[gmx-users] Purpose of the pair list

[Bob Johnson]

Hello everyone,
My system consists of an infinite nanotube, a single stranded DNA molecule and
some water molecules. I have a separate topology file (.itp file) for each
molecule (nanotube.itp, ssdna.itp, and water.itp). I am using the Amber force
field for the DNA and water. I am performing MD simulations with the generate
pairs option. No pair list is specified in the nanotube and water topology
files. However, there is a pair list in the DNA topology file.

Since I am using the generate pairs option, it seems that I should be able to
get rid of the pair list in the DNA topology file. However, when I remove the
list, some of the DNA bonds go crazy and the simulation crashes. The system is
completely stable if I leave the DNA pair list untouched. This doesn't make
sense to me because the generate pairs option should override any pair lists.
Is that correct? Does anyone know why removing the pair list in the DNA would
cause the simulation to crash?

I generated the nanotube topology using x2top. I generated the nanotube and
water topology using Amber's XLeap and then converted them to GROMACS format.
My simulation uses pbc = full to simulate an infinite nanotube.
Thanks,
Bob Johnson