To equlibrate a structure you want to use a sequence of steps like this:
5000
cycles with only the water molecules and ions permitted to move
5000
cycles with only water molecules, ions, and hydrogen atoms in motion.
5000
cycles with the protein backbone fixed & all other atoms in motion
This was followed by heating the system to 300K, with
all the backbone fixed, over 500,000 steps (0.5nsec). The next stage used 500,000
steps of MD in the NPT ensemble to adjust the density of the solvent box, with
the protein backbone still held fixed. There followed four steps of molecular
dynamics with the protein backbone permitted increasing degrees of harmonic
motion, each iteration also 0.5 nsec long. Restraint forces of 4, 3, 2, and 1 kcal/mole/A**2,
respectively, were used before production MD was undertaken. I don't run sander, I use NAMD with the AMBER force field & AMBER parameters, so the restraint forces that are easiest to set may differ for you. I don't think that matters particularly, just that you use a similar procedure. When I use this sequence, the backbone RMSD usually levels off after 2-5 nsec of production MD. I didn't invent this sequence; I learned it from members of the McCammon group. Irene Newhouse
> Date: Wed, 8 Feb 2012 04:13:55 -0800
> From: khaledbodoor.yahoo.com
> To: amber.ambermd.org
> Subject: [AMBER] what do reprocessing steps do in md simulation?
>
> Hello,
>
> I'm running a simulation to do GB implicit calculation of the free energy of hydration for a ligand. The simulation involves two md runs: one in vacuum and another in water. Then, four different "reprocessing" steps are executed. The free energy of hydration is obtained later using the BAR method. The md script that I use is below:
>
> #!/bin/tcsh
> sander -i water.in -p mol.prmtop -c mol.crd -o water.out -r water.rst -x water.c
> rd -v water.vel -e water.ene -inf water.inf -O
>
> sander -i vacuum.in -p mol.prmtop -c mol.crd -o vacuum.out -r vacuum.rst -x vacu
> um.crd -v vacuum.vel -e vacuum.ene -inf vacuum.inf -O
>
> #Do reprocessing
> sander -i water_rerun.in -p mol.prmtop -c mol.crd -o water_water.out -r water_wa
> ter.rst -y water.crd -v water_water.vel -e water_water.ene -inf water_water.inf
> -O
>
> sander -i vacuum_rerun.in -p mol.prmtop -c mol.crd -o water_vacuum.out -r water_
> vacuum.rst -y water.crd -v water_vacuum.vel -e water_vacuum.ene -inf water_vacuu
> m.inf -O
>
> sander -i water_rerun.in -p mol.prmtop -c mol.crd -o vacuum_water.out -r vacuum_
> water.rst -y vacuum.crd -v vacuum_water.vel -e vacuum_water.ene -inf vacuum_wate
> r.inf -O
>
> sander -i vacuum_rerun.in -p mol.prmtop -c mol.crd -o vacuum_vacuum.out -r vacuu
> m_vacuum.rst -y vacuum.crd -v vacuum_vacuum.vel -e vacuum_vacuum.ene -inf vacuum
> _vacuum.inf -O
>
> I'm n ot sure what the reprocessing steps do. It is clear they take as input the trajectories generated in the first two steps (vacuum.crd and water.crd). As examples, I include the input files vacuum.in and vacuum_rerun.in:
>
> vacuum.in:
>
> trajectory segment
> &cntrl
> imin = 0, nstlim = 5000000, ntwr = 5000,
> nscm=100,
> igb = 0, gbsa = 0,
> cut=24.0,
> tempi = 300, ntt = 3, temp0 = 300,
> ntpr=5000,
> gamma_ln=1,
> vlimit = 0.000,
> ntc=2, ntf = 2, tol = 1.0d-8
> dt = 0.002, nrespa = 2,
> ntb = 0, ntave = 0,
> ioutfm = 0, ntwx = 5000, ntwe = 5000,
> ntr = 0,
> ig=27145
> &end
> END
> END
>
> vacuum_rerun.in:
>
> trajectory segment
> &cntrl
> imin = 5, maxcyc = 1, nstlim = 5000000, ntwr = 5000,
> nscm=100,
> igb = 0, gbsa = 0,
> cut=24.0,
> tempi = 300, ntt = 3, temp0 = 300,
> ntpr=5000,
> gamma_ln=1,
> vlimit = 0.000,
> ntc=2, ntf = 2, tol = 1.0d-8
> dt = 0.002, nrespa = 1,
> ntb = 0, ntave = 0,
> ioutfm = 0, ntwx = 5000, ntwe = 5000,
> ntr = 0,
> ig=27145
> &end
> END
>
> The only difference between vacuum.in and vacuum_rerun.in is that in the latter:
>
> imin=5, maxcyc=1, nrespa=2
>
> while imin=0, nrespa=1 in the former.
>
> Thanks for the help.
>
> Khaled
>
>
>
>
>
>
> _______________________________________________
> 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 Wed Feb 08 2012 - 09:30:06 PST
