This is try11-opt2.repack-nonPC, whose backbone was determined by
undertaker, but with the sidechains repacked by Rosetta.
It was not the best-scoring model with the undertaker cost function, but
Rosetta hated it least of the various repacked-sidechain models it
scored. 

It is based on a possible alternative placement for the C terminal helix.
It seems less likely to us and scores worse with the undertaker cost
functions than the helix placement of model 1.
We are also not fond of the burial of E14 in this model.