ARAs returning non-target disease results (ARAX, ROBOKOP)

[sharatisrani]

"What drug treats Ehlers-Danlos" gives [high-scored] results for all sorts of other diseases, pivoted on drug. Especially versatile drugs.

This appears to be happening with ARAX and ROBOKOP (maybe ARAGORN)

Is this ok? Here is the list of diseases that were returned to that question:

MONDO:0001071 - Hypertrophic cardiomyopathy MONDO:0002025 - Breast cancer MONDO:0002081 - Hypertension MONDO:0002254 - Alzheimer's disease MONDO:0003847 - Asthma MONDO:0003900 - Autism spectrum disorder MONDO:0004335 - Cystic fibrosis MONDO:0004995 - Diabetes mellitus type 2 MONDO:0005066 - Epilepsy MONDO:0005093 - Gaucher disease MONDO:0005240 - Hemophilia MONDO:0005267 - Huntington's disease MONDO:0005307 - Inflammatory bowel disease MONDO:0005328 - Leukemia MONDO:0005336 - Liver cirrhosis MONDO:0005381 - Lymphoma MONDO:0005385 - Major depressive disorder MONDO:0005560 - Multiple sclerosis MONDO:0006999 - Parkinson's disease MONDO:0007523 - Rheumatoid arthritis MONDO:0007526 - Schizophrenia MONDO:0011142 - Systemic lupus erythematosus MONDO:0015286 - Thrombophilia MONDO:0016002 - Von Willebrand disease MONDO:0017314 - Myocardial infarction MONDO:0019052 - Prostate cancer MONDO:0019567 - Endometriosis MONDO:0020066 - Chronic obstructive pulmonary disease (COPD) MONDO:0020121 - Osteoporosis

[sharatisrani]

ARAX clarified that with _with knowledgetype "inferred", it behaves as expected. But here is a PK where ROBOKOP returns all sort of other diseases when Q'd for Ehlers-Danfos: c8c2369b-a722-4985-ad42-f0be4f8a8d6d?

But pls hold off pending a discussion with @dkoslicki

[sharatisrani]

This issue can be tied to #422 and #445 identified by @suihuang-ISB in his report on "Embarrassing Results"

[sharatisrani]

Issue not observed for a couple of weeks, as the 5 ARAs got finalized and all other agents not factored directly into ARS removed from the scoring. Can reopen issue if needed.