NTR: defense response to parasitoid wasp infection

[hattrill]

Could I have:

defense response to parasitoid wasp infection

is_a defense response to insect GO:0002213

Reactions triggered in response to the presence of parasitoid wasp eggs or larvae that act to protect the cell or organism. (PMID: 22588641, PMID:32396065).

GOC:ha

Thanks

[ValWood]

HI @hattrill
The multi-species group is hooing to phase out these response to roganism/clade terms because they cause inconsistent annotation (people are using them often in place of the correct , more specific term describing the actual process) and the processes aren't really "species' or 'clade' specific. They are largely 'molecule specific'.

Is there a way you could capture the 'pathogen interaction' by connecting the fly gene to the pathogen gene (i.e as a reception substrate for example)?

[hattrill]

For so called antimicrobial peptides (AMP), the mechanism of action is often unknown. So, we don't know what they interact with and how they clear the organism.

In many instances, all we know is that they are small, induced upon infection (usually by an immune-activated pathway, for which we might know the biological agent and their expression has a negative effect on the invading organism. They probably act by punching holes in membranes or coating targets....we just don't know. All we really know that they are often specific to the organism.

[hattrill]

I can see that they can be used inconsistently, as they are very broad.

Is the innate immune response branch going to have the organism specific terms removed as well?

[ValWood]

There are two main issues with the current situation (also independently noted in numerous conversations by the multispecies working group @dsiegele and @mgiglio99 and by PHI-base collaborators).

1. When these terms are used, the
   "GO:0051817 modulation of process of other organism involved in symbiotic interaction" terms are often ignored.
   ( not helped by the fact that this term and the descendants are really difficult to locate, and the "response to species" terms are very prominent at a high level, so more improvements are required).

Ideally, the annotation would say something more specific about how the host is interacting with the pathogen.

and

2. If the annotations to the correct term are made, we have independent annotations are describing orthogonal aspects of the same observation that should be combined.

My previous preferred solution was to make this branch only available in extensions (this has worked well for the phase terms), but the sustainability of requests for every type of pathogen remains.
I previously thought we should not need to refer to the species because if annotations are made correctly the interacting organism will be in the taxon field, but this is quite hidden and I'm not even sure that all tools observe this requirement for pathogen-host interaction annotations.

Perhaps the most sustainable way to deal with this, and to capture any species at any level of granularity, would be to enable an extension "interacting_species" (or similar) on any term/annotation under
GO:0044419 interspecies interaction between organisms

This would resolve all of the current issues:

1. We would be able to cull all "response to species", and would not need to add terms for every species
   ...the current situation is even worse than "response to favourite species":

This is one example:

GO:0009615    response to virus	113,781 annotations
GO:0051607    defense response to virus	103,512 annotations
GO:0098586    cellular response to virus	4,974 annotations
GO:0050688    regulation of defense response to virus	6,015 annotations
GO:0050687    negative regulation of defense response to virus	1,996 annotations
GO:0050690    regulation of defense response to virus by virus	870,719 annotations
GO:0050691    regulation of defense response to virus by host	2,111 annotations
GO:0002230    positive regulation of defense response to virus by host	1,750 annotations
GO:0050689    negative regulation of defense response to virus by host
etc.....

This is not remotely sustainable.

2. Annotation would be more consistent, and we would always get the biologically relevant term in addition to the pathogen.

3. It would/should be easy to configure in tools, they would need to make a taxon ID available in this slot.

4 We could get rid of the dual taxon notation that is very hidden and not mandated

@thomaspd
@cmungall

[hattrill]

For the experiments we encounter, the broad-brush categories really capture the experiment at the level they are intended e.g gram-negative bacterial, virus.

If groups want to capture interactions right-down to the species level, then I am ok with taxon ids in extensions.

I would prefer have something in broad classes - classification based on some attributes, such as gram negative, may not be necessarily aligned with taxonic lineage. I am not so sure that there is an explosion, if these are restricted to the broader classes. Could do with a good tidy tho' to harmonize on meaning - i.e. do we need separate defense response and innate response?

[ValWood]

broad classes would probably be sufficient (better actually because options could be provided rather than everyone needing to look up taxon IDs which is tedious).

do we need defense response and innate

unfortunately yes. there are defenses that are not immune response related

[addiehl]

+1 there are defenses that are not immune response related

[pgaudet]

We discussed this on the multiorganism call - and @hattrill if you are OK with taxon ids in extensions, then can you use GO:0009625 response to insect and add the parasitoid wasp taxid ?

In this case we would not create the term.

Thanks, Pascale

[hattrill]

What relation would you suggest?
Unfortunately, there are no tools which allow a search on taxid in extensions so it will get lost.

[pgaudet]

You think searching the taxon is needed ? I was hoping 'insect' would be granular enough as a search term, and users could view the exact taxon.

[hattrill]

They are more likely to look for parasite or parasitic insect, but we don't have a specific parasite branch.

