The documentation for morph reads very protein-centric, and testing it with default options certainly makes it look like ligands are removed. 

Is there a way to have ligands included in a morph (assuming the same atoms are present in both models)?

Cheers,

Alexis