From: curcio (r.curcio_at_bioc.unizh.ch)
Date: Thu Oct 05 2006 - 10:59:10 CDT
Hi all,
as done for decaalanine by Schulten and coworkers I wanted to do some SMD
simulations pulling part of a (small) system first in one direction (from r_0
to r_1) and then in the opposite one (so from r_1 to r_0) using various
pulling velocities. The final aim is to see at which pulling velocity pulling
is a reversible proccess, i.e.
distrib(Work_forward_pulling)=distrib(Work_backward_pulling).
However, it is not clear to me (and not from the namd-tutorial files available
for the decaalanine case either) how the backward pulling has to be
performed.
In principle there are two possible ways:
1) starting backward pulling directly from the end of the forward pulling (so
rom r_1), i.e. using the restart files of the latter SMD-simulation and just
modifying the pulling direction to start the other SMD-run.
2) equilibrate the system at the final position of the forward pulling (so at
r_1) first and then start backward pulling from members of the
equilibration-ensemble, i.e. as done for the forward-pulling where
equilibration was done around r_0.
First test-runs have shown that these two ways don't give the same result,
i.e. force-extension curve, work-profiles etc. differ considerably.
Which of the two ways makes more sense (or is there yet another ("better")
way)?
Thanks and best regards,
Raffaele
This archive was generated by hypermail 2.1.6 : Wed Feb 29 2012 - 15:42:39 CST