I don’t know if this is your problem, but we’ve had big issues with the make_pose_from_sequence command. It seems to attach things in a very strange manner (e.g., I’ve found hydrogens shared between different residues). This has been such a big problem for us that we’ve discontinued using that method of making poses, preferring to make a pdb manually in PyMol or by sequence using DeepView.
I’d try to run your script again after making your aa sequence into a pdb, then loading it using pose_from_pdb.
I hope this helps,