descends_from relationship in VBO

[katiermullen]

Ongoing discussions of adding a descends_from relationship in VBO are summarized here with examples:

Working definition for “descends from”: Name:

• descends from (there might be a better name, especially since “descends from” already exist.
• is descendant of
• results from crossing involving
• Has ancestor
• Originate from
• “Contribute to breed history”
• A derivative of

Definition: A relation between two distinct material entities (breeds or species), a descendant entity and an ancestor entity, in which the descendant entity is the result of mating, manipulation, or geographical or cultural isolation of the ancestor entity, therefore inheriting some of the ancestor’s genetic material.

Examples:

• Breed -has ancestor-> breed (N>=1)
• Breed -has ancestor-> wild animal (can we use NCBITaxon entity?)

Specific examples:

• Breeds where there is only one ancestor breed that is combined with a mixed breed e.g Havana brown (cat) has ancestry breed Siamese (cat) and domestic black cat. https://en.wikipedia.org/wiki/Havana_Brown
• Breeds where there is only one ancestor: Cymric (cat) has ancestry breed Manx (cat). https://en.wikipedia.org/wiki/Cymric_cat
• Breeds where there is an ancestor from a wild (non-domesticated) species: Chausie (cat) has ancestry wild animal species jungle cat (Felis chaus). NCBITaxon:61376 https://en.wikipedia.org/wiki/Chausie
• Breeds that have ancestry breeds of two domestic species: Brangus (cattle) has ancestry breed Angus (cattle) Bos taurus x Brahman (cattle) Bos indicus = Bos indicus x Bos taurus hybrid NCBITaxon:30522 https://breeds.okstate.edu/cattle/brangus-cattle.html

Plan:

• [ ] Gather feedback from @ImkeTammen and @franknic on the above definition.
• [ ] Create relationship for use-case in cats (since we have this information all ready to go).
• [ ] @rando2 and students to use LLM to pull descends_from relationship from wikipedia or other sources for dog and/or cat breeds.
• [ ] Eventually compare to genetic relationships to see if "ontology recapitulates phylogeny."

[franknic]

Thank you very much, Katie. Imke and I will discuss this tomorrow.

[franknic]

Back again. Imke and I worked through this today. We are very grateful for all the effort that has gone into this, and which is planned for the future. Our feedback is as follows. Name After considering all the options, we offer “contributions from” as a better option. It has the advantage of avoiding the implication of a linear relationship (which is the reality in only a subset of breeds); and it also embraces the reality that some breeds are not closed populations, but continue to receive contributions from other populations, sometimes regularly and sometimes intermittently. Definition We appreciate all the effort that has gone into the definition, which does a good job conceptually. Our only question concerns the use of the words “ancestor” and “descendant”, which imply a linear relationship that does not always exist in practice. Also, if we understand correctly, one of the entities is always a breed in VBO and the other entity could be a VBO breed or an identifiable population or a species. Is it feasible to consider revising the definition to incorporate these thoughts? Examples Following from our thoughts above, it is probably better to avoid thinking in terms of ancestor, and, instead, to think in terms of “contributions from” one or more populations. In answer to the question: if the contribution is from a wild animal, then we agreed that it is fine to use the NCBI taxon entity. Specific examples Each of these is fine, if we think in terms of “contributions from”. Plan We are happy with the plan ++++++++++++++ Imke: if I have missed anything or have got anything wrong, please add to this feedback.

[katiermullen]

Thank you for your thoughts. This is extremely helpful and makes a lot of sense.

@sabrinatoro do you have any comments on Frank and Imke's feedback?

[sabrinatoro]

Thank you @franknic and @ImkeTammen. "contributes to" or "contribution for" (or similar words) are already in use to mean something else. So we will have to do this review and see if this word can work.

I would like to understand better the following statement (I lack the basic knowledge): "The reality is that some breeds are not closed populations but continue to receive contributions from other populations, sometimes regularly and sometimes intermittently." What does it mean? Does it mean that, for example, a breed population is maintained by crossing animals with animals from another breed population? If it is happening constantly, is a breed defined with the assumption that its characteristics are maintained because of this constant crossing?

[franknic]

Thank you, Katie and Sabrina. That's a good, questions, Sabrina! A cattle example is provided by the Hereford (VBO0000232) and the Polled Hereford (VBO0000341, both of which are transboundary breeds. In some countries, such as Australia, these two breeds used to be run as distinct breeds, each with its own stud book. But some time ago in Australia and in some other countries, such as the USA, the two breeds combined their stud books. The two breeds are still recognised as such, but it is very likely that any Polled Hereford in Australia or the USA has contributions from the Hereford breed, and vice versa. I suspect the same applies to the Cymric cat breed in relation to the Manx breed, mentioned as an example in a previous post. Incidentally, when I looked for Hereford and Polled Hereford in VBO, they are returned as expected in searches, but neither transboundary is included in the list under the heading "Cattle breed (3656)". Similarly, I can't see the transboundaries Polled Shorthorn (VBO000342) or Polled Sussex (VBO000343) in the list. In fact, there seems to be many transboundary cattle breeds missing from the list, e.g. Aberdeen-Angus (VBO0000092) and Abondance (VBO0000091), to choose two at the top of the alphabet. Maybe I am misunderstanding the list.

