Request for 'termReplacedBy' annotation for deprecated terms in NCBITaxon

[Anoosha-Sehar]

Hi, I am Anoosha, and I work with Damion on FoodOn FoodOn ontology. Recently, we came across a few outdated taxon entries that are still linked to older terms in NCBITaxon, but these terms have been deprecated. It’s been a bit tricky to figure out what terms should replace them.

I wanted to check if it would be possible to add a "termReplacedBy" annotation for deprecated terms in the NCBITaxon.owl file. This would make it a lot easier to spot which terms are no longer in use and what they’ve been replaced with. That would be very helpful.

Thanks,

Anoosha

[matentzn]

@anitacaron would you want to look into this? Else, we would have to find another volunteer to see if that information is readily available in

ncbitaxon/Makefile

Line 16 in f0b391c

curl -L -o $@ https://ftp.ncbi.nih.gov/pub/taxonomy/taxdump_archive/taxdmp_$(DATE).zip

[anitacaron]

I can have a look on this.

[anitacaron]

I've checked the files available in the .zip file.

In cases where two terms were merged, we use the annotation oboInOwl:hasAlternativeId with the deprecated term (example). However, there isn't a proper declaration of which terms were deprecated to add the replacement there. This is still on the TODO list.

@Anoosha-Sehar Do you think the outdated terms you've found were updated after November 2024? If so, I might need to run a new release.

[allenbaron]

@anitacaron there are at least a number of outdated virus names, so I think a release is warranted (and I would appreciate it). If there's something I can do to help with getting term replaced by in NCBITaxon, I'm happy to do something to move that forward as well (though to this point I haven't looked at the code that is creating NCBITaxon).

Is the plan to address #33?

[anitacaron]

@allenbaron I'm working right now on issue #33, and I would like to have this before a new release, if possible.

[dosumis]

Hi all,

We've just been hit by this issue too. It looks to me like the conversion still uses the ancient GO merge mechanism - storing alternative_ids but not keeping obsolete IRIs as resovlable & not adding replaced_by assertions. From Claude:

NCBI's taxdump expresses "merged/replaced" via merged.dmp (old_tax_id → new_tax_id). The converter handles it like this:

src/ncbitaxon.py:328-331 — reads merged.dmp and maps merged[new_tax_id].append(old_tax_id).
src/ncbitaxon.py:197-198 — for each new (surviving) node, every merged-in old ID is emitted as oboInOwl:hasAlternativeId "NCBITaxon:<old_id>".
So merged/replaced IDs are not materialized as separate deprecated classes with IAO:replaced_by — they collapse into the surviving class as alternative_id (oboInOwl). delnodes.dmp (truly deleted taxa) is explicitly a TODO at src/ncbitaxon.py:390 — not handled.

@anitacaron - please can this be modernised. Most other OBO ontologies migrated from this several years ago. Thanks!