1. There are no raw input_files for this demo, the PDBs are already clean and Rosetta-readable (in the rosetta_inputs directory).  The sequences are not identical, and neither are the docked orientations.  (Note: each complex has two chains, each chain has 46 amino acids.  Each complex also has two zinc atoms).

2. The clustering protocol should be used to group the 166 starting structures according to their structural similiarity.

3. Once you write the demo, we recommend that you view the PDBs in a few of the clusters to check the results by eye.

4. A common question on the forums is how to cluster a very large number of structures, and what options should be used optimize clustering of 100 structures versus 1000 versus 10,000 etc. For example: memory problems arise when the number of clusters gets too high, therefore, a process where you increase the RMSD angstrom cutoff in order to cluster all structures, then re-cluster each resulting cluster with a smaller RMSD cutoff is one way to get around this. 

5. Any insight you can give in the README.dox would be very helpful.  For this demo, though, we wanted to give you a manageable set of 166.