I need to refine some flexible regions (mainly long loop and linker
regions) of my models using some enhanced sampling engines implemented
in modeller like simulated annealing.
OK, just select them with select_loop_atoms and use loopmodel.
Could you suggest me some example of the realization of this method in
the modeller python script indicating its differences from the method
proposed in the advanced tutorial (listed below).
Sorry, I don't understand this question. All Modeller models are refined
by simulated annealing, so there's nothing special about loop modeling
in that respect. The main difference between the construction of
comparative models and the refinement of loops is that the latter turns
on a statistical potential for nonbonded pairs. This is helpful in
regions where you have no template structures to provide this information.