Thank you Jason! I've changed ntt and performed new simulations and now it
seems all fine.
Best,
Guqin
2015-10-16 12:55 GMT-04:00 Jason Swails <jason.swails.gmail.com>:
> On Fri, Oct 16, 2015 at 12:48 PM, Guqin Shi <shi.293.osu.edu> wrote:
>
> > Hi Ross,
> >
> > thanks for the explanation. Let me re-interpret your email to make sure I
> > understand what you've mentioned...
> >
> > 1. I am using ntt=1 which does not use random numbers. Does that mean
> > whatever number I assign to ig, it doesn't matter at all? As long as I
> use
> > ntt=1, each run will always be identical? (all the rest just read .rst
> file
> > from last run)
> >
>
> Not quite. The temperature control does not depend on random numbers. So
> given a fixed set of starting velocities and coordinates, the random number
> generator will have no effect on dynamics.
>
> However, if you set irest=0, ntx=1, and tempi to some number *greater* than
> zero, then the velocities will be non-zero and set randomly from the
> Maxwell-Boltzmann distribution (so it *will* depend on ig for these initial
> velocities).
>
> 2. If I change to ntt=3, I also need to change tempi to non-0, for instance
> > 1 K, in order to use ig=-1. Right?
> >
>
> No. tempi=0 is perfectly fine for ntt=3.
>
>
>
> > 3. I don't have a specific reason to use ntt=1.
>
>
> Then don't use ntt=1. There are known problems with the Berendsen
> thermostat (it has been proven NOT to generate a correct canonical ensemble
> in the general case). ntt=3 is a much better choice.
>
>
>
> > If any, it might because I
> > learned and obtained input files from senior labmates... My system is a
> > very common, 100 residues, protein-small molecule complex. Currently in
> my
> > lab, I am using Amber 14. The final goal for these multiple independent
> > runs is to calculate binding energies. Can you give a common advice about
> > the ntt I should use for equilibrium and run under NPT? I found some
> > suggestion in tutorial (
> > http://ambermd.org/tutorials/basic/tutorial1/section3.htm) but it's for
> > NVT. Thank you very much!
> >
>
> The suggestions still apply to NpT and for the same reasons.
>
> HTH,
> Jason
>
> --
> Jason M. Swails
> BioMaPS,
> Rutgers University
> Postdoctoral Researcher
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber
>
--
Guqin SHI
The Ohio State University
College of Pharmacy
500 W. 12th Ave.
Columbus, OH, 43210
(614)688-3531
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Fri Oct 16 2015 - 13:00:02 PDT
