Not using SHAKE for the solute intramolecular hydrogen bonds may be okay
by using a small time step, 1 fs or less; but what about with TIP3P
waters. Wasn't the TIP3P water model designed on the basis of a
restrained geometry, namely using SHAKE? I assume that the DNA is
solvated in the simulation.
Good point. One would have to accept that it is no longer TIP3,
if it worked at all. My impression is that the restrained geometry
of TIP3 was intended to allow longer time steps, because given the
closeness of the H's to the single vdw sphere of the O, it is too
easy for the vdw-less H to stretch beyond the O and into another
atom's vdw and fuse. Perhaps with short enough time steps this
wouldn't happen; ideally this would depend on the innate maximum
O-H stretch of water. Otherwise, one would be forced to go to a
water model that has vdw on H's.
Bill Ross
Received on Wed Feb 28 2001 - 11:25:50 PST
