strange subclass_of edge in KG2.8.1pre

[saramsey]

From Kaiwen He (reported by Slack DM):

Hi @ramseyst ! I notice something weird from an entry in RTX-KG2 2.8.1:

MONDO:0008491---rdfs:subClassOf---None---None---None---MONDO:0005015---OBO:mondo.owl    False   MONDO:000501biolink:subclass_of subclass_of infores:mondo       {}              subclass_of rdfs:subClassOf MONDO:0008491   2023-02-20 21:03:32 GMT biolink:subclass_of MONDO:0008491   MONDO:0005015

For this edge, it claims that MONDO:0008491 "stiff-person syndrome" is a subclass of MONDO:0005015 "diabetes mellitus". According to the description of stiff-person syndrome, it does not sound like a subtype of diabetes. Stiff-man syndrome (SMS) is a rare neurological disorder comprising fluctuating trunk and limb stiffness, painful muscle spasms, task-specific phobia, an exaggerated startle response, and ankylosing deformities such as fixed lumbar hyperlordosis. From this link, there is a tree showing the ontology of stiff-person syndrome , but diabetes mellitus is not one of the superclasses. [Also attached the screenshot of the tree below] Could you please verify this edge? I ask because it is related to one of the issues from NCATSTranslator/Feedback: https://github.com/NCATSTranslator/Feedback/issues/519, given that progressive encephalomyelitis with rigidity and myoclonus is a subclass of stiff-person syndrome . Thank you for your time!

https://github.com/NCATSTranslator/Feedback/issues/519

[saramsey]

Kaiwen confirms, they saw this edge in KG2.8.1pre (the pre-canonicalized graph).

[acevedol]

On KG2.8.6pre I don't see this edge. I ran the query match (n {id:"MONDO:0008491"})-[r]->(m {id:"MONDO:0005015"}) return * and returned no results, as shown below. No results were returned for the same query against 2.8.4pre also @saramsey Is there another way I should investigate this?

[saramsey]

In KG2.10.0c, the only edge between MONDO:0008491 and MONDO:0005015 is biolink:coexists_with. Closing this issue.