that seems unusual, but possible if the
rmsd decrease corresponded to a structural
transition that happened near the start of the
second run.
1) are you sure you used the same reference coordinates
for calculating RMSD for both trajectories?
2) compare the top few lines of the sander
output for each run to make sure that sander
indeed used the restart from the first
run as the inpcrd of the second run.
3) you might also compare the final energy from
run1 to the initial energy in run2 and see if they
are similar.
4) make sure it isn't due to some imaging, I don't
know what atom selection you use for the rmsd
but perhaps there is a re-imaging done in the
second MD start.
carlos
----- Original Message -----
From: "Youyi Peng" <pengyo.UMDNJ.EDU>
To: <amber.heimdal.compchem.ucsf.edu>
Sent: Thursday, May 08, 2003 11:02 AM
Subject: RMSD fluctuation too much
> Hi, AMBER fans:
> I tried to run 1 nm MD on a protein-ligand complex in vacco. After
> 465ps, the job stopped without any error message in the output file. I
> restarted the job as a continuous MD run (ntx=5,irest=1), 100ps later
> the job stopped again. I don't know what cause the job to stop. I used
> carnal to check the RMSD and found something abnormal. At the end of
> first MD run, RMSD is fluctuated around 3 A. After restarting MD, at the
> beginning of the second MD run, RMSD goes down to 1.5 A. The RMSD should
> fluctuated aroung 3 A in the second run if my system is stable because
> the coordinates and velocity were read from the restart file from the
> previous run. Could anybody give me some hints to solve the problem.
> Thank you in advance, Youyi.
> This is the input file:
>
> &cntrl
> nmropt = 0,
> ntx = 5, irest = 1, ntrx = 1, ntxo = 1,
> ntpr = 1000, ntwx = 1000, ntwv = 0, ntwe = 0,
>
> ntf = 1, ntb = 0,
> cut = 13.0, nsnb = 10, dielc=4.0
>
> ibelly = 0, ntr = 0,
>
> imin = 0,
> nstlim = 535000,
> nscm = 500,
> dt = 0.001,
>
> temp0 = 300.0, tempi = 300.0,
> heat = 0.0,
> ntt = 1,
> tautp = 2.0,
> vlimit = 20.0,
>
>
> ntc = 1, tol = 0.00001,
> &end
>
> &ewald
> eedmeth=5,
> &end
>
>
>
>
Received on Thu May 08 2003 - 16:53:01 PDT
