this was asked before- you aren't using the same refc for all replicas. this
will not work unless you implement a Hamiltonian exchange method. it may not
be the only problem, but it certainly isn't going to be correct.
also a force constant of 5.0 is not a weak restraint like your label says,
it's a very strong one.
why do you have &ewald flags in the remd stage and not during equilibration?
On Sat, Dec 18, 2010 at 2:23 PM, Bongkeun Kim <bkim.chem.ucsb.edu> wrote:
> Quoting Carlos Simmerling <carlos.simmerling.gmail.com>:
>
> > we're still not sure what you restrained. did you use the same refc for
> all
> > of the remd runs? it isn't clear since you said below it was from the NPT
> > run- was that the equilibration right before remd, or something else?
> > why would you get a vacuum bubble? these are all NVT right?
> > On Sat, Dec 18, 2010 at 11:52 AM, Bongkeun Kim <bkim.chem.ucsb.edu>
> wrote:
> >
> I used one pdb file from the 100ns npt run and ran 200ps equilibrium
> md with the following input file.
> ++++++++++++++++++++++++++++++++++++++++++++=
> Equilibration
> &cntrl
> irest=0, ntx=1,
> nstlim=1000000, dt=0.0002,
> ntt=3, gamma_ln=1.0,
> temp0=XXXXX, ig=RANDOM_NUMBER,
> ntc=2, ntf=2, nscm=1000,
> ntb=1, ntr=1,
> cut=12.0,
> ntpr=500, ntwx=500, ntwr=1000,
> /
> Hold the GRA fixed
> 5.0
> RES 1 1346
> END
> END
> +++++++++++++++++++++++++++++++++++++++++++++++
>
> A part of group file looks like this.
> +++++++++++++++++++++++++++++++++++++++==
> -O -rem 0 -i equilibrate.mdin.001 -o equilibrate.mdout.001 -c
> gr3252_min2.rst -r equilibrate.rst.001 -x equilibrate.mdcrd.001 -inf
> equilibrate.mdinfo.001 -p gr3252.prmtop -ref gr3252_min2.rst
> -O -rem 0 -i equilibrate.mdin.002 -o equilibrate.mdout.002 -c
> gr3252_min2.rst -r equilibrate.rst.002 -x equilibrate.mdcrd.002 -inf
> equilibrate.mdinfo.002 -p gr3252.prmtop -ref gr3252_min2.rst
> -O -rem 0 -i equilibrate.mdin.003 -o equilibrate.mdout.003 -c
> gr3252_min2.rst -r equilibrate.rst.003 -x equilibrate.mdcrd.003 -inf
> equilibrate.mdinfo.003 -p gr3252.prmtop -ref gr3252_min2.rst
> ++++++++++++++++++++++++++++++++++++++++++++++++++
>
> The input and groupfile of REMD is following:
> ++++++++++++++++++++++++++++++++++++++++++++++++++++++
> Equilibration
> &cntrl
> irest=1, ntx=5,
> nstlim=500, dt=0.002,
> ntt=3, gamma_ln=1.0,
> temp0=XXXXX, ig=RANDOM_NUMBER,
> ntp=0, pres0= 1.0, taup=2.0,
> ntc=2, ntf=2, nscm=1000,
> ntb=1, igb=0,
> hybridgb=5, numwatkeep= 500,
> ntr=1,
> cut=12.0,
> ntpr=500, ntwx=1000, ntwr=1000,
> nmropt=0,
> numexchg=4000,
> /
> &ewald
> nfft1 = 96,
> nfft2 = 96,
> nfft3 = 96,
> order = 4,
> verbose = 0,
> ew_type = 0,
> nbflag = 1,
> skinnb = 2.0,
> netfrc = 1,
> vdwmeth = 1,
> column_fft = 1,
> /
> Keep GRA fixed with weak restraints
> 5.0
> FIND
> * * * GRA
> SEARCH
> RES 1 1346
> END
> END
>
> ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++==
> -O -rem 1 -remlog rem.log -i remd.mdin.001 -o remd.mdout.001 -c
> equilibrate.rst.001 -r remd.rst.001 -x remd.mdcrd.001 -inf
> remd.mdinfo.001 -p gr3252.prmtop -ref equilibrate.rst.001
> -O -rem 1 -remlog rem.log -i remd.mdin.002 -o remd.mdout.002 -c
> equilibrate.rst.002 -r remd.rst.002 -x remd.mdcrd.002 -inf
> remd.mdinfo.002 -p gr3252.prmtop -ref equilibrate.rst.002
> -O -rem 1 -remlog rem.log -i remd.mdin.003 -o remd.mdout.003 -c
> equilibrate.rst.003 -r remd.rst.003 -x remd.mdcrd.003 -inf
> remd.mdinfo.003 -p gr3252.prmtop -ref equilibrate.rst.003
> +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
>
> Please let me know what I can do to solve this problem.
