Consider a non-isomorphic mappings to NCBI Taxon database

[cmungall]

The NCBI taxonomy database is a database, not an ontology. As such that automatic translation includes various things that are not ideal from an ontology modeling perspective

If we were to make the OWL non-isomorphic here are some things we might change:

• change root class name Propose changing the label on the NCBI top level term to 'NCBI root' #10
• depluralize taxa common names "human" not "humans"
• exclude non-taxon entries, or rewire them to a different hierarchy
  • e.g. https://www.ebi.ac.uk/ols/search?q=sample&ontology=ncbitaxon
  • metagenomes (e.g "soil metagenome")
  • environmental samples, e.g http://purl.obolibrary.org/obo/NCBITaxon_743727 environmental samples <archaea,phylum Candidatus Aenigmarchaeota>
  • random terms like "mixed sample"
  • host samples, e.g. "human sample"

We could also use this as an opportunity to merge into a single ncbitaxon representation see OBOFoundry/OBOFoundry.github.io#1383

[cmungall]

Additional concerns if they diverge: biopragmatics/bioregistry#1044