distinction between diseases and phenotypes

[nicolevasilevsky]

This is an example use case

The sub-classification of MONDO_0010850 Tessier number 4 facial cleft as a subclass of MONDO_0015824 Oculomaxillofacial dysostosis seems incorrect (although I can't identify the cause - the xref of orphanet_141258 to OMIM_600251 seems okay.

What action should I take here? Should I just remove the parent class assertion in Protege and make a pull request?

Thanks!

[kanems]

Is this term equivalent to https://hpo.jax.org/app/browse/term/HP:0031576 ? Based on the names and cross-references, I'm not sure how to map this concept for MedGen.

UMLS has different identifiers for the HPO term and the OMIM presentation, and the names/synonyms used by the various groups seems to be variably interchangeable...
The OMIM number is on UMLS C1838348, "oculomaxillofacial dysostosis" which is the name on the parent term in MONDO->MONDO:0015824. OMIM treats "tessier type 4" as a phenotype (in line with HPO) and describes 4 affected patients from a Richieri-Costa and Gorlin 1994 paper under "clinical variability" on the "OCULOMAXILLOFACIAL DYSPLASIA WITH OBLIQUE FACIAL CLEFTS" entry (600251)

For additional X-refs: GARD 4046 treats Oculomaxillofacial dystosis as equivalent to : " Richieri Costa Gorlin syndrome; Oculomaxillofacial dysplasia with oblique facial clefts; Oblique facial clefts" https://rarediseases.info.nih.gov/diseases/4046/index

So perhaps the name "Tessier number 4..." should not be the preferred name on this term, at a minimum, to clarify the clinical presentation as a whole is not equivalent to the precise presentation in the HPO term?

@pnrobinson @ahamosh @ana-rath-orphanet

[pnrobinson]

This again gets to the distinction between diseases and phenotypes. If a person has a tessier type 4 cleft as their only clinical problem, it is possible to treat it as a disease. On the other had, various kinds of clefts can be observed as features of other diseases.
In general, I would propose that a disease consists of a clincial course, response to treatment, and an etiology. This distinguishes it from a phenotypic feature. So, a good MONDO term might be "Idiopathic Tessier type 4 cleft" or "Isolated Tessier type 4 cleft". The corresponding HPO terms might include Tessier number 4 facial cleft HP:0031576 and perhaps terms related to common associated findings such as difficulty with feedings, difficulty swallowing, nasal voice and recurrent otitis media.
@cmungall @mellybelly @ahamosh I think it is really important not to make ad hoc decisions about these issues but to have a comprehensive written statement about MONDO's philosophy. We really need to have some workshops!

[nicolevasilevsky]

see also: #86

[nicolevasilevsky]

see also #559

[nicolevasilevsky]

another example: #501

[ahamosh]

These are all excellent examples of the problem with phenotypic feature (which belongs in HPO) and a diagnosis (which belongs in Mondo). The problem becomes even more complicated when a diagnosis is a feature of a disease (e.g. diabetes as part of MELAS (mitochondrial encephalopathy lactic acidosis and stroke). The Discitis example is weird: it is an inflammation of the disc which could be a feature or a diagnosis, especially a diagnosis if the etiology is infection. I think that all of this is/will be very hard. The Tessier type 4 cleft is the description of the cleft. It is definitely a feature and probably of the syndrome, but it sounds like in some people the cleft is the extent of the syndrome...

[ana-rath-orphanet]

Hello,
We are trying to add "isolated" to a morphological abnormality when it is the disorder (not complete yet), and all the disorder. We then attribute a classification_level=disorder and a typology:morphological anomaly. Having a disease code (i.e. Orpha) for those should help distinguishing them from phenotypic abnormalities that have an HPO code. Plenty of examples: you can download and list them by filtering by morphologuical anomaly out of the Orphanet nomenclature @annieolry

[ahamosh]

That is a good idea, but we will have a lot of overlap.

[pnrobinson]

I agree this is a good idea -- I do not think it makes more overlap, instead it should help to disambiguate the existing overlap.

[nicolevasilevsky]

#822