I'm studying an homotetramer and i want to identify new sidechain
orientations on a specific helix segment without any more changes in the
protein structure of my initial pdb. So i tried the loopmodel routine:
loopmodel wasn't really designed or benchmarked for sidechain
optimization, but what you're doing looks pretty reasonable here.
?- Is it possible to preserve the symmetry of the sidechains over all
the subunits? or...to perform a total symmetry of the models by a
post-processing step of refinement?
If you want the sidechains to be identical, it may make more sense to
optimize a single chain, then construct the other chains by symmetry
operations. You can apply a symmetry restraint, but it will compete with
the loopmodel statistical potential, of course, so Modeller may not be
able to satisfy it as well as you might want.