search

PopulationAndCommunityOntology / pco

An ontology about groups of interacting organisms such as populations and communities
Creative Commons Zero v1.0 Universal
16 stars 4 forks source link

NTR: community and ecological community #32

Closed ramonawalls closed 4 years ago

ramonawalls commented 8 years ago

(requested by Barry via email) def = a collection of organisms connected by social or biological relations

with a comment: these relations can include shared values, occupying the same spatial relation

We will need to add some kind of axioms for the social and biological relations. Probably can get the biological relation from RO.

This will be a subclass of collection of organisms.

ramonawalls commented 8 years ago

Added term http://purl.obolibrary.org/obo/PCO_0000028, but needs some work.

ramonawalls commented 8 years ago

@pbuttigieg - copying the email thread here. rwalls@iplantcollaborative.org wrote:

I added the term http://purl.obolibrary.org/obo/PCO_0000028, but it and the rest of the hierarchy will need some more work.

Is there an GitHub issue we can work with to track this? Would be best.

[...]

Now that I am working on this, I see that you were perhaps suggesting a new subclass for collection of organisms, which is called community. It took me working with the ontology to get your message properly. Yes, I think that is a fine addition. The label will surely generate some controversy, but I can live with that.

As it should - the label is used very liberally and the definition can be refined once we have some subclasses to work with. I recall that there was quite a lot of debate over rights allocated to an 'indigenous community' vs a 'community of landowners'. The nature of the relations between community members (economic, political, cultural, genealogical, etc.) will be key in resolving these classes sensibly. First, however, we need to populate the hierarchy to have something to work with.

collection of organisms

  • community -- ecological community -- pair of interacting organisms -- collection of humans -- family -- etc.
  • single-species collection of organisms -- collection of humans (inferred) -- family (inferred) -- etc.

Seems good. I think the class 'ecological population' should be asserted under 'single-species collection of organisms'. Also, is a 'pair of interacting organisms' a community? I think that should be made a subclass of 'collection of organisms' for now. It is perhaps the basic unit of any given community or population.

collection of organisms

I guess the family definition is anthropocentric, which makes sense for SDGIO, but perhaps we should label it 'human family'. I also would say 'human economic community' as Barry's definition may suggest that the livestock involved in a pastoral operation would be part of the community linked by economic relations.

[...]

I think we can do that, given the scope of the ontology is not intended to include random collections of organisms. To complete the logical definition of the term, we would then need to define a social or biological relation. We can probably use something from RO for the biological relation, but will need something from the social ontology workshop you have coming up in the winter for the social relation.

Those social relations will be very important. As mentioned above, we can use these to infer a meaningful hierarchy from the incoming mass of entities identified as communities. Would these social relations also live in RO? I think this would be good if subsets of RO are generated for different communities.

I would still steer away from using community as the primary label, and keep it as a synonym.

Pier, does this sit with you?

I think we can use 'community' as the primary label, and make sure 'ecological community' is defined clearly. PCO could then export subsets as part of the release process for ecologists, social scientists, etc.

[...]

Barry Smith phismith@buffalo.edu wrote:

perhaps we can be a bit more ambitious, defining

community = a collection of organisms connected by social or biological relations

I would say ecological interactions rather than or in addition to biological interactions. This would also be more in line with SEEA thinking. I agree that we should be very general, but is there a way to add a temporal aspect? These interactions must hold for some minimum window of time to be meaningful: what does it take to become part of any given community?

[...]

rwalls@iplantcollaborative.org> wrote: [...]

"ecological community". Ecological community still needs a logical definition, which I will add. I agree that there could be a more general term for multi-species collection of organisms, but I would not call it community, as that is too vague of a name. Yes, and when we're dealing with ecological communities the assumption is that the organisms are interacting or able to interact in some shared ecosystem. That is, the ecosystem increases the chances of a disposition of organisms from multiple species to interact within their lifetimes. The magnitude of this probability of interaction required for some 'multi-species collection of organisms' to be called a community is probably up for debate.

Perhaps "single-species collection of organisms" would be a better name than "collection of organisms of the same species". It is at least a little shorter. Agreed. Just waiting for the discussions around species to explode.

"Collection of organisms of the same species" has "collection of humans" as a subclass. I can add human population as a synonym, and human community as either a synonym or a subclass, depending on how it is defined.

As touched on above, I think we should be prepared to handling thresholding issues here: collection of organisms with 'strong enough' relations? 'close enough' spatial proximity, etc. This can come later, but I think it will come.

Best, Pier

pbuttigieg commented 8 years ago

Thanks for transferring!

ramonawalls commented 8 years ago

Updated ecological community (PCO:0000002) with logical axiom that says is must have member min 2 populuation of organisms.

ramonawalls commented 8 years ago

I'm reading over this again, and I'm concerned that what is proposed for "community" (A collection of organisms connected by social or biological relations.) is not disjoint from "ecological community" (A collection of organisms of at least two different species, living in a particular area. Must have at least two populations or organisms as members.). Non-disjointedness is okay, but I don't want to confuse users by having multiple axes of classification that partially overlap.

