AMBER Archive (2007)

Subject: AMBER: [rsachdeva@imtech.res.in: Pls help - REMD with torsional restraints]

From: David A. Case (case_at_scripps.edu)
Date: Thu Feb 22 2007 - 20:03:47 CST

----- Forwarded message from Ruchi Sachdeva <rsachdeva_at_imtech.res.in> -----

     Im trying to run REMD with torsional restraints on alpha helix
residues using AMBER8 (ff99) on Linux cluster. Although exchanges were
being allowed between replicas but target temperatures were not
getting swapped between the replicas, as indicated in the rem.log
file. Same torsional restraints were applied to all the replicas.

    But when torsional restraints were not applied, REMD was running
fine & target temperatures got exchanged.

    I did this as follows:

Step1 - First I did initial minimization of 25000 steps in implicit
solvent, igb=1, without any torsional restraints.

Step2 - The system was heated up to 300K for 100 ps using berendsen
thermostat. Torsional restraints were also applied.

Step3 - Equilibration was carried out at 300K for 1000ps. Again
torsional restraints were applied.

Step4 - 4 copies of the resulting coordinates ( i.e. md2.rst) were
made. Each replica was independently equilibrated at its own
temperature i.e. 275, 282, 289, 296K for 1000ps, alongwith torsional
restraints.

   I am showing the input file for replica-0 only. Input files for
other replicas differed only in the target temperature.

equil0.in
Equilibration of replica-0 at 275K for 1000 ps, igb = 1, with
torsional restraints on residues forming alpha helix & beta sheet
 &cntrl
  imin = 0, ntb = 0,
  irest = 1, ntx = 5,
  igb = 1, ntpr = 1000, ntwx = 1000, ntwr = 1000,
  ntt = 1, tautp = 1.0,
  tempi = 275.0, temp0 = 275.0,
  ntc = 2, ntf = 2,
  nstlim = 500000, dt = 0.002,
  cut = 999, vlimit = 25, nmropt = 1
 /

    &wt type='REST', istep1=0, istep2=10000, value1=0.1,
                   value2=1.0, /
    &wt type='REST', istep1=10001, istep2=500000, value1=1.0,
                   value2=1.0, /

    &wt type='END' /
   LISTIN=POUT
   DISANG=RST.ci

Step 5 - The final configurations (i.e. equil.rst.000, equil.rst.001,
equil.rst.002, equil.rst.003) of these equilibrations were taken as
input coordinates for REMD. Same restraints i.e. same RST file was
used for each replica.

    4 input files were made. Only the target temperature was
different in these files, rest of the variables were same. I am
showing just 2 files :

rem.in.000
REMD run with torsional restraints on alpha helix residues, igb=1, AMBER8
 &cntrl
        imin = 0, nstlim = 2500, dt = 0.002,
        ntx = 5, tempi = 0.0, temp0 = 275.0,
        ntt = 1, tol = 0.000001, saltcon = 0.2,
        ntc = 2, ntf = 1, ntb = 0, nsnb = 20,
        ntwx = 100, ntwe = 0, ntwr = 200, ntpr = 100,
        scee = 1.2, cut = 99.0,
        ntr = 0, tautp = 0.1, offset = 0.09,
        nscm = 500, igb = 1, numexchg = 10,
        irest = 1, ntave = 0, vlimit = 25, nmropt = 1
  /

    &wt type='REST', istep1=0, istep2=100, value1=0.1,
                   value2=1.0, /
    &wt type='REST', istep1=101, istep2=2500, value1=1.0,
                   value2=1.0, /

   &wt type='END' /
   LISTIN=POUT
   DISANG=RST

rem.in.001
REMD run with torsional restraints on alpha helix residues, igb=1, AMBER8
 &cntrl
        imin = 0, nstlim = 2500, dt = 0.002,
        ntx = 5, tempi = 0.0, temp0 = 282.0,
        ntt = 1, tol = 0.000001, saltcon = 0.2,
        ntc = 2, ntf = 1, ntb = 0, nsnb = 20,
        ntwx = 100, ntwe = 0, ntwr = 200, ntpr = 100,
        scee = 1.2, cut = 99.0,
        ntr = 0, tautp = 0.1, offset = 0.09,
        nscm = 500, igb = 1, numexchg = 10,
        irest = 1, ntave = 0, vlimit = 25, nmropt = 1
  /

