search

monarch-initiative / monarch-legacy

Monarch web application and API
BSD 3-Clause "New" or "Revised" License
42 stars 37 forks source link

inferred genes to disease not quite right #286

Closed nlwashington closed 10 years ago

nlwashington commented 10 years ago

if i look on this page for Alzheimers Disease: http://monarchinitiative.org/disease/OMIM_104300 There are 25 genes listed.

But if i navigate to its parental term, Alzheimers Disease: http://monarchinitiative.org/disease/DOID_10652 There are only 19 genes.

Similarly for a more-distant disease, 'Dementia': http://monarchinitiative.org/disease/DOID_1307 still only 19 genes.

Something is not right with the inheritance here. @cmungall Is it an ontology issue or a data issue?

cmungall commented 10 years ago

Hmmm

The parent gets 19: http://neuinfo.org/servicesv1/v1/federation/data/nif-0000-03216-7.json?exportType=data&q=DOID:10652&includeSubclasses=true

The child gets 16: http://neuinfo.org/servicesv1/v1/federation/data/nif-0000-03216-7.json?exportType=data&q=OMIM:104300&includeSubclasses=true

This is as expect

So if fed returns 16 for the child, where do the 25 on the page come from?

cmungall commented 10 years ago

ahh, clinvar. Need to add this to the closure set

cmungall commented 10 years ago

actually not sure what's going on here. But the IDs without labels in the gene sets in the child look suspicious

mellybelly commented 10 years ago

Can this be fixed for 1.0 release?

nlwashington commented 10 years ago

i think this is good now. cognitive disorder (parent of dementia, http://tartini.crbs.ucsd.edu/disease/DOID_1561) lists 118 genes now > Alzheimer's at 83. reopen if there is a problem.