RMSD calculation

GROMACS version: 2020.4


I am working with a complex includes protein, DNA and small molecule. After my 500 ns simulation, I used to want to calculate RMSD changes of this system during the simulation. I generated an index file with: gmx make_ndx -f complex.gro. And I gave an index to Protein+DNA+SmallMolecule. For RMSD calculation, I selected this group twice, however my RMSD values above 3.5 nm. If I choose Backbone or Protein groups twice, RMSD values are decrasing to ~0.2 nm.
I calculated each group seperately (Protein, DNA and Small Molecule) and the values are ~0.2 nm.
Is the problem about making index file or generating new group?

Thanks in advance

Probably you have pieces of your system jumping across periodic boundaries so the complex isn’t always together. Re-image with trjconv and you should be able to compute the RMSD of the whole complex.

Thank you for your answer.
However I applied image correction to my system and I am sure the system is not jumping. May my system has any other problem?

Without seeing your exact commands, selections, and the plots themselves, there’s not really much to go on.