Re: [AMBER] Lipid14 simulation :MC Barostat vs Berendsen Barostat and NTR=1

From: Himanshu Joshi <himanshuphy87.gmail.com>
Date: Sat, 14 Nov 2015 01:33:56 +0530

Dear Ross,
Thank you very much for the clarification.
.
Regarding the ntr=1 question, here is my input file followed by some lines
from mdout file.
I am running pmemd.cuda.MPI with recent updates on AMBER14. After the
anomaly, I am
trying some cpu runs also, I will update you soon with the proceedings.

Lipid equilibration
 &cntrl
  imin=0,
  ntx=5,
  irest=1,
  ntc=2,
  ntf=2,
  tol=0.0000001,
  nstlim=1000000,
  ntt=3,
  gamma_ln=1.0,
  temp0=300.0,
  ntpr=5000,
  ntwr=5000,
  ntwx=10000,
  dt=0.002,
  ig=-1,
  ntb=2,
  ntp=2,
  cut=10.0,
  ioutfm=1,
  ntxo=2,
  ntr=1,
 /
 /
 &ewald
  skinnb=5, ! Increase skinnb to avoid skinnb errors
 /
Hold complex fixed
10.0 ! Force constant (kcal/(mol Angstroms^2))
RES 1 684
END
END




 NSTEP = 5000 TIME(PS) = 115.000 TEMP(K) = 300.59 PRESS =
-19.2
 Etot = -1354836.4560 EKtot = 232224.1094 EPtot =
-1587060.5653
 BOND = 11059.7680 ANGLE = 35413.1898 DIHED =
30219.4001
 1-4 NB = 10644.2777 1-4 EEL = -39686.4662 VDWAALS =
138379.2243
 EELEC = -1777768.0290 EHBOND = 0.0000 RESTRAINT =
4678.0700
 EAMBER (non-restraint) = -1591738.6354
 EKCMT = 94470.4812 VIRIAL = 96232.2583 VOLUME =
4260611.2260
                                                    Density =
0.9639
 ------------------------------------------------------------------------------


 NSTEP = 10000 TIME(PS) = 125.000 TEMP(K) = 301.34 PRESS =
-6.9
 Etot = -1355582.3661 EKtot = 232804.5312 EPtot =
-1588386.8973
 BOND = 11119.2994 ANGLE = 35156.3936 DIHED =
30306.2670
 1-4 NB = 10634.9034 1-4 EEL = -39622.6550 VDWAALS =
137181.2647
 EELEC = -1777996.0569 EHBOND = 0.0000 RESTRAINT =
4833.6864
 EAMBER (non-restraint) = -1593220.5838
 EKCMT = 94996.1801 VIRIAL = 95628.3471 VOLUME =
4219058.6682
                                                    Density =
0.9734
 ------------------------------------------------------------------------------




On Fri, Nov 13, 2015 at 10:41 PM, Ross Walker <rosscwalker.gmail.com> wrote:

> Hi Joshi,
>
> Yes we have seen similar behavior with the MC barostat vs Berendsen and
> are still looking into it to figure out what is going on. We think it is
> related to the use of anisotropic boundary with the MC barostat. For now
> the suggestion is to stick with the Berendsen barostat for Lipid bilayer
> simulations.
>
> With regards to the NTR=1 question - can you provide more details please.
> Your mdin and mdout files and also the command lines you are using for each
> stage of the run would be helpful.
>
> All the best
> Ross
>
> > On Nov 13, 2015, at 3:42 AM, Himanshu Joshi <himanshuphy87.gmail.com>
> wrote:
> >
> > Dear Friends,
> > While running POPC bilayer simulation in AMBER14 (LIPID14, pmemd.cuda),
> > I found very different Area per head group (APH) using two
> > different barostat. I ran two identical lipid (64 lipid per leaflet)
> > system.
> > I see very less APH for MC babrostat (61.2) as compared to Berendsen
> > barostat(~65.3 A) during the course of simulation. The berendson barostat
> > seems to reproduce the APH reported in Lipid14 force-field paper but not
> > the MC barostat.
> >
> > Is it expected ? Please refer to the attached figure for the variation of
> > APH with simulation time. Rest of the input parameters are exactly
> similar
> > except the barostat for both the systems. I follow the procedure given in
> > the tutorial by Benjamin Madej and Ross Walker on the official AMBER
> > website to do the simulation.
> >
> > Another thing I observed (in lipid +complex system) that if I use NTR=1
> > option (Rest input parameter similar to tutorial) to restrain the system
> > during the equilibration, the simulations are not stable.
> >
> > Can you please comment on above issues.
> > Thanks for the kind concern.
> >
> >
> > --
> >
> >
> >
> > *With Regards,HIMANSHU JOSHI Graduate Scholar, Center for Condense Matter
> > TheoryDepartment of Physics IISc.,Bangalore India 560012*
> > _______________________________________________
> > 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
>



-- 
*With Regards,HIMANSHU JOSHI Graduate Scholar, Center for Condense Matter
TheoryDepartment of Physics IISc.,Bangalore India 560012*
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Fri Nov 13 2015 - 12:30:03 PST
Custom Search