> Thank you.
> Bongkeun Kim
>
> >>
> >> Quoting Jason Swails <jason.swails.gmail.com>:
> >>
> >> > The restraint energy is still outrageous, but I'm not quite sure what
> was
> >> > happening with your previous simulation without restraints. Since I
> >> didn't
> >> > write the hybrid remd code and I've never used it, I don't know how
> much
> >> > help I can be of here.
> >> >
> >> > However, Dave Case's point was a good one, and you never really
> answered
> >> > it. Did you use the same reference structure for each replica? If
> you
> >> > didn't, then this will cause your restraint energies to be very high
> as
> >> > well.
> >> I used the same structure from the single npt simulation about 100ns
> run.
> >> And I ran a group equilibrium md for all replicas in order to
> >> equilibrate each replica in each target temperature to avoid a vacuum
> >> bubble. the final rst files from this step were used to run remd. So
> >> do you think I have to run a hybrid REMD without running equil MD to
> >> use the same reference structure? If so, how do I avoid bubble in the
> >> space?
> >> Thank you.
> >> Bongkeun Kim
> >> >
> >> > Good luck!
> >> > Jason
> >> >
> >> > On Sat, Dec 18, 2010 at 12:00 AM, Bongkeun Kim <bkim.chem.ucsb.edu>
> >> wrote:
> >> >
> >> >> Hello,
> >> >>
> >> >> I applied this patch and used ntr=1 again.
> >> >> ...........................................................
> >> >> =================HYBRID REMD: energy calc for exch
> >> >> 1=================
> >> >> HYBRID REMD: Stripping waters
> >> >> wrapping first mol.: 83.13530 0.00000 41.15202
> >> >> HYBRID REMD: New natom= 4820
> >> >> HYBRID REMD: Calling force.
> >> >>
> >> >> NSTEP = 0 TIME(PS) = 127.600 TEMP(K) = 0.00 PRESS
> >> >> = 0.0
> >> >> Etot = 0.0000 EKtot = 0.0000 EPtot =
> >> >> 116431771.4632
> >> >> BOND = 22423.0668 ANGLE = 546043.1091 DIHED =
> >> >> 388.7305
> >> >> 1-4 NB = 52961.7061 1-4 EEL = 1814.5928 VDWAALS =
> >> >> -5494.9108
> >> >> EELEC = -6066.5669 EGB = -3141.0101 RESTRAINT =
> >> >> 115822842.7455
> >> >> EAMBER (non-restraint) = 608928.7177
> >> >> TEMP0 = 501.0000 REPNUM = 32 EXCHANGE# =
> >> >> 0
> >> >>
> >> >>
> >> >>
> >>
> ------------------------------------------------------------------------------
> >> >>
> >> >> HYBRID REMD: myEptot= ************ myTargetTemp= 501.00
> >> >> HYBRID REMD: Restoring...
> >> >> =========================END HYBRID REMD energy
> >> >> calc.=========================
> >> >> REMD: myEptot= 116431771.4632 myTargetTemp= 501.00 mytemp= 0.00
> >> >> ==========================REMD EXCHANGE
> >> >> CALCULATION==========================
> >> >> Exch= 1 RREMD= 0
> >> >> Replica Temp= 501.00 Indx= 32 Rep#= 32 EPot= **********
> >> >> Partner Temp= 285.10 Indx= 1 Rep#= 1 EPot= **********
> >> >> Metrop= 0.000000E+00 delta= 0.181951E+04 o_scaling=
> -1.00
> >> >> Rand= 0.821770E+00 MyScaling= -1.00 Success= F
> >> >> ========================END REMD EXCHANGE
> >> >> CALCULATION========================
> >> >> REMD: checking to see if bath T has changed: 501.00->501.00
> >> >> | # of SOLUTE degrees of freedom (RNDFP): 59256.
> >> >> | # of SOLVENT degrees of freedom (RNDFS): 0.
> >> >> | NDFMIN = 59256. NUM_NOSHAKE = 0 CORRECTED RNDFP =
> >> >> 59256.
> >> >> | TOTAL # of degrees of freedom (RNDF) = 59256.
> >> >> vlimit exceeded for step 8; vmax = 135.1554
> >> >> vlimit exceeded for step 10; vmax = 98.9017
> >> >> ...........................................................
> >> >>
> >> >> So, there is only one ****** on HYBRID REMD: myEptot= ************
> >> >> and EPtot is still large but is not ****. Because EPtot is too large,
> >> >> I'm afraid that exchanges occur or not.
> >> >> How do you think this output.
> >> >> Thank you.
> >> >> Bongkeun Kim
> >> >>
> >> >> Quoting Jason Swails <jason.swails.gmail.com>:
> >> >>
> >> >> > Hello,
> >> >> >
> >> >> > I think your problem lies in the use of coordinate restraints with
> >> hybrid
> >> >> > REMD. One of the steps that hybrid REMD has to do is reimage all
> of
> >> the
> >> >> > water molecules and place it at the location of its closest
> periodic
> >> >> image
> >> >> > since GB methods are inherently non-periodic and know nothing about
> >> >> > alternate images.
