Workflow B: sytematically troubleshoot B.1 three-hop

[karafecho]

The workflow B.1 three-hop queries are no longer returning expected results that were interesting, such as dexamethasone, resveratrol, silybin, and methylprednisone. This issue is to suggest a systematic approach to troubleshooting, perhaps by working backward and/or forward and testing each individual one-hop query.

[karafecho]

Comment: The B.1 ARS and/or ARAX queries that I ran during the week of September 13 returned dexamethasone, resveratrol, silybin, curcumin, and dextran sulfate, which are all of interest (the first three more than the last two). Methylprednisone was returned by direct ARAGORN query. I saved all PKs and ARAX response keys (ARAGORN does not offer response keys), as well as additional notes, in my running weekly progress report on Workflow B. Casey and I are now using my notes and the former results to track down the missing answers.

Notable changes over the past two weeks:

• RTX KG 2.7.1 -> 2.7.3
• Lost genes for MMP2, PPIG, IL2, possible others
• Reduced number of input CURIES
• Swapped out biolink:related_to in favor of biolink:interacts_with
• Multiomics now contributes to first hop, which may add to the combinatorial explosion of results