AMBER Archive (2005)
Subject: AMBER: langevin, GB and simulated annealing - 3

From: pascal.baillod_at_epfl.ch
Date: Tue Apr 26 2005 - 05:37:16 CDT

Next message: Anthony Cruz: "AMBER: AMBER Test"
Previous message: Thomas Steinbrecher: "Re: AMBER: A problem in TI approach - test case: meth2eth and eth2meth"
Next in thread: David Case: "Re: AMBER: langevin, GB and simulated annealing - 3"
Reply: David Case: "Re: AMBER: langevin, GB and simulated annealing - 3"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

Dear Prof. Case,

I thank you very much for your answer.

I actually use Langevin dynamics on my rna system because the sampling stays
pretty much closer to the nmr structure as compared to Berendsen dynamics,
result that by the way matches to simulations reported in the litterature. But
the SA simply does not work, it jumps very quickly from 0 to 300 degrees K. It
works fine for the Berendsen dynamics, though. I would like to know whether I
forgot anything in the input file (end of the email), or whether SA simply is
not made to work with langevin dynamics in amber8 (as suggested by one of your
messages concerning this issue, in the mailing list archives:
http://structbio.vanderbilt.edu/archives/amber-archive/2004/3024.phtml ).

I would guess the thermostat temperature control for langevin dynamics is done
via the random collision force, that is taken from a temperature dependant
gaussian distribution, as suggested in van Gunstern et al., 1981. If that is the
implementation in amber8, I suppose simulated annealing could be performed by
slowly varying the temperature of the gaussian distribution yielding the random
collision force? But that might not be implemented..

One last question, if you will: Supposing there's no SA for langevin dynamics, I
should apply the trick you suggested me in your last email, that I recopy here
below:
To use Langevin dynamics, you could set a series of target temperatures,
(say 600, 500, 400, 300, 200, 100, 10, etc.) to take effect at different times
during the simulation. This could have the effect of fairly slow cooling.
But the Berendsen approach seems to me to be a more natural way of carrying
out simulated annealing.

Practically speaking, you mean there would be one simulation for each of these
temperatures, so that in the end I would have to paste the trajectory and energy
files? Or are there any tools to do that in one single simulation in amber?

Thank you very much!

Pascal

------------------------------------------------------

equilibr
&cntrl
     IMIN = 0,
     NMROPT = 1,
     NTX = 1,
     IREST = 0,
     NTPR = 10,
     NTWR = 10000,
     NTWX = 500,
     NTF = 1,
     NTB = 0,
     DIELC = 1.0,
     CUT = 1000.0,
     SCNB = 2.0,
     SCEE = 1.2,
     NSTLIM = 32000000,
     NSCM = 1000,
     DT = 0.0015,
     NTT = 3,
     GAMMA_LN = 91,
     TEMPI=0.0
     NTC = 2,
     TOL = 0.00001,
     NTR = 0,
     IBELLY = 0,
     IVCAP = 0,
     IGB = 1,
&end

&wt
     TYPE='TEMP0', istep1=0, istep2=400000, value1=0.0, value2=300.0,
&end
&wt
     TYPE='TEMP0', istep1=400001, istep2=32000000, value1=300.0, value2=300.0,
&end
&wt
     TYPE='END',
&end

-----------------------------------------------------------------------

          -------------------------------------------------------
          Amber 8 SANDER Scripps/UCSF 2004
          -------------------------------------------------------

| Run on 04/19/2005 at 22:25:28

Here is the input file:

equilibr
&cntrl
     IMIN = 0,
     NMROPT = 1,
     NTX = 1,
     IREST = 0,
     NTPR = 10,
     NTWR = 10000,
     NTWX = 500,
     NTF = 1,
     NTB = 0,
     DIELC = 1.0,
     CUT = 1000.0,
     SCNB = 2.0,
     SCEE = 1.2,
     NSTLIM = 32000000,
     NSCM = 1000,
     DT = 0.0015,
     NTT = 3,
     GAMMA_LN = 91,
     TEMPI=0.0
     NTC = 2,
     TOL = 0.00001,
     NTR = 0,
     IBELLY = 0,
     IVCAP = 0,
     IGB = 1,
&end


