search

NCI-Thesaurus / thesaurus-obo-edition

OBO Library edition of NCIt
22 stars 8 forks source link

Would NCI-Thesaurus accept more animal names like "heron"? #51

Open ddooley opened 5 years ago

ddooley commented 5 years ago

We're looking at having a growing roster of animal names that aren't necessarily sequenced and therefore not in NCITaxon, but which we would dbxref other taxonomic identifiers to. The terms wouldn't fit CDISK aims probably, but they are needed for lab test metadata for viral and bacterial pathogens for various projects. FoodOn would probably migrate existing animal name ids to a comprehensive ontology source of these to encourage data normalization; alternately FoodOn could/would become the home for them, if that is the most opportune way forward. Thanks for any info... Damion

balhoff commented 5 years ago

@ddooley why not use NCBITaxon:8899 (Ardeidae)?

https://www.ncbi.nlm.nih.gov/Taxonomy/Browser/wwwtax.cgi?mode=Info&id=8899

ddooley commented 5 years ago

So far the strategy is to offer up an entity (class) with common name label for each food related organism, and link each such thing to one or more NCBITaxon via 'has taxonomic identifier' object property or other ITIS etc schemes via annotation. It allows a 1-many relation between food common name and the taxa it applies to. Admittedly this abstraction isn't necessary for many plants and animals, but it solves all sorts of reference issues around fish names, shrimp / prawn , 5 kinds of "spinach" etc. And it keeps separate the NCBITaxon reference taxonomy hierarchy from the mix and mash common name multi-homed graph. (And it avoids historical issue of taxonomy reassignment).

balhoff commented 5 years ago

@ddooley thanks, that makes sense. There are some taxon unions defined in go-plus, e.g. 'Fungi or Dictyostelium'. This may not be quite the same as what you want to do.

ddooley commented 5 years ago

Not to nitpick, but 'Fungi or Dictyostelium' - i now see that kind of conjunction exists in the go-plus file but not yet findable in OLS, and Ontobee only indirectly has it through KTAO? http://purl.obolibrary.org/obo/NCBITaxon_Union_0000022 seems to break. How would NCBITaxon_Union sit with respect to lookup services?

Also, might NCI-Thesaurus go in the direction of using NCBITaxon ids instead, or at least 1-1 mapping between those ids and NCI-Thesaurus? That would be in the OBOFoundry spirit!