What to do with NCBITAXON? (BioPortal version vs. OBO version)

[syphax-bouazzouni]

No description provided.

[jonquet]

The existence of two versions of the NCBI Taxonomy (both using the acronym “NCBITAXON”) arises from its independent development and use within two distinct communities: BioPortal (via UMLS) and the OBO community. Here’s an explanation of this situation:
1. BioPortal NCBI Taxonomy (UMLS version): https://bioportal.bioontology.org/ontologies/NCBITAXON (also used in AgroPortal)
BioPortal historically hosts an early version of the NCBI Taxonomy in RDF, which was imported from the Unified Medical Language System (UMLS) with the UMLS2RDF tool. This version uses URIs minted within the context of BioPortal by the UMLS2RDF tool e.g., http://purl.bioontology.org/ontology/NCBITAXON/131567 The focus here was to provide integration with other biomedical ontologies and facilitate data interoperability within the BioPortal ecosystem.
2. OBO Community NCBI Taxonomy: https://obofoundry.org/ontology/ncbitaxon.html
Later, the Open Biological and Biomedical Ontology (OBO) community developed its own representation of the NCBI Taxonomy in RDF. This version adheres to the OBO Foundry principles, which emphasize standardized practices for ontology development, including consistent URI schemes The OBO version uses URIs following the OBO PURL pattern, such as http://purl.obolibrary.org/obo/NCBITaxon_. The OBO version is tailored for use in the OBO Foundry ecosystem, emphasizing interoperability and alignment with other OBO ontologies.

Having two versions with the same acronym but different URI schemes create challenges for interoperability and semantic integration.
AgroPortal has historically stored the BioPortal version. Now BiodivPortal likes to use the OBO version. This creates a conflicts in the federation of AgroPortal and BiodivPortal.

We are coming to a moment where we could consider dropping the BioPortal version and adopt the OBO version in AgroPortal as we have more request for this one.

Opinions? @cmungall @naouelkaram @jvendetti @alanruttenberg @jamesaoverton @nlharris
Feel free to CC more folks.

[jamesaoverton]

I believe that both the BioPortal and OBO versions of NCBI Taxonomy are about two decades old now. So both have any number of dependencies built on them, and we have no way of knowing who is using them.

One more factor: A few years ago, the NCBI Taxonomy Browser added a message to the top of each taxon page, e.g. "Taxonomy ID: 2 (for references in articles please use NCBI:txid2)" https://www.ncbi.nlm.nih.gov/Taxonomy/Browser/wwwtax.cgi?mode=Info&id=2. I guess NCBI:txid2 the identifier form that they now prefer.

The good part is that everyone uses the same numeric part of the ID.