Dear Modeller Collegues, After sticking for a long time to the old TOP language, I am trying to switch to Python because I need to use some of the high-end Modeller features such as dope_loopmodel. Here is my problem : I want to refine several loops in an existing model using dope_loopmodel. Some of the loops have disulfide bridges with the rest of the model or between themselves (the disulfides do not exist in the template). A previous case of using restraints or patches in loop refinement already appeared in the discussion list but there, loopmodel was used on the fly after automodel so that the redefined patches and restraints automodel subclasses were active during loop refinement. What happens when one uses only loopmodel or dope_loopmodel to refine an existing model ? I suppose it is possible to import patches subclasses from automodel. I tried to fiddle but got nowhere. Could somebody help me ? To make myself clear, I am pasting below an example (on one loop) of what I was doing with the old TOP language (it worked, at least computationally). Thanks in advance for your help. Best regards. Michel