High TM-score (91.3) and well-matched result but with very low DockQ value (0.05)

Hi!

When I tried to test a complex (PDB ID: 6M0J), I got a high TM-score (91.3) and well-matched result but with a very low DockQ value (0.05),

For computing DockQ value, I just download the pdb file (6m0j.pdb) from [PDB: 6M0J](https://www.rcsb.org/structure/6M0J), and change the chain name from 'E' to 'B' in Pymol with code:
* alter sele, chain='B'

Then I used renumber_pdb.pl in DockQ/scripts to renumber the atoms of chain B in the 6m0j.pdb, and got 6m0j.pdb.renum.
* renumber_pdb.pl 6m0j.pdb

Finally just running:
* DockQ.py ./6M0J/ranked_0.pdb ./6m0j.pdb.renum

I have tried to repeat some examples in the AlphaFold-Multimer paper, such as Fig 4c and Fig5, and got a similar result with no problems, but only the result of this instance looks really strange.

Does anyone know the reason for it?
Thanks