Weird Protein Structure after NPT Production

GROMACS version: 2018.3
GROMACS modification: No

I have been running simulations on protein-protein interaction, and ran NPT production for 100ns. Then I tried to see how the system looked on VMD, and I found it not looking right.

First off, parts of the proteins were sticking out of the simulation box at the last time point.
Second, when I used trjconv to visualize the system at different time points, I saw some parts of the proteins “torn” out and some oriented in an odd way.

Below I put a link to some of my files. Could anyone help me solve the problem?

Thank you.

zip file with .gro, .xtc files and a .gro file at 10ns

It is due to PBC. Make molecule whole with -pbc whole while using trjconv.