Please recommend alternative homes for out-of-scope obsoletions

[matentzn]

Many terms are obsoleted in GO, and many for the reason of being "out of scope".

For example: #24930

Regardless of whether "aging" is in scope in GO or not, it is an important concept we want to talk about. I do not think it is scalable to debate every time whether or not something is or is not in scope, especially if one is a small curation effort (small compared to GO) that just needs to use some terms from GO;

What I would suggest here is to amend the SOP for the benefit of ontology curators.

1. Please add a new field to your metadata: Consider following "in scope" ontology.

If one was motivated, which other (OBO) ontology would be an appropriate home for the obsoleted term?

When the answer is, "there is none", perhaps maybe avoid obsoletions a bit more, or help understanding what we should be doing. For example, FlyBase switched to using an EFO term for "aging" for their ontology curation - and we really should not use application ontologies as dependencies for reference ontologies.

2. Please provide a more detailed and concrete reason/motivation

... for why a term should be obsoleted, one that is detailed enough so that a user can rethink their own curation decisions. #24930 only contains some boilerplate text for the obsoletion reason, which is IMO not enough..

Side note: Why am I personally impacted?

I am building many application and reference ontologies and fight for a comprehensive interoperable system of ontologies in the OBO Foundry; having application ontology classes, from outside OBO at that, creeping into OBO ontologies is a bit of a danger to that goal, so I would like to stop it in its tracks if possible.

Thanks for your time :)

[pgaudet]

Hi @matentzn

I understand it's a pain to transfer mappings. However after many years of discussion, ontology editors have decided that 'aging' is not a normal process, but a result of cellular malfunctions that accumulate over time, including oxidative damage, etc.

Is there no other ontology that you could refer to in order to define your terms?
It seems there is a need for some sort of a meta ontology that would group 'normal' biological processes (ie, GO), plus other phenomena relevant to living organisms, like aging, and possibly others. Is this right?

Thanks, Pascale

[cmungall]

I think choosing where the replacement lives is a great and interesting topic for a COB ticket! I'd be happy to comment over there.

There are some things we can do to improve the overall process here to avoid causing frustration. GO has an SOP for obsoletion that allows for comments and objections. We announce on go-friends, the announcement tracker, as well as the issue itself. All GO curators are aware of these mechanisms but we could do more to advertise them on the OBO slack.

I think @pgaudet's description is perfect "ontology editors have decided that 'aging' is not a normal process, but a result of cellular malfunctions that accumulate over time, including oxidative damage, etc." (though I'd change "normal process" to "genetically encoded program"). But the problem is people don't see this when the click on the github issue link that accompanies the obsolete term, you have to navigate deep in many issues and know a lot of surrounding context. I think we should try and make the original issue better reflect our reasoning. This helps us too, when we go back years after the fact. But please also appreciate that we have a lot of legacy, over 20k issues, and need to prioritize what we are funded to do.

A lot of this comes down to when we clarified the definition of the root BP term - it is hard disseminating this kind of information outwards, I have been trying my best on OBO/COB channels, but sometimes information doesn't stick. Once you read the definition it's clear why aging and some other obsoleted terms don't belong. But perhaps we should put something more prominently in the main GO documentation about this.

[matentzn]

I am extremely sad that aging is not a normal process, I will try to do better! Maybe more green leaves will do :P

Sorry couldn't help it.

[ValWood]

It would also be really useful to promote and use the more precise description of what a GO "biological process" is. Of course ageing is a process, cell death and cancer are processes, but they are not "evolved processes" that can be captured by the normal programmed roles of gene products that we aim to capture with GO.

[matentzn]

Thank you @ValWood, @pgaudet and @cmungall for your comments!

Do I see this correctly that the recommendation here is:

• Write a more restrictive / precise definition of biological process more in line with "genetically encoded program"
• Define a generalised notion of a "biological process" in some ontology like COB and have GO:bp be a subclass of it
• Create an "Ontology of Biological processes which are not completely genetically encoded" (BOUNCE) and add terms like aging and cell death there?

[ValWood]

This sounds like a good plan to me.

[pgaudet]

I think the BP definition that Paul and I worked on a couple of years ago is clear about ''genetically encoded program':

A biological process is the execution of a genetically-encoded biological module or program. It consists of all the steps required to achieve the specific biological objective of the module. A biological process is accomplished by a particular set of molecular functions carried out by specific gene products (or macromolecular complexes), often in a highly regulated manner and in a particular temporal sequence.

But I am happy to include any improvements you suggest.

for the next 2 proposals, this should be moved to the trackers of the respective ontologies.

[pgaudet]

I have also improved the obsoletion comment of 'aging:

The reason for obsoletion is that it is not a normal, genetically encoded biolgical program, but a phenotype resulting from cellular malfunctions that accumulate over time, including oxidative damage, etc.

I am not sure we should add 'consider' terms (for example response to oxidative damage or cellular senescence), since that may lead to overannotations to these terms, when people want to capture phenotypes rather than the underlying process.

Pascale

[ValWood]

The definition is good, but it might be useful to add a information about what GO BP excludes to this high-level information about GO
https://www.geneontology.org/docs/ontology-documentation/

Just reading about GO on the website, and without reading, and comprehending, the definition of the root node, this important point is a bit hidden.

[ValWood]

A case in point is the misinterpretation of GO biological process. Used here by the experimental factor ontology to group
heart plant rejection?
HIV mother to child transmission ?
mental process?
seroconversion?
lack of efficacy?
disease course ?
and all manner of non-biological process terms