Open gaurav opened 7 years ago
I think that's a good thing to do anyway. Abusing labels for identifiers is fine for getting a proof of concept started, but is still an abuse. So I agree let's start to think more thoroughly about what the model for describing OTUs should look like.
There's some work I've done previously that's somewhat related, in that it proposes a model for annotating OTUs (in this case with TNRS matches): https://github.com/phylotastic-legacy/ontologies/tree/master/tnrs
DendroPy cannot load a phylogeny with duplicate taxon labels, which means that phylo2owl.py can't, either. Fixing this will mean moving away from our previous model of naming leaf nodes from the taxon name (e.g. "
pg_2357:Alobevania_gattiae
") to using the same Node-number scheme we use for internal nodes (e.g. "pg_2357:Node_134
"), and then providing them with either anrdfs:label
or making them explicit instances ofcdao:TU
and figuring out the best name of assigning taxonomic names to those.