GO:0004901 (GM-CSF-R) is NOT a kinase

[gocentral]

GO:0004901 (granulocyte macrophage colony-stimulating factor receptor activity) is currently listed as a child of GO:0004714 (transmembrane receptor protein tyrosine kinase activity). This is most likely incorrect. GM-CSF-R does not have an intrinsic protein kinase domain, or intrinsic protein kinase activity, though one subunit of the receptor (gp130) associates with the non-receptor Jak kinases, and the pathway also signals through other kinases.

I propose that the link between 4901 and 4714 be removed.

As an aside, I noticed that GO:0004901 is a child of GO:0004900 (erythropoietin receptor activity). I'm not at all an expert in this area, but I believe that GM-CSF, while it is a cytokine, would not generally be classified as an erythropoetin (for instance, erythropoetin's best known function is as a growth factor for red blood cells, while GM- CSF, as the name suggests, is a growth factor for granulocytes and macrophages. It would probably be correct to instead make 4901 a child, rather than a grandchild of GO:0004896 (hematopoietin/interferon-class (D200-domain) cytokine receptor activity). The lack of a decent definition for these terms (other than the title) certainly hinders evaluation of the correctness of these assignments.

Reported by: ceolas

Original Ticket: "geneontology/ontology-requests/1341":https://sourceforge.net/p/geneontology/ontology-requests/1341

[gocentral]

Logged In: YES user_id=436423

OK, I have removed the relationship between 4901 and 4714. I'll leave this item open, because we'll have to look into the erythropoietin term and topic a bit more closely to figure out what to do. At a mimimun we should be able to define 4896 and 4900 -- or make them obsolete.

m

Original comment by: mah11

[gocentral]

Logged In: YES user_id=473796

On a general note, a lot of these receptor terms look suspiciously like gene products - the relationships they have with other terms encode gene product rather than function information. The generic GO definition of receptor is 'The function of combining with --- to initiate a change in cell function.', so 'erythropoietin receptor activity' would be OK defined as "The function of combining with erythropoietin to initiate a change in cell function.", but I don't think that we can justify the continued existence of 'hematopoietin/interferon-class (D200-domain) cytokine receptor activity", since it is refers to a class of gene products AND a sequence feature. The term should be split into the constituent activities (hematopoietin receptor activity, interferon receptor activity).

There are quite a lot of receptor terms linked to catalysis activity terms - these relationships encode gene product information which would be better captured by concurrent annotation, and so should be removed.

Original comment by: girlwithglasses

[gocentral]

Logged In: YES user_id=436423 Originator: NO

with luck, the receptor terms will be dealt with as part of the signaling overhaul

Original comment by: mah11

[gocentral]

• milestone: --> TermGenie
• assigned_to: nobody --> gomidori
• status: open --> closed-fixed

Original comment by: mah11