I agree with Justin, 1 ns is just nowhere near long enough. Something in
the tens of ns should do it.
I have used pretty much the same settings as you (using the CHARMM
water, as you have) and get APL's matching those published in the Klauda
paper. You are correct that you should not use the dispersion
correction, this was discussed in their paper.
Cheers
Tom
On 18/05/11 23:20, Justin A. Lemkul wrote:
Peter C. Lai wrote:
Hello
I am trying to equilibrate from scratch a 196 POPC bilayer using Tom's
charmm36.ff. My box has a lot of TIPS3P (charmm) waters above and below
the membrane with a box size around 8.5x8.5x12.7. My desired end state
is to only use the setup for g_membed but I'd also like to get a pristine
tall bilayer that might be useful for other things or colleagues. I ran
1ns NPT so far with:
dt = 0.002
continuation = yes
constraint_algorithm = lincs
constraints = all-bonds
lincs_iter = 1
lincs_order = 4
ns_type = grid
nstlist = 5
rlist = 1.2
rlistlong = 1.4
rcoulomb = 1.2
rvdw = 1.2
vdwtype = switch
rvdw_switch = 0.8
coulombtype = PME
pme_order = 4
fourierspacing = 0.16
tcoupl = Nose-Hoover
tc-grps = POPC SOL
tau_t = 0.5 0.5
ref_t = 300 300
pcoupl = Parrinello-Rahman
pcoupltype = semiisotropic
tau_p = 4
ref_p = 1.01325 1.01325
compressibility = 4.5e-5 4.5e-5
DispCorr = no
comm_mode = Linear
comm_grps = POPC SOL
The metrics look roughly stable:
Energy Average Err.Est. RMSD Tot-Drift
---------------------------------------------------------------------
Temperature 300 0.00018 0.957697 0.000466096 (K)
Density 1017.26 0.24 1.5953 1.4803 (kg/m^3)
Pressure 1.00795 0.065 99.5783 -0.0431925 (bar)
Box-X 8.58491 0.036 0.0773839 -0.235388 (nm)
Box-Y 8.59611 0.036 0.0774849 -0.235696 (nm)
Box-Z 12.4357 0.1 0.216337 0.659896 (nm)
My APLs are ~10A^2/lipid above what they "should be" according to Klauda
and experimental (I get 76-77A^2/lipid vs 65-58A^2/lipid). I suppose
with TIPS3P water, I could get closer LJ packing if I turned on DispCorr
but I thought that you generally left that off in charmm36 bilayer runs...
Now, I also could extend for several ns until density/box drift gets
smaller, but any other thoughts (yes this is probably a continuation of the
CHARMM36 lipid bilayers thread from October
(http://www.mail-archive.com/gmx-users@gromacs.org/msg34582.html)?
Later in this thread it is suggested that one use TIPS3P ("CHARMM TIP3P") water,
so start there, although that discussion found that the APL was consistently
underestimated, not overestimated, as is your case.
1 ns is not nearly enough to make solid conclusions about membranes. Lipid
rotational relaxation is on the order of 5 ns, and translational relaxation
about 10 ns. I'd say you need at least 20 ns of simulation to make any real
conclusions, with some of that of course discarded as equilibration.
-Justin
--
Dr Thomas Piggot
University of Southampton, UK.
--
gmx-users mailing list gmx-users@gromacs.org
http://lists.gromacs.org/mailman/listinfo/gmx-users
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-requ...@gromacs.org.
Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
