[gmx-users] Restart files

[Bob Johnson]

Hello everyone,
My system involves an infinite nanotube, a single stranded DNA molecule, and
some water. After performing several simulations, the DNA molecule drifts along
the nanotube and half of the DNA ends up in the next periodic box. Now I want to
add additional water molecules (or methanol molecules) to the system. When I add
just one water molecule, it seems that the simulation loses memory of the fact
that the DNA on one side of the box is bonded to the DNA on the other side.
Normally when the DNA slides into the next periodic box, VMD will draw long
bonds stretching across the nanotube connecting the truncated DNA molecule with
the portion that now appears on the opposite side of the box. However, upon
addition of one water molecule, no bonds are drawn. It seems that the
simulation now treats the DNA as two truncated pieces instead of realizing that
they are, in fact, bonded together. Eventually the simulation crashes. All I am
doing is adding one extra water to the topology file like so,

Initial topology file entry:
[ molecules ]
nanotube       1
ssdna          1
water        496

Final topology file entry:
[ molecules ]
nanotube       1
ssdna          1
water        497

I then add the coordinates of one water molecule to the restart.g96 file. I then
perform a simulation using restart.g96 as my initial configuration. The
simulation proceeds as I have described above. What is going on here?
Thanks,
Bob Johnson