    &wt type='REST', istep1=0, istep2=100, value1=0.1,
                   value2=1.0, /
    &wt type='REST', istep1=101, istep2=2500, value1=1.0,
                   value2=1.0, /

   &wt type='END' /
   LISTIN=POUT
   DISANG=RST

        I got following rem.log file:

 # replica exchange log file
 # Replica #, Velocity Scaling Factor, T, Eptot, Temp0, NewTemp0
 # exchange 1
 1 -1.00 279.93 -2015.51 275.00 275.00
 2 -1.00 278.93 -1951.14 282.00 282.00
 3 -1.00 288.59 -1960.63 289.00 289.00
 4 -1.00 285.23 -1913.25 296.00 296.00
 # exchange 2
 1 -1.00 272.73 -1993.13 275.00 275.00
 2 1.01 272.77 -1924.55 ** 282.00 ** 289.00
 3 0.99 289.04 -1951.70 ** 289.00 ** 282.00
 4 -1.00 299.82 -1950.95 296.00 296.00
 # exchange 3
 1 -1.00 270.77 -2011.05 275.00 275.00
 2 -1.00 284.21 -1970.67 ** 282.00 ** 282.00
 3 -1.00 292.72 -1977.76 ** 289.00 ** 289.00
 4 -1.00 301.78 -1946.56 296.00 296.00
 # exchange 4
 1 -1.00 278.82 -2038.74 275.00 275.00
 2 -1.00 286.11 -1989.72 282.00 282.00
 3 -1.00 290.02 -1955.69 289.00 289.00
 4 -1.00 302.56 -1949.02 296.00 296.00
 # exchange 5
 1 -1.00 277.76 -2038.56 275.00 275.00
 2 -1.00 283.52 -1967.68 282.00 282.00
 3 -1.00 293.98 -1965.70 289.00 289.00
 4 -1.00 288.43 -1914.41 296.00 296.00
 # exchange 6
 1 -1.00 274.67 -2043.14 275.00 275.00
 2 1.01 281.29 -1974.33 282.00 289.00
 3 0.99 294.38 -1971.34 289.00 282.00
 4 -1.00 295.38 -1937.41 296.00 296.00
 # exchange 7
 1 -1.00 272.07 -2015.42 275.00 275.00
 2 -1.00 283.13 -1963.17 282.00 282.00
 3 -1.00 299.59 -1991.40 289.00 289.00
 4 -1.00 291.96 -1923.93 296.00 296.00
 # exchange 8
 1 -1.00 269.27 -2014.54 275.00 275.00
 2 1.01 289.24 -1980.72 282.00 289.00
 3 0.99 284.76 -1954.30 289.00 282.00
 4 -1.00 295.82 -1930.07 296.00 296.00
 # exchange 9
 1 -1.00 273.25 -2028.57 275.00 275.00
 2 -1.00 282.23 -1977.14 282.00 282.00
 3 1.01 278.57 -1912.78 289.00 296.00
 4 0.99 304.64 -1956.62 296.00 289.00
 # exchange 10
 1 -1.00 277.65 -2040.60 275.00 275.00
 2 1.01 285.22 -1982.82 282.00 289.00
 3 0.99 291.23 -1976.79 289.00 282.00
 4 -1.00 292.41 -1918.82 296.00 296.00

             As we can see from the rem.log file that during exchange
attempt 2, exchange is allowed between replica pair 2 & 3 (marked as
star). But after exchange attempt 3, the current target temperature of
replica 2 is still 282K and of replica 3 is still 289K (marked as
star). Rather they must be 289 & 282K respectively as the exchange has
been allowed prior to Sander call. We can notice such kind of
behaviour in further exchange attempts and with other replicas.

            This indicates that the target tempratures are not
getting swapped even when the exchange is allowed.

******* However when I didnt apply torsional restraints, REMD was
running fine. Temperatures swapping did occur. *********

Have I followed the right protocol? Please check out the files. Can u
give some suggestions.

Thanks in advance!!!!

Ruchi

----- End forwarded message -----
-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber_at_scripps.edu
To unsubscribe, send "unsubscribe amber" to majordomo_at_scripps.edu