On Thu, Jul 17, 2014 at 11:02 PM, Niels Geudens <Niels.Geudens.ugent.be>
wrote:
> Hi all,
>
> I would like to see the cis-trans isomerisation of a proline dipeptide
> (with non-proteinic caps). For this I am trying to use scaled MD in
> Amber14. I first ran a simulation without the scaledMD flag in the
> input file:
>
> &cntrl
> iwrap = 1
> imin = 0, irest = 1, ntx = 7,
> ntb = 2, pres0 = 1.0, ntp = 1,
> taup = 2.0,
> cut = 10.0, ntr = 0,
> ntc = 2, ntf = 2,
> tempi = 300.0, temp0 = 300.0,
> ntt = 3, gamma_ln = 1.0,
> nstlim = 5000000, dt = 0.002,
> ntpr = 100, ntwx = 500, ntwr = 1000
> ioutfm = 1
> /
>
> This simulation ran without any errors, but no isomerisation is
> present. Next, I added the scaledMD (with lamba = 0.7). At first, this
> simulation also runs without any problems. However, after a variable
> number of frames (I retried a few times), the molecule seems to
> "disappear" when viewed with vmd. The last meaningfull entry in the
> scaledMD.log file is:
>
> 500 354500 -9646.920713734136 0.700000000000
> -6752.844499613895 2894.076214120241
> 500 355000 NaN 0.700000000000
> NaN NaN
>
Your simulation went off to NaNdyland (i.e., it blew up). There are many
things that could cause this. Since scaled MD is quite new, it's entirely
possible that there is a bug. It could also be atoms collapsing onto each
other as a consequence of either the input parameters or force field
parameters you used.
However, there are currently 3 different implementations of Scaled MD in
Amber 14 -- the one in sander, the one in pmemd (CPU), and the one in
pmemd.cuda (GPU). Which one are you using? Can you try one of the others
and see if you get similar behavior? (If the problem lies with your
system, the CPU may take longer to blow up than the GPU version given its
higher tolerance for large numbers).
Good luck,
Jason
--
Jason M. Swails
BioMaPS,
Rutgers University
Postdoctoral Researcher
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Fri Jul 18 2014 - 00:00:02 PDT
