[gmx-users] T-coupling of water and ion

[zouj01 at mails.tsinghua.edu.cn]

> Jian Zou wrote:
>> I am interested in how to set the T-coupling group.
>> 
>> for system containing Protein(or nucleic acid), water, and ions(eg. Na), many
>> people do T-coupling seperately with water and ion, ie. Protein SOL ION.
>> 
>> While others put water and ion in one tc_group, ie. Protein SOL_ION.
>> 
>> What's the difference between these two tc_grps selections?
>> Do they influence much the simulation results?
>
David wrote:
> Very slightly because when combining the groups you rely on correct
> exchange of heat between solvent and ions, which is not given. You
> shouldn't combine non-polar and polar groups in one (or protein and
> water). You can test it with a short simulation and look at the average
> temperature and the fluctuations of each group (if you save velocities
> you can still compute the T using g_traj).

Hi, David

Thanks for your commemnt.

The system is pulling 8 ADE in a water box, Na added.

I ran 2 tests for 200ps 400K 3bar PME.
then I use
> g_traj -f md.trr -s b4md.tpr -n index -ot temp.xvg
to output temperature information for several groups defined in my index.
while the results seem a little strange to me.


Test 1

tc_group     =  RNA SOL_Na
constraints  =  none
dt           =  1fs

====md.mdp====
include                  =  -I../top
integrator               =  md
tinit                    =  0.0
dt                       =  0.001
nsteps                   =  200000
comm_mode                =  None
nstxout                  =  10000
nstvout                  =  10000
nstfout                  =  10000
nstlog                   =  10000
nstenergy                =  1000
nstxtcout                =  1000
xtc_grps                 =  RNA
energygrps               =  RNA SOL_Na
nstlist                  =  5
ns_type                  =  grid
rlist                    =  0.9
coulombtype              =  PME
rcoulomb                 =  0.9
rvdw                     =  1.4
fourierspacing           =  0.1
pme_order                =  4
optimize_fft             =  yes
constraints              =  none
tcoupl                   =  berendsen
tc-grps                  =  RNA SOL_Na
tau_t                    =  0.1 0.1
ref_t                    =  400 400
pcoupl                   =  berendsen
pcoupltype               =  isotropic
tau_p                    =  2.0
compressibility          =  5.0e-5
ref_p                    =  3.0
gen_vel                  =  no
====EOF====

====temp.xvg====
####
#       System  RNA      SOL     Na      SOL_Na
####
 0      265     402      263     265     263
 10     271     401      268     334     268
 20     269     396      266     306     266
 30     269     405      266     345     266
 40     270     398      267     346     267
 50     268     373      266     299     266
 60     270     397      267     371     268
 70     267     377      265     188     265
 80     271     437      268     469     268
 90     267     408      265     463     265
 100    271     407      268     512     268
 110    268     403      266     516     266
 120    269     412      266     559     266
 130    270     412      267     394     267
 140    272     367      269     522     270
 150    268     387      265     403     265
 160    269     420      266     348     266
 170    269     382      267     434     267
 180    269     401      266     472     266
 190    270     394      267     366     267
 200    269     405      266     504     266
====EOF====


Test 2

tc_group     =  RNA SOL Na
constraints  =  all-bonds
dt           =  2fs

====md.mdp====
include                  =  -I../top
integrator               =  md
tinit                    =  0.0
dt                       =  0.002
nsteps                   =  100000
comm_mode                =  None
nstxout                  =  10000
nstvout                  =  10000
nstfout                  =  10000
nstlog                   =  10000
nstenergy                =  1000
nstxtcout                =  1000
xtc_grps                 =  RNA
energygrps               =  RNA SOL Na
nstlist                  =  5
ns_type                  =  grid
rlist                    =  0.9
coulombtype              =  PME
rcoulomb                 =  0.9
rvdw                     =  1.4
fourierspacing           =  0.1
pme_order                =  4
optimize_fft             =  yes
constraints              =  all-bonds
tcoupl                   =  berendsen
tc-grps                  =  RNA SOL Na
tau_t                    =  0.1 0.1 0.1
ref_t                    =  400 400 400
pcoupl                   =  berendsen
pcoupltype               =  isotropic
tau_p                    =  2.0
compressibility          =  5.0e-5
ref_p                    =  3.0
gen_vel                  =  no
====EOF====

====temp.xvg====
####
#       System  RNA     SOL     Na      SOL_Na
####
 0      263     247     263     265     263
 20     266     257     267     525     267
 40     266     244     267     319     267
 60     269     265     269     544     270
 80     270     263     270     437     270
 100    268     244     268     439     268
 120    269     268     269     293     269
 140    269     230     270     266     270
 160    269     259     270     376     270
 180    267     257     267     330     267
 200    268     248     268     350     268
====EOF====


from these 2 tests, it seems that
1. putting water and Na in one tc_group will give more accurate temperature for
RNA.
2. water or Na could be well coupled to ref_t, for either of these 2 tc_group
selections.


I use different time step and bond option for these 2 tests.
And I use afm pulling options.
I'm not sure whether these mdrun options will also influence the T-coupling.


could someone give some advice or comment on this?


Thank you very much in advance.


Regards,

Jian Zou