Chemical Treats Rattus Norvegicus

NCATSTranslator / Feedback

A repo for tracking gaps in Translator data and finding ways to fill them.

7 stars 0 forks source link

Chemical Treats Rattus Norvegicus #707

Closed sstemann closed 3 months ago

sstemann commented 9 months ago

This is a question of answer quality - the overall result is not in question but the supporting evidence in the supporting paths which have Rattus Norvegicus seem fairly general

UI Test > MVP1 > Common Cold: https://ui.test.transltr.io/main/results?l=Common%20Cold&i=MONDO:0005709&t=0&r=0&q=a49c78af-b108-490e-b48f-f7aa7a5e568c

Epoprostenol Treats Rattus Norvegicus

Rattus Norvegicus Has Part Ins

Ins Associated With Common Cold

Looks to be coming from ARAX - ARAX GUI: https://arax.ncats.io/beta/?r=a49c78af-b108-490e-b48f-f7aa7a5e568c

gglusman commented 9 months ago

Interesting that the first publication allegedly supporting the 'RN has_part INS' edge is... in mice, and rats aren't mentioned in it at all.

sstemann commented 8 months ago

@edeutsch this result in question is from ARAX via semmed, can you take a look?

edeutsch commented 8 months ago

yeah, it's coming from the xDTD machine learning model that was trained with data that includes a silly assertion from SemMedDB. On the plus side, there do seem to be ample publications that indicate that many researchers do indeed treat rats with Epoprostenol.

bill-baumgartner commented 8 months ago

One potential solution to this issue could be to leverage the domain and range of Biolink predicates. In the case of biolink:treats its range is specified as disease or phenotypic feature. Assertions like chemical - treats - taxon are, therefore, technically not compliant with the Biolink model and could be filtered automatically.