The differences as currently defined are: 1) community can be single or multi-species ecological community is only multi-species 2) community has members that interact socially or biologically ecological communities just need to live in the same area

Also, social interactions are biological (biotic) interactions. Not only are the participants organisms, but they are moderated by other biological or biochemical entities such as hormones.

As mentioned somewhere in this thread or another, the PCO is not concerned with random collections of organisms, but rather on collections of interacting organisms. We already have very general classes for collections of organisms that don't include interactions, such as collection of organisms (POC:0000000), etc., and I don't want to get rid of those at this point, but it might be better to start building our hierarchy based on interacting organisms.

ramonawalls commented 8 years ago

I propose the following hierarchies:

1) Any collections of organisms collection of organisms

single species collection of organisms multi-species collection of organisms

2) Collections of interacting organisms community (=collection of organisms that are interacting)

ecological community (=community with two or more populations, all living in the same space) population (=single specie collection of organisms all living in the same space)

The first hierarchy is mostly a convenience, and something we already have. The second hierarchy is what is generally of interest to scientists. We should be able to use RO:biotically interacts with to define this hierarchy.

If we get the logical definitions right, there will be some overlap among the hierarchies, so I am a little concerned this may lead to confusion among users. We may therefore ultimately end up just deprecating those original terms that don't include any interactions.

I am going to put these classes into PCO now. They will go into the next release, and we can get some feedback on whether or not they work.

ramonawalls commented 8 years ago

@cmungall, @pbuttigieg Can I use RO:colocalizes with to talk about organisms that live in the same place? Seems generally correct. Should we make a subpropety called "cohabitates"?

cmungall commented 8 years ago

The intent is that localizes is a bit more specific and to do with active localization.

I think cohabitates is a good relation to add - but not as a subproperty. We could later add a grouping relation, something like co-located-with, if useful.

Do you want an RO ID range? If we sync on Protege 5 then we should be safe for making pull requests (or I can just add it...)

On 11 Jun 2016, at 15:12, Ramona Walls wrote:

@cmungall, @pbuttigieg Can I use RO:colocalizes with to talk about organisms that live in the same place? Seems generally correct? Should we make a subpropety called "cohabitates"?

You are receiving this because you were mentioned. Reply to this email directly or view it on GitHub: https://github.com/PopulationAndCommunityOntology/pco/issues/32#issuecomment-225397355

pbuttigieg commented 8 years ago

I think cohabitates is a good relation to add - but not as a subproperty. We could later add a grouping relation, something like co-located-with, if useful.

This may be a bit strong - the organisms may not have overlapping habitats, that would probably be a specification of co-located-with.

Do you want an RO ID range? If we sync on Protege 5 then we should be safe for making pull requests (or I can just add it...)

I think we should sync.

ramonawalls commented 8 years ago

I have requested a relation for cohabitates with at oborel/obo-relations#111, but I agree with Pier that cohabitates with is wrong for defining communities. For those, I would prefer to use RO:biotically interacts with, which is just as strong, but has a different meaning.

ramonawalls commented 8 years ago

I was able to define an ecological community as a community that occupies a site, but have to add the new occupies relation to RO.

Until we have a way to specify that all members of a community participate in the same biotic interaction process, community will be equivalent to collection of organisms. I'll work on that next.

ramonawalls commented 8 years ago

Note I used "located in" for now, rather than create "occupies" at this time.

ramonawalls commented 5 years ago

Given that we don't have a convenient way to describe interactions yet, and we don't have an actual use case for the term community, I am going to deprecate community and just keep ecological community.

ramonawalls commented 5 years ago

See BiodiversityOntologies/bco#86 for explanation of why "located in some site" is wrong for the definition of ecological community. I am going to remove that axiom. The logical definition will be underspecified, just as for population of organisms.

cmungall commented 5 years ago

Pedantically speaking, a logical definition is either fully specified or it is wrong. If you can't state N+S conditions accurately better to just state N conditions.

Also, a logical def that deviates from the text def is often a bad smell; see https://douroucouli.wordpress.com/2019/07/08/ontotip-write-simple-concise-clear-operational-textual-definitions/

Are you sure you don't want to switch your equivalentClass to subClassOf?

Right now you have:

ecological community = multi-species collection of organisms and has-member min 2 populations of organisms

which is weaker than the text def: A multi-species collection of organisms of at least two different species, living in a particular area. Must have at least two populations of different species as members.

This isn't just fussiness - you'll end up with confounded inferences further down the line, and I don't think having an equivalence axiom does you any favors here.

ramonawalls commented 5 years ago

Thanks, @cmungall. That makes perfect sense. Let me try it with subclass axioms instead.