If your nstlim is that high then you are not in serious danger from random
seed bugs. The problem is that the NET force imparted to each particle by
Langevin thermostating tends towards zero as 1/sqrt(nstlim). If nstlim is
10000 you'll see a big problem, 100000 you'll see a problem but may not
notice it, and 1000000+ the effect is negligible. Still, ig = -1 is
usually a good policy unless you are intent on exactly reproducing a
trajectory, it which case you should set ig to a specific sequence of
numbers (i.e. 8572, 8573, 8574, ...) for each segment.
Dave
On Wed, Nov 16, 2016 at 10:58 AM, Elvis Martis <elvis_bcp.elvismartis.in>
wrote:
> Hi,
> I guess you must first try what Daniel and Dave have suggested, maybe you
> can play around with the gamma_ln values later when you rule out everything
> else.
> all the best.
>
> Regards
>
> *Elvis Martis* Ph.D. Student, Bombay College of Pharmacy
> [image: photo] Email: elvis_bcp.elvismartis.in <elvis_bcp.elvismartis.in>
> Website: http://www.elvismartis.in <http://www.elvismartis.in/>
> group website: http://www.profeccoutinho.net.in/
> Address:
>
> Kalina, Santa Cruz [E]
> , Mumbai, 400 098, INDIA
>
> Alt. e-mail: emartis12.live.com
> <https://in.linkedin.com/in/elvisadrianmartis>
>
>
> On 16 November 2016 at 21:25, Thakur, Abhishek <axt651.miami.edu> wrote:
>
> >
> > Elvis thank you for such a quick response I am using default value for
> > gama_ln. so I guess it must be 0. Do you suggest someother value that I
> > should define>
> >
> >
> > David I don't think that I am using ig=-1, do you recommend me to try
> with
> > that?
> >
> >
> >
> > This is the script I am using in both case,
> >
> > Just with the difference of time step
> >
> >
> > imin=0,irest=1,ntx=5,
> > nstlim=2500000,dt=0.004,
> > ntf=2,ntc=2,ntb=0,
> > igb=8,cut=9999.0,rgbmax=25.0,
> > ntpr=500, ntwx=500, ntwr=10000,
> > saltcon=0.2,
> > temp0=300.0,
> > ntt=1,tautp=1.0,
> > nscm=0,ntxo=2,
> >
> >
> >
> > ________________________________
> > From: David Cerutti <dscerutti.gmail.com>
> > Sent: Wednesday, November 16, 2016 4:46:06 AM
> > To: AMBER Mailing List
> > Subject: Re: [AMBER] HMassRepartition
> >
> > These are both very short runs, and without knowing any other details
> > (assuming you have done everything correctly) I would expect the dimer to
> > open with 1fs time steps soon enough. Just to make sure, are you running
> > with a clock-based or otherwise varying random seed? I don't know what
> the
> > status of all our bulletproofing is, but if you're running with ascii
> > restarts then there could still be a way to get repeating pseudo-random
> > sequences into your system which would tend to unfold proteins if the
> > restarts occur frequently enough (or cause milder "opening" effects if
> the
> > restarts happen less frequently).
> >
> > Dave
> >
> >
> > On Wed, Nov 16, 2016 at 10:38 AM, Thakur, Abhishek <axt651.miami.edu>
> > wrote:
> >
> > > Hi,
> > >
> > > I have carried out two different GB/SA simulation, igb=5 and all the
> > > parameters same
> > >
> > > 1. With HMassRepartition with dt=0.004 and in 20ns I found that my
> dimer
> > > opened.
> > >
> > > 2.Simple Gb/SA with dt=0.001 and in 20ns I have found my dimer has not
> > > moved at all.
> > >
> > >
> > > I cannot get the reason behind this, as both have same duration of
> > > simulation. Can this be possible or I have done some mistake>
> > >
> > >
> > >
> > > Thanking you,
> > >
> > > Abhishek
> > > _______________________________________________
> > > AMBER mailing list
> > > AMBER.ambermd.org
> > > https://urldefense.proofpoint.com/v2/url?u=http-3A__lists.
> > ambermd.org_mailman_listinfo_amber&d=DgICAg&c=y2w-uYmhgFWijp_IQN0DhA&r=
> > 0Hah93XWYYBgmROOVcaccg&m=gKg7iBIwv0hggSRetsewgEBbJXuAmVQiURHsRTj_hbY&s=
> > QnS03sljOKJO8FUVe8Hs48ra_zzP4kJ4Qnm8wSJJJms&e=
> > >
> > _______________________________________________
> > AMBER mailing list
> > AMBER.ambermd.org
> > https://urldefense.proofpoint.com/v2/url?u=http-3A__lists.
> > ambermd.org_mailman_listinfo_amber&d=DgICAg&c=y2w-uYmhgFWijp_IQN0DhA&r=
> > 0Hah93XWYYBgmROOVcaccg&m=gKg7iBIwv0hggSRetsewgEBbJXuAmVQiURHsRTj_hbY&s=
> > QnS03sljOKJO8FUVe8Hs48ra_zzP4kJ4Qnm8wSJJJms&e=
> > _______________________________________________
> > AMBER mailing list
> > AMBER.ambermd.org
> > http://lists.ambermd.org/mailman/listinfo/amber
> >
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber
>
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Wed Nov 16 2016 - 08:30:02 PST