> >> >> >
> >> >> > One of the things done in "subroutine stripwat" in remd.f is the
> >> >> centering
> >> >> > of the entire system with respect to its center of mass. The
> reason I
> >> >> think
> >> >> > you are having issues is that the coordinates are adjusted by the
> >> center
> >> >> of
> >> >> > mass position, but the reference coordinates are never adjusted.
> >> >> >
> >> >> > I'm not positive this will work, but I included a patch which will
> >> adjust
> >> >> > the reference coordinates in exactly the same way as the normal
> >> >> > coordinates. Apply it to the latest, fully patched amber11 from
> >> >> AMBERHOME.
> >> >> >
> >> >> > cd $AMBERHOME
> >> >> > patch -p0 -N < hybrid_remd.patch
> >> >> >
> >> >> > Then recompile. Also please report back whether it worked or not.
> >> One
> >> >> > other thing to try that probably should've been mentioned before is
> >> >> testing
> >> >> > to see that it worked if you don't include restraints, just to make
> >> sure
> >> >> > that it is in fact the restraints that are causing the issue.
> >> >> >
> >> >> > Hope this helps,
> >> >> > Jason
> >> >> >
> >> >> > On Fri, Dec 17, 2010 at 11:43 AM, Bongkeun Kim <bkim.chem.ucsb.edu
> >
> >> >> wrote:
> >> >> >
> >> >> >> Quoting case <case.biomaps.rutgers.edu>:
> >> >> >>
> >> >> >> > On Fri, Dec 17, 2010, Bongkeun Kim wrote:
> >> >> >> >>
> >> >> >> >> I just run the hybrid remd and checked the output file.
> >> >> >> >> It said ********** on EPot. Does it mean only the number is too
> >> big
> >> >> to
> >> >> >> >> print or is there any problem on this run?
> >> >> >> >
> >> >> >> > There seems to be a problem with the restraint energy:
> >> >> >> >
> >> >> >> >> EELEC = -6589.2279 EGB = -3004.6220 RESTRAINT
> >> >> >> >> =57092146.9150
> >> >> >> >
> >> >> >> > You can't run replica exchange simulations with ntr>0 (unless, I
> >> >> guess,
> >> >> >> that
> >> >> >> > all replicas have the same reference structure. Also, be sure
> that
> >> >> you
> >> >> >> can
> >> >> >> > sucessfully run non-exchange simulations at each temperature of
> >> >> interest.
> >> >> >> >
> >> >> >>
> >> >> >> I already tried to run it with full explicit water and it gave me
> no
> >> >> >> problem. But this run demands too many resources.
> >> >> >>
> >> >> >> With hybridgb, EPtot = 58053860.3203 and
> >> >> >> Without hybbridgb, EPtot = 546283.2545.
> >> >> >>
> >> >> >> Also, RESTRAINT = 57444151.6840 with hybridgb and
> >> >> >> RESTRAINT = 177.4334 without hybridgb.
> >> >> >> It seems like the hybridgb method overestimates restraint energy
> of
> >> my
> >> >> >> graphite surface too much. I checked the structure of whole
> system.
> >> >> >> The peptide lies on the graphite surface now. I think explicitly
> >> >> >> considered water molecules by numwatkeep=500 are on the other side
> of
> >> >> >> graphite surface, that is across the surface. Do you think this
> makes
> >> >> >> problems?
> >> >> >> Thank you.
> >> >> >> Bongkeun Kim
> >> >> >>
> >> >> >>
> >> >> >>
> >> >> >> _______________________________________________
> >> >> >> AMBER mailing list
> >> >> >> AMBER.ambermd.org
> >> >> >> http://lists.ambermd.org/mailman/listinfo/amber
> >> >> >>
> >> >> >
> >> >> >
> >> >> >
> >> >> > --
> >> >> > Jason M. Swails
> >> >> > Quantum Theory Project,
> >> >> > University of Florida
> >> >> > Ph.D. Graduate Student
> >> >> > 352-392-4032
> >> >> >
> >> >>
> >> >>
> >> >>
> >> >>
> >> >>
> >> >> _______________________________________________
> >> >> AMBER mailing list
> >> >> AMBER.ambermd.org
> >> >> http://lists.ambermd.org/mailman/listinfo/amber
> >> >>
> >> >
> >> >
> >> >
> >> > --
> >> > Jason M. Swails
> >> > Quantum Theory Project,
> >> > University of Florida
> >> > Ph.D. Graduate Student
> >> > 352-392-4032
> >> > _______________________________________________
> >> > 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
> >
>
>
>
>
>
> _______________________________________________
> 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 Sun Dec 19 2010 - 09:30:02 PST
