Is there a reason for different runs of the same .py file give exactly
the same model, with the same DOPE score and all? Does Modeller make
aways the same pathway?
Of course - it wouldn't be very good science if it were not
reproducible! If you want to get different models every time, you will
need to set a different random seed. See
http://salilab.org/modeller/9v1/manual/node99.html