Group cell, subcellular structure, and gross anatomical entity under 'anatomical structure' grouping

[cmungall]

E.g.:

• anatomical entity
  • material anatomical entity
    • gross material anatomical entity
    • cell
    • subcellular anatomical structure

Note: there is an elephant in the room we haven't discussed which is immaterial anatomical entities.

FAO @matentzn I recall you saying that the phenotype group do include cell but don't include subcellular under AE. Are you sure you have this right?

[matentzn]

This decision was made due to the fact that cellular components life in GO, while cells and above life in CL and UBERON. I dont think anyone in our effort is strongly opposed to extending the anatomical patterns conceptually to cover sub-cellular stuff. Our DPs should be flexible enough to re-use special GO relationships where it is appropriate, but that does not mean the patterns between AEs and CCs should not be aligned (in fact; they are: increasedNumberOf and similar patterns are identical, only with different filler range).

The only thing I can think of top of my head that was a bit more "conceptual" was that some groups measured the "concentration" of cellular components rather than counting them (increased amount). This would not make sense for AEs. @dosumis is off email atm, but maybe he can also add his perspective here.

[matentzn]

I am just realising that for our automated matching, it is a bit hard to distinguish cell (which currently resides under anatomy) and cellular components, because GO asserts a cell to actually BE a cellular component. Are there any news on this issue here? We would like to follow whatever OBO core suggests here, and it is not problem to do so; right now all cells are matched by both anatomical entity AND cellular components patterns - both of which we still treat as conceptually independent (not to say disjoing). @dosumis

[cmungall]

Cells are AEs. But you can ignore the GO cc class and treat it as ad hoc grouping

On Tue, Nov 19, 2019, 03:33 Nico Matentzoglu notifications@github.com wrote:

I am just realising that for our automated matching, it is a bit hard to distinguish cell (which currently resides under anatomy) and cellular components, because GO asserts a cell to actually BE a cellular component. Are there any news on this issue here? We would like to follow whatever OBO core suggests here, and it is not problem to do so; right now all cells are matched by both anatomical entity AND cellular components patterns - both of which we still treat as conceptually independent (not to say disjoing). @dosumis https://github.com/dosumis

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/OBOFoundry/Experimental-OBO-Core/issues/45?email_source=notifications&email_token=AAAMMOK7GK6RLJGITP7NSBLQUPFF3A5CNFSM4ILDFOZ2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEEN4EGA#issuecomment-555467288, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAAMMOP3TDRYFGOPL4SX7PTQUPFF3ANCNFSM4ILDFOZQ .

[matentzn]

Thanks! Practically, this means that before I run dosdp-match, I will remove the GO:Cell sub GO:CC axiom. This will cause GO:Cell to classify under AE (because of GO:Cell = CL:Cell axiom).

[sbello]

@matentzn @cmungall We already include subcellular stuff in the MP (axons, dendrites, nuclei, mitochondria), as I recall we were drawing the line at proteins or macromolecules. See for example: http://purl.obolibrary.org/obo/MP_0008415 abnormal neurite morphology http://purl.obolibrary.org/obo/MP_0011635 abnormal mitochondrial crista morphology http://purl.obolibrary.org/obo/MP_0005060 accumulation of giant lysosomes in kidney/renal tubule cells

[bpeters42]

'Anatomy' is often used as ~'Parthood'. For example there are 475k google scholar articles mentioning the anatomy of social networks: https://scholar.google.com/scholar?hl=en&as_sdt=0%2C5&q=anatomical+structure+social+network&btnG=

So I would strongly prefer that we have labels that spell out what kind of 'anatomy' is being considered. 'gross anatomy part' already does that (referring to above cellular level). So why not also 'subcellular anatomy part' (or something like that), and make them distinct from gross anatomy? And we can retain 'anatomical entity' as a grouping term, spelling out all the children covered by it?

I suspect the remaining issue will be things like bodily fluids, which don't readily fall into the larger/smaller than individual cells distinction. But that is what we are here to figure out:)

[cmungall]

Discussed on GO call today

We will start using {animal,plant,fungal} gross anatomical part {development, ...} so would like these terms materialized in the various AOs:

https://github.com/geneontology/go-ontology/issues/22994