&wt
     TYPE='TEMP0', istep1=0, istep2=400000, value1=0.0, value2=300.0,
&end
&wt
     TYPE='TEMP0', istep1=400001, istep2=1000000, value1=300.0, value2=300.0,
&end
&wt
     TYPE='END',
&end

--------------------------------------------------------------------------------
1. RESOURCE USE:
--------------------------------------------------------------------------------

| Flags: MPI
| New format PARM file being parsed.
| Version = 1.000 Date = 03/18/05 Time = 18:11:40
NATOM = 389 NTYPES = 14 NBONH = 133 MBONA = 286
NTHETH = 301 MTHETA = 445 NPHIH = 658 MPHIA = 812
NHPARM = 0 NPARM = 0 NNB = 2166 NRES = 12
NBONA = 286 NTHETA = 445 NPHIA = 812 NUMBND = 35
NUMANG = 72 NPTRA = 39 NATYP = 24 NPHB = 0
IFBOX = 0 NMXRS = 34 IFCAP = 0 NEXTRA = 0
NCOPY = 0

--------------------------------------------------------------------------------
2. CONTROL DATA FOR THE RUN
--------------------------------------------------------------------------------

General flags:
imin = 0, nmropt = 1

Nature and format of input:
ntx = 1, irest = 0, ntrx = 1

Nature and format of output:
     ntxo = 1, ntpr = 10, ntrx = 1, ntwr = 10000
     iwrap = 0, ntwx = 500, ntwv = 0, ntwe = 0
     ioutfm = 0, ntwprt = 0, idecomp = 0, rbornstat= 0

Potential function:
     ntf = 1, ntb = 0, igb = 1, nsnb = 25
     ipol = 0, gbsa = 0, iesp = 0
     dielc = 1.00000, cut =1000.00000, intdiel = 1.00000
     saltcon = 0.00000, offset = 0.09000, gbalpha= 1.00000
     gbbeta = 0.00000, gbgamma = 0.00000, surften = 0.00500
     rdt = 0.00000, rgbmax = 25.00000
     scnb = 2.00000, scee = 1.20000

Frozen or restrained atoms:
ibelly = 0, ntr = 0

Molecular dynamics:
nstlim =32000000, nscm = 1000, nrespa = 1
t = 0.00000, dt = 0.00150, vlimit = 20.00000

Langevin dynamics temperature regulation:
ig = 71277
temp0 = 300.00000, tempi = 0.00000, gamma_ln= 91.00000

SHAKE:
ntc = 2, jfastw = 0
tol = 0.00001

NMR refinement options:
iscale = 0, noeskp = 1, ipnlty = 1, mxsub = 1
scalm = 100.00000, pencut = 0.10000, tausw = 0.10000

--------------------------------------------------------------------------------
3. ATOMIC COORDINATES AND VELOCITIES
--------------------------------------------------------------------------------

begin time read from input coords = 0.000 ps

Begin reading energy term weight changes/NMR restraints
WEIGHT CHANGES:
TEMP0 0 400000 0.000000 300.000000 0 0
TEMP0 4000011000000 300.000000 300.000000 0 0

RESTRAINTS:
** No restraint defined **

Done reading weight changes/NMR restraints

Number of triangulated 3-point waters found: 0
| Atom division among processors:
| 0 196 389
| Running AMBER/MPI version on 2 nodes

--------------------------------------------------------------------------------
4. RESULTS
--------------------------------------------------------------------------------

NSTEP = 0 TIME(PS) = 0.000 TEMP(K) = 0.00 PRESS = 0.0
Etot = -2779.4852 EKtot = 0.0000 EPtot = -2779.4852
BOND = 19.1296 ANGLE = 82.2276 DIHED = 272.9131
1-4 NB = 95.4341 1-4 EEL = -1594.9962 VDWAALS = -231.1730
EELEC = 982.8168 EGB = -2405.8371 RESTRAINT = 0.0000
------------------------------------------------------------------------------