Creative "one" hops

[dkoslicki]

From the meeting 3/11/22: we need an operation that, when presented with a one hop query, signals to an ARA that this query can be satisfied with any number of hops (or a variable number of hops, not all expressible with a single query graph) in any fashion the ARA deems appropriate, as long as the "spirit" of the predicate on the one-hop-query edge is "roughly" satisfied.

This is for the flagship, template queries to let the ARAs flex their unique reasoning capabilities.

[dkoslicki]

Clarifications:

• This is for ARAs only, KPs need not worry about this operation as they will never be asked to respond to this operation
• It is completely acceptable to respond to this query identically to lookup (i.e. just find bioentities that strictly adhere to the query graph structure)
• Diversity of answers is encouraged

[dkoslicki]

First up for these "in spirit" operations

• up-regulates in some way
• down-regulates in some way
• may treat/ameliorate

[dkoslicki]

Current test query is: "What drugs could possibly be used to treat Alkaptonuria"

Attached are the results of the ML model developed by @chunyuma
alkaptonuria_subgraphs.pdf

[cbizon]

Here is a json file with some supposed trapi output from ARAGORN for this kind of query. Note that this is TRAPI I wrote by hand, so I am not 100% sure of its validity.
creative_aragorn_output.json.txt

[finnagin]

Also linking a json file of ARAX's response message here: dtd_creative_alkaptonuria_results_message.json