search

open-physiology / apinatomy-models

Working and published ApiNATOMY models
2 stars 4 forks source link

Meaning of phenotype terms #36

Open dbrnz opened 1 year ago

dbrnz commented 1 year ago

We still have:

ilxtr:neuron-type-sstom-11 phenotype ['ILX:0105486'] is unknown
ilxtr:neuron-type-sstom-12 phenotype ['ILX:0104003'] is unknown
ilxtr:neuron-type-sstom-13 phenotype ['ILX:0105486'] is unknown
ilxtr:neuron-type-sstom-14 phenotype ['ILX:0104003'] is unknown

What is the meaning of these terms?

@tgbugs commented about this here but the terms still appear...

tgbugs commented 1 year ago

Here is the mapping.

IlxTerm('ILX:0104003', label='Excitatory')
IlxTerm('ILX:0105486', label='Inhibitory')
dbrnz commented 1 year ago

Thanks.

The other phenotype terms though are like ilxtr:SensoryPhenotype, in the ilxtr: namespace. Should these ILX ones not be ilxtr:Excitatory and ilxtr:Inhibitory?

tgbugs commented 1 year ago

I have never used them as such, because I used the ILX: form directly in the python code since it already existed (I usually only use ilxtr: in cases where a term does not exist yet or when I don't know whether the semantics I need in a particular context match the modelling in some other context).

dbrnz commented 1 year ago

So are you saying that the ilxtr:SensoryPhenotype and similar ilxtr: terms that SCKAN uses for phenotypes will eventually become ILX: terms??

tgbugs commented 1 year ago

Yes, they will eventually become ILX: terms, however we will not switch the identifiers in the ontology so no changes would need to be made on your end. When the receive a corresponding ILX: term they will redirect to that term.

dbrnz commented 1 year ago

At the moment we get this mixture of ilxtr: and ILX: terms for phenotypes -- does it not make sense for them to all be in the same namespace? (I know from a RDF point-of-view it makes no difference; just me wanting consistency!)

tgbugs commented 1 year ago

The use cases for the various prefixes are determined by requirements that are mostly orthogonal to this, precisely because from an RDF standpoint it makes no difference, and if we started trying to clean up and make everything consistent and human readable/accessible we could spend all our time on that :).

dbrnz commented 1 year ago

As an aside, parts of the codebase rely on namespace prefixes having a consistent meaning, which is not assured by RDF. Hopefully this won't trip anyone up, but things are likely to stop working if different prefix is used for a namespace...