[gmx-users] Can't center box with trjconv

[Ignacio Fernández Galván]

I must be doing something wrong, but I can't get trjconv to center the simulation on a residue.

The simulation box is cubic, the system is formed by an 8-atom frozen molecule and 800 water molecules. I try to center the box with:

  $ trjconv -f system.cpt -s system.tpr -o conf.gro -center

(system.cpt and system.tpr are the result of a simulation)

I select group "2" (the frozen molecule) as a center group, and I get this in the output conf.gro:

    1ACR     C1    1   0.074   0.042   1.443  0.0000  0.0000  0.0000
    1ACR     H1    2   0.130   0.139   1.443  0.0000  0.0000  0.0000
    1ACR      O    3   0.133   2.822   1.443  0.0000  0.0000  0.0000
    1ACR     C2    4   2.811   0.056   1.443  0.0000  0.0000  0.0000
    1ACR     H2    5   2.771   0.158   1.443  0.0000  0.0000  0.0000
    1ACR     C3    6   2.731   2.834   1.443  0.0000  0.0000  0.0000
    1ACR    H3a    7   2.622   2.843   1.443  0.0000  0.0000  0.0000
    1ACR    H3b    8   2.775   2.734   1.443  0.0000  0.0000  0.0000

as you see, the molecule is clearly broken (a difference of more than 2 nm in coordinates), and when visualized, it's indeed in the box's edge.

I've tried with different combinations of -pdb res/mol, -boxcenter zero/tric/rect, -ur compact, but I always get the molecule broken, sometimes in a corner, sometimes in an edge.

In case it matters, the initial configuration for the simulation has:

    1ACR     C1    1   0.138870   0.039109   0.000000
    1ACR     H1    2   0.195321   0.135855   0.000000
    1ACR      O    3   0.197788  -0.067044   0.000000
    1ACR     C2    4  -0.009210   0.052984   0.000000
    1ACR     H2    5  -0.049761   0.154673   0.000000
    1ACR     C3    6  -0.089322  -0.054436   0.000000
    1ACR    H3a    7  -0.198186  -0.045502   0.000000
    1ACR    H3b    8  -0.045500  -0.154639   0.000000

and, as I said, this residue (and molecule) is set as frozen, and it is actually frozen during the simulation.

Is there some bug/limitation in trjconv? Am I doing something wrong?

Thanks,
Ignacio