[gmx-users] freeze dummy atom completely

[Pim Schravendijk]

Hello,

I am doing a simulation where a molecule-dummy atom z-distance is kept
constant and I want to look at the potential of mean force depending on
this distance. The dummy atom should be fixed at an absolute z-point in my
box, x and y are not important.

In pull.ppa there is (amongst other options):

runtype                  = constraint
pulldim                  = N N Y
group_1                  = MOLECULE
reference_group          = DUMM
reftype                  = com
constraint_direction     = 0.0 0.0 0.1
constraint_tolerance     = 1e-06

in topol.top I have:

[ moleculetype ]
dumm           3

[ atoms ]
     1        DUM      1    DUMM   DUM      1    0.00000   99.00000

and in grompp.mdp

tc_grps                  = mol sol dumm
tau_t                    = 0.1 0.1 0.1
ref_t                    = 298 298 0

freezegrps               = dumm 
freezedim                = N N Y
cos-acceleration         = 0


My problem is that, if I look at snapshots, my dummy atom doesn't remain
at its z-position, it gets pushed away.  I guess the problem
might be the mass of the dummy, I tried making it 0 but then grompp warns:

WARNING 1 [file "topol.top", line 2387]:
  atom DUM (Res DUMM-1502) has mass 0

and mdrun crashes on:

Fatal error: ci = -2147483648 should be in 0 .. 431 [FILE nsgrid.c, LINE 
218]

mass -1 also doesn't help. Is there a definite solution to keep it at its 
position?

Greetings, Pim

--
Pim Schravendijk - PhD Student
Max Planck Institute for Polymer Research
http://www.mpip-mainz.mpg.de/~schraven/