[katiermullen]

@franknic in regards to your search for terms under 'Cattle breed' are you using Protege or OLS to look for these breeds as a subclass of 'Cattle breed' VBO:0400020?

If you are using OLS, I can re-create this issue. I think that perhaps OLS does not list all 3,656 breeds of cattle under 'Cattle breed.' When I search for 'Abondance (Cattle)' VBO:0000091, for example, this breed is a child of 'Cattle breed.' I noticed the same thing would happen if you browse cattle breeds under 'Bos taurus' - not all cattle breeds are displayed.

In Protege, all cattle breeds appear as a subclass of 'Cattle breed' VBO:0400020 and 'Bos taurus' NCBITaxon:9913 as expected.

@sabrinatoro - do you have any further comments on how VBO is displayed in OLS?

[franknic]

Thank you, Katie. I'm using EBI's OLS. And yes, you can certainly recreate it: I've included a screenshot of the top of the 'Cattle breed' list, taken just now, showing neither the transboundaries Aberdeen-Angus or Abondance. And yes, if you search for, e.g. Abondance, then it is definitely in VBO as a child of Cattle breed. If there is anything we can do to help solve this mystery, please let us know. Also, I was interested in your suggestion for browsing cattle breeds under 'Bos taurus'. How do we do this in OLS? Screenshot:

[sabrinatoro]

Thank you for the answer, Frank!

The issue you are reporting is weird. The good news is that the ontology is OK: I can see the breeds you mentioned in the "cattle breeds". For example:

However, these breeds are not there when looking at "all" the cattle breeds (ie when opening the "cattle breeds" branch and browsing, as shown here:

This is a UI problem in OLS (maybe due to the fact that there are a lot of terms under "cattle breeds" and they are omitting some?), and we should create an issue on the OLS GitHub tracker.

[katiermullen]

After discussion we think that we need two separate relations - one for ancestry breeds and one for has contributions from. This issue will continue to address the ancestry breeds and we will create a new issue for contributions from.

"Has foundation stock" add in the comment what the definition of foundation stock is:
refers to animals that are the progenitors, or foundation, of a breed or of a given bloodline within such. Many modern breeds can be traced to specific, named foundation animals, but a group of animals may be referred to collectively as foundation bloodstock when one distinct population (including both landrace breeds or a group of animals linked to a deliberate and specific selective breeding program) provides part of the underlying genetic base for a new distinct population.

https://en.wikipedia.org/wiki/Foundation_stock.

See usage here: https://en.wikipedia.org/wiki/Standardbred

Note: Some breeds with an established phenotype and named foundation stock may still permit outside bloodlines, usually from the foundation breeds from which they came, to contribute further to the genetic base of the breed; these thus have at least partially open stud books. An example of a partially open stud book is that of the American Quarter Horse, which still allows limited registration of animals with one Quarter Horse parent and one Thoroughbred parent. Newer breeds, such as many of the warmblood breeds, may have mostly open stud books, where horses that are registered may be of a variety of bloodlines, but must first pass a rigorous studbook selection process.

https://en.wikipedia.org/wiki/Foundation_stock

This implies that "contributes to" may be a child of "has foundation stock" (in our future work).

[katiermullen]

@matentzn @sabrinatoro

Example for initial testing:

1. Cymric (cat) VBO:0100080 has foundation stock Manx (cat) VBO:0100156
2. Chausie (cat) VBO:0100068 has foundation stock jungle cat (Felis chaus) NCBITaxon:61376
3. Brangus (cattle) VBO:0000161 has foundation stock Angus (cattle) VBO:0000104 and has foundation stock Brahman (cattle) VBO:0000159

[franknic]

Thank you, Katie and Sabrina Imke and I will discuss this on Thursday, in preparation for our meeting with you on Friday morning (Sydney time). In the meantime, it is worth mentioning that Mason's dictionary has a wealth of information about foundation stock. For example, his Brangus entry is:

Brangus: (USA: Texas and Oklahoma) / m. [meat] / black; pd [polled] / orig. 1932 at Jeanerette, Louisiana, from Brahman (c. 3/8) × Aberdeen-Angus (c. 5/8) / var.: Red Brangus / BS [Breed Society] 1949; BS also New Zealand, Argentina, S Africa 1986 / see also Australian Brangus.

The words in square brackets are explanations of the abbreviations used by Mason. Looking forward to our conversation.