cytokine receptor binding TPVs

[gocentral]

From a David/Midori chat:

A large number of receptor binding terms are under 'cytokine activity', e.g. fibroblast growth factor receptor binding, all of the interleukin receptor binding terms, etc. This causes path violations because a given gene product may bind to the receptor but not have cytokine activity (e.g. a protein that binds to the cytoplasmic side of a transmembrane receptor).

We should move the receptor binding terms out from under cytokine activity; they would become direct is_a children of receptor binding (information would be lost for annotated gene products that are cytokines, but this is not as bad as having things that aren't cytokines erroneously annotated). We can add child tems of the 'x receptor binding cytokine activity' sort, and annotations should be reviewed to see which should be "moved" to the child terms.

Reported by: mah11

Original Ticket: "geneontology/ontology-requests/4289":https://sourceforge.net/p/geneontology/ontology-requests/4289

[gocentral]

Logged In: YES user_id=865072 Originator: NO

I actually recommended this change about three years ago and favor a generic "cytokine receptor binding" term to be parent to all these terms.

I consider "cytokine activity" to be a fundamentally different type of function term, and have consequently coannotated to both this term and a "cytokine receptor binding" term if data was available in the paper to support both annotations (I was annotating to the definition and ignoring the DAG structure).

Thus, I favor this change.

Thanks,

Alex

Original comment by: addiehl

[gocentral]

Logged In: YES user_id=436423 Originator: YES

It would literally take only a few seconds to move 'cytokine receptor binding' ... is it worth doing so immediately, or would it be better to wait until I have time to add the child terms as well?

(I'm working on some additions for metacyc2go mapping, and I'd like to get that more or less done before I take on anything else biggish.)

m

Original comment by: mah11

[gocentral]

Original comment by: mah11

[gocentral]

Logged In: YES user_id=436423 Originator: YES

Hi David & Alex,

OK, I've had a bit of time to work on this now. I took the current children of 'cytokine activity' that have 'receptor binding' in the names, and created corresponding 'receptor-binding cytokine activity' terms. The terms are in the attached file (OBO stanzas). Could one or both of you review the terms? In particular, I'd appreciate if you'd check for:

- any terms that shouldn't be included;

- any missing terms;

- synonyms and their scopes (also see below);

- the usual sorts of mistakes, such as typos.

Note that I haven't included all parents -- each term will have the corresponding 'receptor binding' term as an additional is_a parent, but I've shown this parent only for the first term in the file.

Synonyms: The receptor binding terms (at least in this part of the ontology) have 'receptor ligand' narrow synonyms. For example:

[Term] id: GO:0001531 name: interleukin-21 receptor binding namespace: molecular_function def: "Interacting selectively with the interleukin-21 receptor." [GOC:ai] synonym: "IL-21" NARROW [] synonym: "interleukin-21 receptor ligand" NARROW [] is_a: GO:0005126 ! hematopoietin/interferon-class (D200-domain) cytokine receptor binding

I'm wondering whether the synonyms should be moved to the 'cytokine' child terms instead, and if so, should they still be narrow in scope. I've included narrow synonyms in the attached file, but what do you think?

On a related note, GO:0005126, 'hematopoietin/interferon-class (D200-domain) cytokine receptor binding', should be defined; would this do?

def: Interacting selectively with a cytokine receptor that contains one or more D200 domains; the D200 domain is composed of two fibronectin type III domains. [PMID:10220409]

If it's ready to go, either as is or edited as you recommend, before I leave work tomorrow, I can add the terms and do the relationship changes before the muscle content meeting; otherwise I can work on it again when I'm back.

m File Added: cytokine.txt

Original comment by: mah11

[gocentral]

Logged In: YES user_id=865072 Originator: NO

Midori,

I would like to have a bit of time to look more deeply at your file and think about the biology behind the terms. For instance, there is no IL-14 in human, mouse, or any other species. "IL-14" has never been cloned, and the reported sequences had errors in them and were withdrawn. The gene symbols for IL-14 in human and mouse have been renamed to match a different protein, one unlikely to be a cytokine, found at the asserted genetic locus for IL-14.

Also, some of the other terms, like the IL-17 terms need to be revised in the light of the fact that there are a family of IL-17s.

Also, I would like to replace (obsolete) GO:0005126, 'hematopoietin/interferon-class (D200-domain) cytokine receptor binding' which is I believe a very odd term indeed that does not reflect the literature of cytokine receptors.

Your approach is not exactly what I expected. I have felt that "cytokine activity" should stand by itself, and "cytokine receptor binding" would be created to be the proper parent of "cytokine X receptor binding" terms.

While I thank you for your work, I really think I should take this over -- there are too many issues here and I think an immunologist should do this revision. Much of what is in the GO regarding cytokine nomenclature is flawed, and I want it fixed, not just perpetuated in a new set of terms. As my cytokine books are in transit at this point, it may be a few weeks until I can do tis properly.

Sorry,

Alex

Original comment by: addiehl

[gocentral]

Logged In: YES user_id=436423 Originator: YES

Hi Alex,

This is precisely why I have not tried to implement anything at all in the "live" ontology yet.

The content of that file is based my imperfect understanding of David's brief description of the problem and his proposed solution -- in other words, it's been a biology-geek take on the kids' game of "telephone," so I'm not very surprised (and not at all bothered) that it's so far from perfect at this first pass. You are much better versed in this area even than David, let alone me; I'm glad you're involved in this work and I'm happy to wait until you have time to do a thorough job on it.

I didn't mean to suggest that "cytokine receptor binding" would not be created -- though I now realize that I omitted it from the file. One point I would ask you to clarify with David is whether "cytokine activity" should remain an is_a child of "receptor binding"; my impression from our conversation was that David thought that parentage was correct -- but either his understanding or my impression of it could be mistaken.

I will assign this SF item to you so that it will be apparent that the biology is in your ballpark, and so that everyone else will wait to hear more from you before acting on it. When you do have the biology sorted out, I'm quite willing to do some or all of the actual editing.

Thanks again for all of your efforts to improve GO's representation of things immunological!

Midori

Original comment by: mah11

[gocentral]

• assigned_to: nobody --> aledie

Original comment by: mah11

[gocentral]

Logged In: YES user_id=436423 Originator: YES

just a housekeeping note: when we do add terms, the IDs will be different from those in the attached file. m

Original comment by: mah11

[gocentral]

This has been superseded by work done for two or three much more recent SF items.

Original comment by: mah11

[gocentral]

• status: open --> closed-out-of-date

Original comment by: mah11