Logical axioms for non-human neoplasms

[diatomsRcool]

It appears that the current cross-species-analog relationship is based on similarities between labels. We would like to propose that the current cross-species-analog relationship be removed and replaced after adding logical axioms to the non-human diseases. We would like to propose that the data about causal genotypes for non-human cancers that are contained in OMIA be added to mondo. After these data are ingested, we would like to use semantic similarity analysis to add the cross-species-analog relationship.
An example of how this information looks in OMIA can be found here. Here a BRAF variant is known to cause a type of bladder cancer. We would be open to adding more information if found.
The vets I've been working with have been telling me that there are some known cancers in humans and canines that are analogous, but manifest themselves in different parts of the body. Thus, the label matching is not going to work all the time. A computational approach could reveal new relationships.

[sabrinatoro]

I think that we might need to look at 2 different relations.
1st relation: the disease, as defined by its phenotype/symptoms, is "the same" even though the cause is different.
Non-human leukemias are a good example for this. Taking cat as an example, in most cases, leukemia is due to virus infection, BUT the actual disease (the abnormal proliferation of leukocytes) is the same as human leukemia.
2nd relation: the cause is the same (e.g variation in a gene) but the phenotypes/symptoms are different
See the BRAF example above.

Questions:

1. is this a correct representation of the situation? (@katiermullen, could you please comment on this)
2. I don't think the orthologous relation between genes belongs in Mondo.
   For example:

• feline leukemia is analogous to human leukemia (because leukemia is defined as an abnormal proliferation of leukocytes).
  For example 2:
• BRAF-bladder cancer in dog can be associated to BRAF
• BRAF-colorectal cancer in human can be associated to BRAF
• BUT there shouldn't be a relation between BRAF cancer in dog and in human (in Mondo). This relation should be in the annotations, outside of Mondo.

Maybe we need to have a meeting with experts to discuss this

[katiermullen]

Feline leukemia defined as the presence of neoplastic cells of hematopoietic origin in blood and/or bone marrow can be FELV (viral) or non-viral associated.

• FELV infection does not always result in the cat developing leukemia, but it can.
• FELV infection more commonly results in the development of lymphoma than leukemia.
• Non-FELV associated leukemia has been reported in cats.
• FELV infection can result in non-cancerous sequela (anemia, immunosuppression, etc.).

[nicolevasilevsky]

@diatomsRcool maybe it would be helpful to have you attend a future Mondo curation call to discuss this further? We'd need @sabrinatoro @cmungall @katiermullen

[diatomsRcool]

I think that feline leukemia is a complicated example that needs to be discussed further. Maybe let's stick to the canine cancers for this one? Pick the low hanging fruit? FeLeuk is an example of something that we prob need to work through in a workshop setting with more discussion and several subject matter experts.

[diatomsRcool]

Yes, I think I need to attend the next mondo call.....

[nicolevasilevsky]

@diatomsRcool is this still needed? Do you want to attend a future Mondo curation call?

[diatomsRcool]

I have been attending mondo calls that focus on animal diseases. I think this issue can be closed. When we figure this out better I can create another, more specific issue.

[sabrinatoro]

@nicolevasilevsky, please let me take care of this issue. I put it for August so I take a look at what is needed and figure out where it fits in the non-human animal disease work

[nicolevasilevsky]

ok great, thanks @sabrinatoro and @diatomsRcool