[gmx-users] Same starting structure but different trajectories

[Emanuel Peter]

Dear Gromacs users,

At the moment I have some doubt on my run input .mdp file. 
I simulated a protein structure in water and I used parameters as following in my mdp file:

title                    = MD simulation of xxx
;Preprocessor
cpp             = /lib/cpp
;Directories to include in the topology format
include          = -I../top
 
integrator         = md
 
dt             = 0.001

nsteps           = 1000000

nstxout          = 10000
;frequency to write velocities to output trajectory file
nstvout          = 10000
;frequency to write energies to log file
nstlog           = 10000
;frequency to write energies to energy file
nstenergy         = 10000
;frequency to write coordinates to xtc trajectory 
nstxtcout         = 10000
;group(s) to write to xtc trajectory
xtc_grps         = prot sol XXX
group(s) to write to energy file 
energygrps         = prot sol  XXX 
;Frequency to update the neighbor list (and the long-range forces, 
;when using twin-range cut-off's). 
nstlist          = 10
;Make a grid in the box and only check atoms in neighboring grid cells 
;when constructing a new neighbor list every nstlist steps. 
ns_type          = grid
;cut-off distance for the short-range neighbor list
rlist             = 1.4 
;treatment of electrostatic interactions
coulombtype         = PME
fourierspacing           = 0.12
pme_order                = 4
rcoulomb         = 1.4 
;treatment of van der waals interactions
rvdw             = 1.4
; Periodic boudary conditions in all the directions 
pbc                      = xyz
;Temperature coupling
tcoupl           = nose-hoover
tc-grps          = protein NA+ CFP SOL   
tau_t             = 0.1  0.1 0.1   0.1    
ref_t             = 300 300 300 300 
;Pressure coupling
Pcoupl           = Parrinello-Rahman
Pcoupltype               = isotropic
tau_p             = 1.0
compressibility      = 4.5e-5
ref_p             = 1.0
;Velocity generation
gen_vel          = yes 
gen_temp         = 300
gen_seed         = 173529
;Constrain  bonds
constraints         = none

My problem is as following:

I have simulated the protein with exactly the same starting structure, the same electrostatics, the same 
seed for velocity generation and the same thermostating scheme three times, but I did not get exactly the same trajectory. 
I took the following measurement for comparing my trajectories:

g_mindist -f traj.trr -s input.tpr -n index_atom1_atom2.ndx -od dist_atom1_atom2.xvg

The atomic distances always differ from each other by 0.1 to 0.3 angstrom whereas the starting value at t=0 is the same.

Could you tell me why I get in three different simulations with 3 totally equivalent inputs three different trajectories?
In my opinion it should give 3 times a identical trajectory. 

Thank you in advance.

Best regards,

Emanuel Peter