Rearrange capsid hierarchy

[cmungall]

Carried on from:

• NTR: virus-like capsid #26842

In summary the issues are:

• currently nucleocapsid is part of capsid, which is inverted
• we have 3 terms with very similar biology (a protein shell enclosing nucleic acid, with possible the same evolutionary origins) disconnected in the ontology

@hattrill proposed:

_nucleocapsid
|__viral nucleocapsid
___(p)viral capsid
|__retrotransposon nucleocapsid
___(p)retrotransposon capsid (NEW)
|__virus-like nucleocapsid
___(p)retrotransposon capsid (NEW)
__(p)capsid (NEW)
|___viral capsid
|___retrotransposon capsid
|___virus-like capsid

But also wonders if it if it is "over-kill for the handful of virus-like/retrotransposon proteins, but for viruses, would be worth the effort. Think that is a question I would punt back to GO!"

I also worry about overkill, I don't like overly latticed structures. The above also instantiates the single-child-by-one-axis-of-classification antipattern (if current annotations to nucleocapsid are intended to capture nucleic acid interacting proteins rather than capsid compoments then it seems we need a more specific child term rather than leaving things like the usual lopsided GO christmas tree)

However, I think that it is worthwhile grouping the shared biology here