(but I think that an insects for insects is not intuitive - certainly, when I spoke to Nick Brown he thought that parasitic insect was essential for users to find the annotations)

[pgaudet]

I see. Are parasitic insects always parasitic ? We have the same issue with pathogen.

[pgaudet]

Talking with @vanaukenk - Kimberly proposes to use 'has input'; would that work ?

Also, if @mgiglio99 @dsiegele @ValWood agree I could create 'defense response to parasitic insect', as a child of 'defense response to insect'.

Finally - if this is a response to infection, shouldn't there be some link to an immune response ?

[hattrill]

I am not an expert on parasitic wasps, but the laying of eggs in host/presence of larvae is always parasitic act.

I have been quite conservative for this particular example and some of the processes involved in the immune response e.g. melanization and pathways can be described with co-annotation or used in an extension in some way.

I think if there was a link to immune response, it would not be incorrect as it is a response to an infection.

[hattrill]

has_input could work.

[mgiglio99]

I don't know much about parasitic insects, but I do fear that we will have the same issues with 'parasitic' as we had with 'pathogenic'. To me, the fact that the term is 'defense response to...' and not just 'response to...' already implies that the insect is doing harm and is therefore parasitic or pathogenic.
Would it be an option to make 'defense response to parasitic insect' a narrow synonym to 'defense response to insect'? That would solve the searching issue.

[ValWood]

I think this sounds good.

I still have major concerns about 'defense response to blah' , because it leads to very inconsistent annotation, but this is an ongoing issue that will need to be addressed at some point in the future.

[hattrill]

The current definition is very broad: A response to protect an organism from a directly detected or perceived external threat from an insect or insects to that organism.

What we seek here is to differentiate the response to physical injury (i.e. injection - which we can capture using current terms) and a distinctly different response - the response triggered by the internal presence of eggs/larvae. Which is what we proposed in the definition at the top.

[ValWood]

Isn't a 'response to presence of eggs larvae etc' always a type of 'immune response? or can it be some non immune system process? is the process 'detection' for example? or is the process totally not known?

The major problem that keeps arising is that people will use the "response to" and not the terms under
"GO:0051702 interaction with symbiont" which should be used to describe the specific interaction with pathogen/symbiont.

I still think it would be useful to couple these terms and only allow "response" to as an extension

GO:0051702 interaction with symbiont some_relation. "in response to blah"

This might solve your problem @hattrill because it implies some molecular level interaction, not just getting away from (although I agree the definition could also be tightened, I guess it was written this way because some types of movement are allowed here- but I'm not sure what they are).

[hattrill]

Well, could make the assumption that it is part of the immune response and would probably not be wrong.

'symbiont' is not the wording that parasitologists would use. I'd rather have a synonym under 'defense response to insect'

(I am not around for a week, so no more posting from me for a while)

[ValWood]

'symbiont' is not the wording that parasitologists would use.
the definition encompasses parasitism though... but I agree this causes endless problems and results in curators ignoring these terms...

[hattrill]

Following this discussion: using taxon ids in extensions with has_input to specify species:
@alexsign would it be possible to add NCBI:txid to the has_input field in Protein2GO?

[vanaukenk]

Note a parallel discussion here: geneontology/go-shapes#237
I've put these two tickets on the agenda for the 2020-10-19 ontology editors call.

[pgaudet]

On the ontology editor call we agreed that 'has input' is the correct relationship for this.

[ValWood]

I think I misunderstood what this ticket was about. So let me get this straight.
In GO traditional annotation the interacting taxon goes (for any pathogen-host interaction term) in the taxon field with a separator
In GO-Cam the interacting the taxon goes in an extension field using has_input (or is this just in the case o the "response to species" terms), which would be very confusing.

Questions
So do you plan that people no longer use the taxon field for this information longer term?
Or is the plan that we redundantly populate both fields in a GAF file ?
And where does this fit in GPAD (I presume GPAD has a taxon field too?)

[pgaudet]

So do you plan that people no longer use the taxon field for this information longer term?

It's a difference of format between GAF and GPAD.

Or is the plan that we redundantly populate both fields in a GAF file ?

I dont think so - I think the interacting taxon stays in the taxon field. There is no plan to further change the GAF specs.

And where does this fit in GPAD (I presume GPAD has a taxon field too?)

Yes, according to the latest specs: https://github.com/geneontology/go-annotation/blob/master/specs/gpad-gpi-1_2.md

@vanaukenk can confirm.

Can this be closed ?

@hattrill This is now a P2GO issue ?

[ValWood]

My questions are answered.

[pgaudet]

I created a separate annotation ticket for the use of 'has input' + the species in an extension,
geneontology/go-annotation#3514

As far as the ontology goes, this is resolved.

Thanks, Pascale

[mgiglio99]

Hi - I looked at the other issue as well as this one. I just want to make sure I understand the resolution here - the new term requested above was not made.
And the wasp situation will be handled with use of taxon id extensions.
Correct?

[pgaudet]

Yes @mgiglio99 this is right.

We suggested to @hattrill to use GO:0002213 defense response to insect.

Thanks, Pascale