Drift in coordinates of lipid molecule on xy plane

GROMACS version:2020.5
GROMACS modification: Yes (Patched by Plumed 2.7.1)

Dear users,

I’m running a targeted molecular dynamics including a transmembrane channel protein and a lipid bilayer. While checking its outputs, I noticed that coordinates of lipid molecules (actually, phosphorous atom) on the xy plane were drifting in a certain direction. Here is an example:

phos_move.panx1.357_plane

This is a time dependency of coordinates of a lipid molecule (phosphorous atom), on the xy plane.
As you see, it was moving in a certain diagonal direction.

Do you know what cause this kinds of problem? Any help and ideas would be very appreciated.
Here I note important parameters;

integrator = md
dt = 0.002
comm_mode = liner
nstcomm = 100
comm_grps = protein+lipids water+ions
cutoff-scheme = Verlet
nstlist = 10
pbc = xyz
verlet-buffer-tolerance = 0.005
rlist = 1.2
coulombtype = PME
rcoulomb-switch = 0
rcoulomb = 1.2
vdwtype = Cut-off
vdw-modifier = Potential-swicth
vdw-switch = 1.0
rvdw = 1.2
DispCorr = No
tcoupl = Nose-Hoover
tc_grps = protein+lipids water+ions
tau_t = 1.0 1.0
ref_t = 303.15 303.15
pcoupl =Parrinello-Rahman
pcoupltype = semiisotropic
tau_p = 2.0
compressibility = 4.5e-5 4.5e-5
ref_p = 1.0 1.0

I appreciate your reading.