Does GO have a term for pathogen?

[CuzickA]

I am trying to make logical definitions for the 3 phipo namespace terms

Pathogen phenotype = phenotype inheres in pathogen (???)

Host phenotype = phenotype inheres in host (http://purl.obolibrary.org/obo/GO_0018995)

Pathogen host phenotype = phenotype inheres in symbiont process (synonym is host-pathogen interaction) (http://purl.obolibrary.org/obo/GO_0044403)

I can't find a GO term for pathogen.

Next to think about is an external ontology term for 'phenotype' or to create as a phipo class (similiar to fypo). Would this be a parent to the above 3 namespaces? Would this cause a problem with loading the 3 namespaces into PHI-Canto? What namespace would 'phenotype' have?

[mah11]

Hmm. We'd better check whether an ontology term can have a different namespace than its superclass (is_a parent). I don't think I've run into any so far, and it may be disallowed.

It's very much an awkward hack that GO has a term for "host". Given the way the "host" term is used in GO, I'm not sure it's an exact match for what you want to do in PHIPO - it's a subtype of "other organism", and for PHIPO host phenotypes will at least sometimes be due to host mutations. To me it doesn't seem to make sense to call the host "other" in that scenario.

There isn't a term for "pathogen", and that's unlikely to change because (it would be even hackier because) it brings in a lot of information about the effect the pathogen has on the host, which I think GO would regard as closer to the BP ontology scope than CC.

In FYPO the root "phenotype" term has the fission_yeast_phenotype namespace, but that's trivial because it's the only namespace in the ontology. The other phenotype ontologies that I've looked at closely enough to tell also have one namespace each, so PHIPO is a bit exceptional in having more than one. That means I don't know whether there's an existing SOP.

Sorry I don't have quick simple answers!

[ValWood]

Maybe it isn't necessary to have these high level rrot node logically defined as long as definitions are plugged in for the terms that are used for in annotation (these high level terms are really only for completion/ontlogical correctness).

@mah11 do you have terms right to the root logically defined? If so where would be a good place to define "pathogen" and "host". I thought they would be required by PATO maybe?

[mah11]

@mah11 do you have terms right to the root logically defined?

What do you mean by "terms right to the root"?

The FYPO root itself does not have a logical definition. For all other terms, I include a logical definition if I can figure out how; length of path to root is not a criterion.

If so where would be a good place to define "pathogen" and "host". I thought they would be required by PATO maybe?

I don't see how they would fit into PATO - they're roles organisms can play relative to one another, not qualities of anything.

I wonder if Ramona Walls (@ramonawalls) might be able to help. She works on PCO, "Population and Community Ontology", and I found this on its tracker: https://github.com/PopulationAndCommunityOntology/pco/issues/4

[ValWood]

Sorry I didn't explain v. well. But yes that's basically what I meant.

I just wondered it it was critical that the high-level terms needed a logical def, but if not I guess @CuzickA can leave these until a natural way to define them materializes, and concentrate on the ones with more obvious design patterns.

I thought PATO also included things that are not qualiites but are required to describe phenotypes, but clearly not.

[CuzickA]

I though I would try and make logical definitions for the 3 phipo namespace terms as I will also need to define the 'pathogen', 'host' and 'pathogen host interaction' component of the binding and localization phenotypes that I have triplicated and used from FYPO #21.

eg under the host phenotype namespace for 'host binding phenotype'

I have taken this logical definition from FYPO but it is lacking the 'host term'.

The sub-class terms also eg 'decreased host DNA binding'

[jseager7]

For the time being we've decided to use the 'pathogen' term from the Infectious Disease Ontology (IDO:0000528), just so we're able to create logical definitions for chemical sensitivity and resistance.

The term seems sufficiently general that it shouldn't cause major semantic issues, and the term was in use by many other ontologies, which you can see here (scroll down to the IDO definition).