[gmx-users] have trouble to repeat PMF with martini polarizable water model

[Yuan Hu]

Hi All,

This is the first time we use Martini force filed and CG polarizable water
model. So we decide to repeat the result in the recent published paper.
Title: Polarizable Water Model for the Coarse-Grained MARTINI Force Field
link:
http://www.ploscompbiol.org/article/info:doi/10.1371/journal.pcbi.1000810
We use Gromacs version 4.6-beta3, single precision.
we want to repeat figure 9 in this paper, the polarizable martini sodium
result.

We build a system containing 128 DPPC, 1 NA+, 1CL-, 2000 CG polarizable
water, and we equilibrate the system for about 500ns. The area per lipid we
got is around 0.635 nm2. it is the same value reported in the paper. The
number density of different cg groups of DPPC and water also match the
density reported in the paper.
Then we set up 40 windows to compute the pmf of translocation of NA+
through a DPPC bilayer, by using umbrella sampling method.
To set up each window, we place sodium ion at the ideal place and turn off
the interaction between sodium and other beads in the system
, then we slowly grow in the vdw interactions between sodium and the the
rest of the system in 10 ns, after that we grow in the vdwq interaction in
the same way.



The results is shown in this figure.
http://udel.edu/~yuanhu/pmf-plos-repeat.png

see the mdp file, and log file of the center of bilayer window:
http://udel.edu/~yuanhu/tau-us.mdp
http://udel.edu/~yuanhu/us.mdp(grompp output mdp file)
http://udel.edu/~yuanhu/us.log


We add the umbrella sampling code to the end the example mdp file from this
website (
http://md.chem.rug.nl/cgmartini/images/parameters/exampleMDP/old/martini_v2.P_example.mdp),
it's also the same mdp file shown in the supporting information of this
paper.
===============================
; Pull code
pull            = umbrella
pull_geometry = distance
pull_dim        = N N Y
pull_start      = no
pull_ngroups    = 1
pull_group0     = DPPC
pull_group1     = NA+
pull_init1      = 0.9
pull_rate1      = 0.0
pull_k1         = 400

ld_seed=-1
===============================

In order to match the PMF, we tried several test, but all of the tests give
the similar results.



The systems we have tested are as follow:

a. T=320k, tau_p=1.0,  tau-t=1.0,  nstlist=10,
rlist=1.2,compressibility=3e-4, tc-grps=DPPC PW_ION
h. T=323k, tau_p=5.0,  tau-t=1.0,  nstlist=10,
rlist=1.2,compressibility=3e-4, tc-grps=DPPC PW_ION
i. T=323k, tau_p=5.0,  tau-t=1.0,  nstlist=5,
rlist=2,compressibility=3e-4, tc-grps=DPPC PW_ION
j. T=323k, tau_p=3.0,  tau-t=0.3,  nstlist=5, rlist=2,compressibility=3e-4,
tc-grps=DPPC PW_ION
k. T=323k, tau_p=3.0,  tau-t=0.3,  nstlist=5,
rlist=2,compressibility=4.5e-5, tc-grps=DPPC PW_ION
l. T=323k, tau_p=3.0,  tau-t=0.3,  nstlist=5,
rlist=2,compressibility=4.5e-5, pull_vec1= 0 0 1
o. T=323k, tau_p=3.0,  tau-t=0.3,  nstlist=5,
rlist=2,compressibility=4.5e-5,comm-grps= DPPC PW_ION
p. T=323k, tau_p=3.0,  tau-t=0.3,  nstlist=5,
rlist=2,compressibility=4.5e-5,tc-grps=DPPC_ION PW
q. T=323k, tau_p=3.0,  tau-t=0.3,  nstlist=5,
rlist=2,compressibility=4.5e-5,comm-grps= DPPC PW_ION, tc-grps=DPPC PW ION
s. T=323k, tau_p=3.0,  tau-t=0.3,  compressibility=4.5e-5, comm-grps= DPPC
PW ION, tc-grps=DPPC PW ION
r. T=325k, tau_p=3.0,  tau-t=0.3,  nstlist=5,
rlist=2,compressibility=4.5e-5,comm-grps= DPPC PW_ION


we also tried to use gromacs 3.3.1 version (single precision), but the pmf
didn't match either.
We don't know where is the problem. We run out of ideas now. Could someone
help us? Thanks.