[gmx-users] "trjconv -pbc whole" and SPC water

[Michael Patra]

Dear all,

I have a problem with "trjconv" and the option "-pbc whole". The documentation
says that this will put the atoms in the box and then make broken molecules
whole (a run input file is required).

If I understand this correctly, it is possible that a bit of a molecule can
stick out of the simulation box - otherwise that molecule would need to be
broken. Mostly this function seems to work as I expect it to work but
occasionally I see that an SPC water molecule gets broken (i.e. two of the
three atoms get a z-coordinate like 0.05 and the third atom one like 79.90 if
the simulation box has size 80). The source code for the functions to put
particles in the box is somewhat incomprehensible but in mk_mshift() there is
an (incomprehensible) remark on molecules containing only settles and no
bonds - like SPC water.

My question: Does "trjconv -pbc whole" simply not work for SPC water at all, or
am I doing something wrong?

Yours sincerely,

- Michael Patra