[gmx-users] Problems with coupling groups in gas phase simulations
mark.abraham at anu.edu.au
Wed Sep 22 02:40:48 CEST 2010
----- Original Message -----
From: Sascha Rehm <Sascha.Rehm at itb.uni-stuttgart.de>
Date: Tuesday, September 21, 2010 22:34
Subject: [gmx-users] Problems with coupling groups in gas phase simulations
To: gmx-users at gromacs.org
> Dear Gromacs-User (and Developers),
> I tried to set up and simulate some systems with 1000 Argon
> atoms and roughly a thousand water molecules that were randomly
> distributed (like gaseous water in argon gas).
> The starting systems are set up with OPLS/AA and 100nm box side
> length, so the systems are much too large and are expected to
> shrink in the first few picoseconds.
Generating velocities at 5K, annealing to 318K in 2ps at NPT at a density that's far too low is being pretty rough on the system. I'm surprised the integration didn't break completely. I've no idea what your objective is, but I wouldn't expect to be able to make equilibrium observations for many nanoseconds :-)
> Two different coupling groups were used (Ar and SOL) with two
> different compressibilities (1.27e-2 for Argon and 5.3e-5 for Water).
T-coupling works best in the limit of large groups of atoms of similar heat capacity. Standard MD advice is not to couple (say) ions in water to separate groups, because the resulting coupling leads to artifacts. This would seem to apply even more strongly to a mixture of components such as you have.
> The systems are heated up in the first ns to 318K and then
> continued for another 100 ns. During that time the boxes shrink
> only to a box size of 99.5 nm, leading to a totally wrong density.
> When I set up the systems with only one couling group ("System")
> the systems shrink to roughly 38nm and show a correct density in
> a few hundred picoseconds but then the density starts to
> oszilate every few ns. I guess this happens because the
> compressibility is set to 1.2e-2 and the water starts to form
> droplets (which should have a compressibility of 5.3e-5)?
> Increasing tau_p from 0.6 to 2.0 does not show big effects on
> the oszilation.
> 1. Is there an explanation why the systems behave strange when
> using 2 coupling groups? My search on the mailing list did not
> lead to an answer (also, there are only a few topics on
> simulating gas or condensing water droplets)
Yes, see above.
> 2. Are there reasons, why I should not use only one coupling
> group for the whole system? E.g. in protein-solvent simulations,
> the manual says that using only one coulpling group leads to an
> increase of temperature in the protein and a decrease in the solvent.
There, the issue is a difference in heat capacity between two groups of notable size. Search literature for "hot solute cold solvent" for discussion.
Since you're simulating a system that's intended to start mixed and then do something, this situation cannot arise until (perhaps) your phases separate.
> Hopefully someone can help me out of that problem...
> Best Regards,
> Sascha Rehm
> The used MDP file looks like the following:
> integrator = md
> dt = 0.002
> nsteps = 5000000
> annealing = single
> annealing_npoints = 3
> annealing_time = 0 100 5000000
> annealing_temp = 5 318.15 318.15
> nstxout = 0
> nstvout = 0
> nstenergy = 1000
> nstxtcout = 1000
> ;group(s) to write to energy file
> energygrps = system
> nstlist = 10
> ns_type = grid
> ;cut-off distance for the short-range neighbor list
> rlist = 1.5
> ;treatment of electrostatic interactions
> coulombtype = Cut-off
> rcoulomb = 1.5
> ;treatment of van der waals interactions
> vdwtype = cut-off
> rvdw = 1.5
> ; Periodic boudary conditions in all the directions
> pbc = xyz
> ;Temperature coupling
> tcoupl = berendsen
> grps = system
> tau_t = 0.1
> ref_t = 318.15
> ;Pressure coupling
> Pcoupl = berendsen
> Pcoupltype = isotropic
> tau_p = 0.6
> compressibility = 1.27e-2
> ref_p = 1.0
> ;Velocity generation
> gen_vel = yes
> gen_temp = 5
> gen_seed = 173529
> Dipl. Bioinf. Sascha Rehm
> Institut fuer Technische Biochemie
> Allmandring 31
> D-70569 Stuttgart
> gmx-users mailing list gmx-users at gromacs.org
> Please search the archive at
> http://www.gromacs.org/Support/Mailing_Lists/Search before posting!
> Please don't post (un)subscribe requests to the list. Use the
> www interface or send it to gmx-users-request at gromacs.org.
> Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the gromacs.org_gmx-users