AMBER Archive (2004)

Subject: Re: AMBER: question on simulated annealing

From: Carlos Simmerling (carlos_at_ilion.bio.sunysb.edu)
Date: Tue Jul 27 2004 - 13:28:27 CDT

you don't set temp0 in the sander input. in order for the
random numer seed to be used (and the differences in IG to matter),
you need to assign initial velocities. Check your output file and see if
TEMPI is equal to zero, which means that no initial velocities are
assigned and they are just obtained from the forces (which are the same
for all of your simulations). if so, set tempi to any nonzero value that
you want (such as 300K, but even 1K should work to provide divergence).
carlos

===================================================================
Carlos L. Simmerling, Ph.D.
Assistant Professor Phone: (631) 632-1336
Center for Structural Biology Fax: (631) 632-1555
Stony Brook University Web: http://comp.chem.sunysb.edu/carlos
Stony Brook, NY 11794-5115 E-mail: carlos.simmerling_at_stonybrook.edu
===================================================================

Venkata S Koppuravuri wrote:

>Hi All,
>
> I have a question regarding simulated annealing.
> Starting with a PDB file I have protonated and minimized my protein.
>Following the example in the manual I have heated and cooled the protein
>multiple times starting with the minimized protein (replicates ..... not
>sequentially) using the minimized protein restart file. To my surprise
>the structures in each replicate are identical. Since I wondered if
>randomization of momentum occurred prior to annealing, I retried the
>process doing replicate minimizations but got the same structures again. I
>tried using the IG flag to change the random number seed between
>annealing, but still go the same structure on all annealing runs.
>
> I was expecting that there would be some variation in the detailed
>structure, corresponding to the structure finding alternative.... not
>necessarily global .... minima. Is this not so? Or what am I missing in
>doing my annealing?
>
> The scripts that I am using are below:
>
>----------------------------
>#minimization input file
> &cntrl
> imin = 1,
> maxcyc = 5000, ncyc = 2000, ntmin = 2,
> scee = 1.2,
> nsnb = 25, cut = 8.0,
> ntpr = 100, ntwx = 100, ntwv = 100, ntwe = 100,
> ntb = 0, igb = 1,
> lastist = 7500000, lastrst = 7500000,
> &end
>
>END
>
>----------------------------
>
>----------------------------
>#simulated annealing input file
>#
>#
> &cntrl
> nstlim=22000, pencut=-0.001, nmropt=1,
> ntpr=100, ntt=1, ntwx=200,
> cut=15.0, ntb=0, vlimit=20,
> ntr=1,
> ig=61777,
> lastist = 7500000, lastrst = 7500000,
> /
> &ewald
> eedmeth=5,
> /
> &end
>#
> &wt type='TEMP0', istep1=0,istep2=500,value1=0.,
> value2=100., /
> &wt type='TEMP0', istep1=500,istep2=1000,value1=100.,
> value2=100., /
> &wt type='TEMP0', istep1=1001,istep2=1500,value1=100.,
> value2=200., /
> &wt type='TEMP0', istep1=1501,istep2=2000,value1=200.,
> value2=200., /
> &wt type='TEMP0', istep1=2001,istep2=2500,value1=200.,
> value2=300., /
> &wt type='TEMP0', istep1=2501,istep2=3000,value1=300.,
> value2=300., /
> &wt type='TEMP0', istep1=3001,istep2=3750,value1=300.,
> value2=400., /
> &wt type='TEMP0', istep1=3751,istep2=4500,value1=400.,
> value2=400., /
> &wt type='TEMP0', istep1=4501,istep2=5250,value1=400.,
> value2=500., /
> &wt type='TEMP0', istep1=5251,istep2=6000,value1=500.,
> value2=500., /
> &wt type='TEMP0', istep1=6001,istep2=6500,value1=500.,
> value2=550., /
> &wt type='TEMP0', istep1=6501,istep2=9000,value1=550.,
> value2=550., / #
> &wt type='TEMP0', istep1=9001, istep2=11000, value1=550.0,
> value2=450.0, /
> &wt type='TEMP0', istep1=11001, istep2=13000, value1=450.0,
> value2=350.0, /
> &wt type='TEMP0', istep1=13001, istep2=15000, value1=350.0,
> value2=300.0, /
> &wt type='TEMP0', istep1=15001, istep2=22000, value1=300.0,
> value2=300.0, /
> &wt type='TAUTP', istep1=0,istep2=4500,value1=1.0,
> value2=1.0, /
> &wt type='TAUTP', istep1=4501,istep2=5000,value1=2.0,
> value2=2.0, /
> &wt type='TAUTP', istep1=5001,istep2=9000,value1=2.5,
> value2=3.0, /
> &wt type='TAUTP', istep1=9001,istep2=15000,value1=3.0,
> value2=3.0, /
> &wt type='TAUTP', istep1=15001,istep2=21000,value1=1.0,
> value2=1.0, /
> &wt type='TAUTP', istep1=21001,istep2=22000,value1=0.1,
> value2=0.05, /
> &wt type='END' /
>DISANG=dummy
>restraint residues1-124
> 1.0
>RES 1 124
>RES 163 292
>RES 323 505
>RES 575 811
>RES 832 1007
>END
>END
>END
>END
>------------------------
>
>Thanks in advance
>
>-Venkat
>
>-----------------------------------------------------------------------
>The AMBER Mail Reflector
>To post, send mail to amber_at_scripps.edu
>To unsubscribe, send "unsubscribe amber" to majordomo_at_scripps.edu
>
>
-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber_at_scripps.edu
To unsubscribe, send "unsubscribe amber" to majordomo_at_scripps.edu