PMI2 RMF hierarchies of input and output RMFs different

[shruthivis]

There is a difference in the hierarchies between output and input RMFs of the PMI version I am using.
For example, after sampling:
/salilab/park1/shruthi/sampcon/pmibugs/1AVX/1AVX_4.1
output/initial.0.rmf3 for example has a hierarchy like System->State 0-> Molecule A, Molecule B.

But after getting the top 500 models from the run using AnalysisReplicaExchangeMacro (Script: /salilab/park1/shruthi/sampcon/get_top_models_each_run.py)
the hierarchy becomes State 0->Molecule A, Molecule B (missing System).
e.g. all_models.499/0.0.rmf3 in the same example directory.

Apparently Chimera can read both kinds of RMFs but PMI itself cannot.
When I try to run the Analysis macro again to do clustering etc. it fails (script: /salilab/park1/shruthi/sampcon/cluster_k1.py)
PMI does not throw up an error but simply outputs an empty distance matrix and AVERAGE_RMSD in the stat files in the cluster directories is 10000000000.0, because it does not read any coordinates from the RMFs.