand an issue I've had with a simple molecule in
Gromacs (just 20 atoms) where gen_pairs=yes, nrexcl=3 and all of the atoms
carry some charge.  The system is non-periodic with no cutoff.  This is a
minimal test case derived from a more complex system (a lipid bilayer) where
I first saw this issue. 

 

My issue is that there is no "Coulomb-14" or "LJ-14" energy term, and the
potential energy does not depend on my setting of fudgeLJ and fudgeQQ.  When
I run gmxdump on my .tpr file, it does not output any LJ14 interactions.  It
does print out the value of fudgeQQ but this value doesn't have an effect on
the potential energy.

 

I thought that this molecule should have some 1-4 interactions because there
are several pairs of atoms that are separated by three bonds, and both atoms
have nonzero charge and LJ parameters.  I also wrote a Python script to
calculate the Coulomb and LJ energy just for this system, and I confirmed
that Gromacs was behaving as if fudgeLJ and fudgeQQ were both set to zero.
My example calculations are archived and documented here:
https://dl.dropboxusercontent.com/u/5381783/zerotor_excl3.tar.gz

 

If someone more experienced with Gromacs could look at my files and explain
how I'm misunderstanding the generation and treatment of 1-4 pairs in
Gromacs, I would highly appreciate it.

 

Thanks,

 

-          Lee-Ping Wang