AMBER Archive (2009)

Subject: Re: [AMBER] Umbrella test

From: David A. Case (case_at_biomaps.rutgers.edu)
Date: Sat Apr 04 2009 - 08:12:13 CDT


On Thu, Apr 02, 2009, Arnaud wrote:
>
> I runed umbrella sampling with amber8 and amber10 with the parameters file
> given in amber manual :
>
> ************************************
> test of umbrella sampling of a chi torsion angle
> &cntrl
> nstlim=50000, cut=20.0, igb=1, saltcon=0.1,
> ntpr=1, ntwr=100000, ntt=3, gamma_ln=0.2,
> ntx=5, irest=1,
> ntc=2, ntf=2, tol=0.000001,
> dt=0.001, ntb=0,
> nmropt=1,
> /
> &wt type='DUMPFREQ', istep1=10 /
> &wt type='END' /
> DISANG=chi.RST
> DUMPAVE=chi_vs_t.170
> EOF
> **************************************
>
> chi.RST, prmtop and md4.x files needed for this calculation were those
> proposed in test directory of amber package.
>
> Begin and end results in chi_vs_t.170 file obtained with amber10 were :
> 0 28.408
> 10 76.692
> 20 103.804
> 30 106.999
> 40 101.039
> 50 93.077
> 60 109.138
> 70 119.145
> 80 142.406
> 90 133.365
> 100 128.226
> 110 117.133
> ...
> 49900 169.076
> 49910 171.938
> 49920 170.164
> 49930 170.135
> 49940 170.565
> 49950 171.870
> 49960 173.956
> 49970 172.371
> 49980 172.124
> 49990 170.298
>
>
> Begin and end results in chi_vs_t.170 obtained with amber8 were :
> 0 28.408
> 10 76.756
> 20 103.865
> 30 107.183
> 40 101.340
> 50 94.314
> 60 109.482
> 70 119.648
> 80 141.769
> 90 132.950
> 100 128.029
> 110 118.266
> ...
> 49900 166.738
> 49910 162.640
> 49920 166.876
> 49930 172.380
> 49940 173.665
> 49950 167.582
> 49960 163.551
> 49970 166.160
> 49980 166.259
> 49990 164.893
>
> *******************************
>
> Why these differences of results ?

Molecular dynamics is numerically chaotic, see the discussion on p. 18
of the Amber10 manual. You should see if the distributions of the chi
angle are different in the two runs. It is also possible that the
random number generator in the Langevin thermostat is slightly different
in Amber 8 vs 9 and 10: try a short run with ntt=0 to see how and when
initially identical trajectories start to diverge.

...hope this helps....dac

_______________________________________________
AMBER mailing list
AMBER_at_ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber