edit GO:0042605

[gocentral]

Hi, For the use of term GO:0042605 'peptide antigen binding' by the IEDB (once imported into OBI), we are requesting edits to the definition. The current definition "Interacting selectively and non-covalently with an antigen peptide, a fragment of a foreign protein derived by proteolysis within the cell." is too narrow for us. It excludes the binding of epitopes to BCRs, TCRs, antibodies and MHC molecules when they are not part of a foreign protein, i.e. are self (autoimmunity), or are not part of proteins, but instead are non-peptidic epitopes (carbohydrates, lipids, etc). Additionally, some epitopes are not processed by APC & therefore are not derived by proteolysis within the cell. Because this term has the Comment "Note that this term can be used to describe binding to the peptide expressed by an MHC molecule" we feel it should include all types of epitopes that bind to MHC molecules (self, non-peptidic, and not processed ones). Also the comment should probably be edited to read "Note that this term can be used to describe the binding of a peptide to an MHC molecule." Please let us know if we have misunderstood anything or are attempting to edit the wrong term for our purposes. Thanks! -Randi Vita

Reported by: rvita

Original Ticket: geneontology/ontology-requests/8409

[gocentral]

Randi,

I agree that the definition should be shortened to "Interacting selectively and non-covalently with an antigen peptide," and that the comment should be modified as you suggest.

Thanks,

Alex

Original comment by: addiehl

[gocentral]

• assigned_to: nobody --> paolaroncaglia

Original comment by: paolaroncaglia

[gocentral]

I was wondering if there were any updates to this request? The IEDB/OBI need it to move ahead. Thanks, Randi

Original comment by: rvita

[gocentral]

Dear Randi,

Apologies for the delay. This is now done - I've shortened the definition to "Interacting selectively and non-covalently with an antigen peptide" same as Alex suggested, and I've modified the comment to "Note that this term can be used to describe the binding of a peptide to an MHC molecule".

However, since GO:0042605 peptide antigen binding is a child of peptide binding, it would not be correct to have it include non-peptidic epitopes such as carbohydrates or lipids. The more generic parent term GO:0003823 antigen binding might be more suitable to your needs in that case.

Thanks, Paola

Original comment by: paolaroncaglia

[gocentral]

• status: open --> closed-accepted

Original comment by: paolaroncaglia

[gocentral]

Hi, I like your suggestion of using the more generic parent term GO:0003823 'antigen binding' for non-peptidic epitope binding, but 'antigen binding' GO:0003823 has three children 'lipid antigen binding' GO:0030882, 'MHC protein complex binding' GO:0023023, and 'peptide antigen binding' GO:0042605. So we would use 'peptide antigen binding' GO:0042605 for peptide epitopes and 'lipid antigen binding' GO:0030882 for lipid epitopes. In order to have all of our different types of epitopes represented equally, should we also request terms such as carbohydrate binding, element binding, glycopeptide binding, peptidoglycan binding etc or is that too much? Thanks for your help.

Original comment by: rvita

[gocentral]

Dear Randi,

It may certainly make sense to add broad terms such as "carbohydrate antigen binding". However, I'm not sure if more granular terms such as "element antigen binding", "glycopeptide antigen binding", "peptidoglycan antigen binding" and similar would add much to GO, as, after all, everything could be considered an epitope if presented to the immune system… Perhaps, in such cases, the use of existing general terms such as GO:0030246 carbohydrate binding GO:0001948 glycoprotein binding GO:0043394 proteoglycan binding could be just as informative? (Not sure about "element binding".) However I realize this is a complex issue and I'd welcome further feedback from Alex Diehl if he gets a chance.

Thanks, Paola

Original comment by: paolaroncaglia

[gocentral]

Hi Randi and Paola,

I would favor adding terms like

carbohydrate antigen binding glycoprotein antigen binding proteoglycan antigen binding

in analogous way to peptide antigen binding and lipid antigen binding.

I'm not sure what is meant by "element binding" either. Is there some relation to phenomena such as nickel allergy?

Thanks,

Alex

Original comment by: addiehl

[gocentral]

I guess what we really wanted was a child of 'antigen binding' GO:0003823 that we can use regardless of the type of structure binding to the immune receptor (antibody or T cell) for all epitopes, regardless of the epitope's chemical type. We do not want to have to add each chemical type to GO. Because 'peptide antigen binding' was already a child of 'antigen binding', it seemed that adding a sibling of 'peptide antigen binding' that is "non-peptide antigen binding" made sense. The other GO binding terms such as 'carbohydrate binding' are not children of 'antigen binding' and the "antigen" part is critical for our usage. If 'antigen binding' had not had any children, we would probably have just used it for both peptidic and non-peptidic epitopes and been happy. I will propose we use the parent 'antigen binding' for non-peptidics and 'peptide antigen binding' for peptidic epitopes and if that is accepted, I'll let you know. Thanks for the help.

Original comment by: rvita

[gocentral]

• status: closed-accepted --> open-accepted

Original comment by: paolaroncaglia

[gocentral]

If that's the solution you prefer, that is fine with me.

-- Alex

Original comment by: addiehl

[gocentral]

• status: open-accepted --> open

Original comment by: paolaroncaglia

[gocentral]

The editing of 'peptide antigen binding' was great and we need no other terms related to this request. Thank you!

Original comment by: rvita

[gocentral]

• status: open --> closed

Original comment by: rvita