#11250 Subclasses of 'cytokine secretion'

[Sebastian Köhler]

Hi GO editors,

I have seen that GO has a class for "cytokine secretion". A subclass of that is "chemokine secretion".
For logical definitions in HPO I would like to refer to the GO classes "interferon secretion" and "interleukin secretion" (I suggest both to be subclasses of "cytokine secretion"). At the moment these two classes are not present GO. Is that something GO might want to consider?

Thank you very much,
Sebastian Köhler

[Chris Mungall]

I think we should make a template for these

[Chris Mungall]

'cytokine secretion' being discussed now... may be some changes afoot...

[Alexander Diehl]

Hi Sebastian,

The argument for 'chemokine secretion' and other chemokine terms is that chemokines represent a true protein family with a high degree of homology and similar functions among the various members.

Interferons too, represent a protein family with structural homology and somewhat more diverse functions. The GO already has terms for 'type I interferon secretion' (alpha and beta interferons), 'interferon-gamma secretion', and 'type III interferon secretion' (interferon-lambda). Essentially, you are arguing for a grouping term based on the lexical similarity of these three terms, ignoring the dissimilarity in their biology. Personally, I think you would serve the users of the HPO better by splitting up your 'Abnormality of interferon secretion' term into three terms reflecting the three major families of interferons. An abnormality in type I interferon secretion manifests a distinct phenotype from an abnormality in interferon-gamma secretion and grouping them together just jumbles up well researched and clinically relevant biology.

An 'interleukin secretion' term would be just another lexical grouping, with even less meaning than 'interferon secretion'. It is fairly arbitrary that certain cytokines are called interleukins and others are not. Things that are called interleukins represent a diverse array of proteins from multiple protein families and having a very wide range of functions, cellular origins, and cellular targets. Calling something an 'interleukin' is mostly an exercise in seeking prestige for one's latest discovery, and has been for years. Suggesting that 'Abnormality of interleukin secretion' conveys any real biological meaning beyond 'Abnormality of cytokine secretion' seems almost ridiculous. Things that happen to be called interleukins are secreted by several different mechanisms following different stimuli by different cell types (some immune, some not). Personally, I think you would serve your users best by merging 'Abnormality of interleukin secretion' into 'Abnormality of cytokine secretion' as a narrow or related synonym.

You might consider reading PMID:12103349 and PMID:12464567 in regards to interleukins. Also, the COPEwithCytokines site has a very good discussion of nomenclature issues surrounding interleukins and cytokines in general: http://www.copewithcytokines.de/cope.cgi?key=Interleukins and http://www.copewithcytokines.de/cope.cgi?key=cytokines

Personally, I oppose the GO adding an 'interleukin secretion' grouping term, or indeed any grouping term based on interleukin as a distinct class of physical entities, since it is primarily an arbitrary lexical grouping that has no place in a realism-based ontology. I also don't think 'interferon secretion' is needed as a grouping term, as the GO has terms representing secretion of the three types of interferon already.

Best regards,
Alex

[Sebastian Köhler]

Hi Alex,

thank you very much for your detailed answer. This was very helpful. We will think about options how to deal with this in HPO.

Best wishes,
Sebastian