search

EBISPOT / efo

Github repo for the Experimental Factor Ontology (EFO)
https://www.ebi.ac.uk/efo/
56 stars 13 forks source link

Add terms from WC_20190325_disease #415

Closed twhetzel closed 5 years ago

twhetzel commented 5 years ago

Add terms from WC_20190325_disease [1].

[1] https://docs.google.com/spreadsheets/d/1Ss5ssjt8Ci1Df8__r__GjyRPpBqQN4bj4VbDigi1Btk/edit#gid=2091625908

twhetzel commented 5 years ago

@paolaroncaglia should "Modic type vertebral endplate changes", which is a term in HP, be added to EFO with an EFO IRI and xref to HP? The suggested parent in EFO is "lumbar disc degeneration", which is a disease term.

twhetzel commented 5 years ago

I found an example of this case, e.g. acute pancreatitis, and will leave the term with the EFO URI and add the xref to HP.

paolaroncaglia commented 5 years ago

@twhetzel This goes back to the broader issue of disease versus phenotype we’ve come across recently (see e.g. https://github.com/EBISPOT/efo/issues/261). In your particular case here, the definition of the term (“An abnormal magnetic resonance tomography signal from a vertebral endplate according to a widely used classification published by Dr. Michael Modic.”) points quite strongly to a phenotype rather than a disease. An HP term for this already exists, so we should import the HP term as such rather than creating a new EFO term. I’d place it under EFO HP:0000924 ‘abnormality of the skeletal system’, as that's the closest ancestor of the HP term that is already in EFO.

For simplicity, let’s ignore the fact that the definition of the HP term is about the signal specifically, so if I were to create a new term for it, I’d be tempted to place it under EFO:0000720 ‘test result’ ;-) (also see https://github.com/EBISPOT/efo/issues/342)

Thanks! (I moved the ticket back to the "In progress" queue as a reminder if you wish to look into it further.)

twhetzel commented 5 years ago

@paolaroncaglia thanks for the feedback. After initially writing I found examples of other HP terms, e.g. acute pancreatitis, peritonitis, that are in EFO as new terms with an xref back to HP and in the disease branch so I added it as originally proposed.

If it should be moved as you describe above per EFO policy let me know (and if so are you suggesting two parent terms, e.g. "test result" and "abnormality of the skeletal system"?)

paolaroncaglia commented 5 years ago

@twhetzel thank you for your comments. I'd suggest to bring this ticket up at our scrum meeting tomorrow if you don't mind. If there is a preference to keep 'Modic type vertebral endplate changes' in the disease branch, I think we should at least modify its current definition (“An abnormal magnetic resonance tomography signal from a vertebral endplate according to a widely used classification published by Dr. Michael Modic.”) to point to the disorder in the affected anatomical part rather than to the abnormal reading from a medical test. As for the examples you quoted (pancreatitis and peritonitis), in EFO as diseases but with xrefs to HP terms, those are inflammatory diseases, a grouping that we are indeed planning to have in the disease branch (rather than phenotype) consistently, see https://github.com/EBISPOT/efo/issues/372. But the line may be more blurred for other phenotype terms in EFO, so discussing as a group may help. As per your last question, no, I wouldn't advice having terms as subclasses of both 'test result' and a phenotypic abnormality term. We can discuss that too if we wish. Thanks again! :-)

twhetzel commented 5 years ago

Thanks, let's chat about tis tomorrow. As for the dual subclass question, I wasn't sure what you were suggesting for: "I’d place it under EFO HP:0000924 ‘abnormality of the skeletal system’, as that's the closest ancestor of the HP term that is already in EFO.

For simplicity, let’s ignore the fact that the definition of the HP term is about the signal specifically, so if I were to create a new term for it, I’d be tempted to place it under EFO:0000720 ‘test result’"

twhetzel commented 5 years ago

Following the EFO call we have decided to: