Create a clearer distinction between concrete diseases and groups of diseases

[matentzn]

From @pnrobinson

In general, I think that the definitions for groups of diseases in Mondo to be made more clear that they are groups of specific diseases and not themselves a specific disease. For instance

Fanconi renotubular syndrome
A genetic or acquired disorder characterized by impairment of the function of the proximal tubules of the kidney. It results in decreased reabsorption of electrolytes, glucose, amino acids, and other nutrients.

This should make it more clear that it is a grouping term. By saying "A disorder" this makes it clear only to ontologists, but many people could think it is one disease (clinically these are very distinct)

[cmungall]

I thought we had an issue about this already. We definitely need to distinguish between disease entities and groupins

We discussed it as part of:

• 6593.

And we even include "disease entity" in the FAQ but don't actually define that.

We had a lot of discussion on this, on whether eg Parkinson was one entity or many, and we had links to the literature about this.. can't find any of it... did it get lost in the slack memory hole?

[pnrobinson]

There is no question that Parkinson is multiple entities. But this is what I mean by the need for documentation: We need to define how to cut the cheese. Do we divide solely on clinical basis like OPRHA, do we divide on gene basis like OMIM or do we continue subtyping and on what basis. I think we can provide reasonable answers like this. Having an mkdocs site that we could make PRs to to add documentation would save a lot of churn, like this maybe: http://obophenotype.github.io/human-phenotype-ontology/annotations/genes_to_phenotype/

[matentzn]

I have most of our discussions in my memory somewhere, but we never defined a clear strategy.

There is a lot of stuff documented in the Mondo curation docs about grouping classes, and the tech docs.

Also you mentioned: https://github.com/cmungall/multi-axial-ontology-test

Here is my suggestion to make progress on this:

1. Create a google docs for developing a strategy
2. Develop a meaningful definition of grouping vs real disease conceptually with Peter and other medical experts.
3. Create a library of examples (maybe 10, 20) for illustration.
4. Define patterns that can recognise diseases that folllow the definition of grouping defined above.