Date: Thu, 3 Nov 2011 06:39:47 -0400 (EDT)
Hi all,
to add to my previous quick estimate, here are simulation results on
switching off one SHAKE constraint:
The System is a benzene molecule, gaff parameterized in 818 TIP4PEW water.
Standard Temp and density equilibration was followed by 100 ps
unrestrained NPT dynamics. This was followed by 100 ps of NVE dynamics
with 1 or 2 fs timestep and with or without one C-H SHAKE contraint
removed. Total energy plots for the NVE step are attached, summarily total
energies are:
AVG RMS
1 fs -7546.6245 0.2297
2 fs -7547.4597 0.3608
1 fs noshake -7546.6529 0.2349
2 fs noshake -7547.1003 0.3931
rms increases ca 60% if the timestep is increased, but only ~5% if a SHAKE
constraint is removed. All four cases show more or less comparable and
acceptable energy conservation for all practical purposes.
C-H bond length are pretty much unaffected:
1fs 1.088 +- 0.03
2fs 1.089 +- 0.02
shaked 1.087 +- 0.0004
the longer timestep gives average bond length .2% above the minimum
distance compared to .1% for the shorter timestep. A plot of the actual
vibration (attached) shows that the oscillation is integrated a bit more
roughly, but with the same period and similar amplitudes. Visually, it is
impossible to distinguish the unshaked H from the other 5.
Admittedly, I should have done this check before suggesting that running
2fs timesteps in TI with missing SHAKEs will be ok...
Kind Regards,
Thomas
Dr. Thomas Steinbrecher
formerly at the
BioMaps Institute
Rutgers University
610 Taylor Rd.
Piscataway, NJ 08854
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
(image/png attachment: shake.png)
(image/png attachment: shake2.png)
