Gene to Pathway - Some ARS Agents Currently Returning 0 Results

[sstemann]

https://github.com/NCATSTranslator/testing/blob/main/ars-requests/not-none/pathway.json

• the query has been updated since this morning to correctly capitalize NCBIGENE:1017
• the parent PK is 5e3f38b6-2b23-479c-b04e-81fce973ca52
• as of weds 3/10 : 4525e749-aef6-4025-9c54-b90cfd064305
• as of fri 3/19: 56085e63-c878-4634-8a70-69810f858e1e

Based on the updated query: ARAX, ARAGORN, and BTE are now returning various results.

Adding the non-responsive ARA reps to the assignees, but the KPs are the domain specialists and should serve as a resource to explore what could be done to answer/expand on queries like the one proposed.

@NCATSTranslator/translator-standup for KPs to take a look at and determine if they can fill gaps on the query for ARAs, assigning to non-responsive ARA reps to the assignees to provide feedback here or open tickets in their repos and add the link here, but the KPs are the domain specialists and should serve as a resource to explore what could be done to answer/expand on queries like the one proposed.

[sstemann]

Explanatory Agent - no support for this particular predicate (gene-pathway association) but are hitting MolePro Improving Agent - capitalization we confirmed change, need to implement gene-pathway, not finding the CURIE also

Unsecret - hits Molepro, Vlado exposed pathways (but 0.9.2 compliant) - so both have results but marked as non-compliant

[sstemann]

@cbizon are the ARAGORN message IDs:

Weds @ 9am: 9aa08c33-0cd0-4401-b834-f2b59b585f14 Fri @ 950am: 3e32f299-c335-4b79-9b3e-443661bc626f

[cbizon]

@cbizon are the ARAGORN message IDs:

Weds @ 9am: 9aa08c33-0cd0-4401-b834-f2b59b585f14 Fri @ 950am: 3e32f299-c335-4b79-9b3e-443661bc626f

We recently updated the URL that the ARS should be hitting. For some reason (maybe the change just hadn't gone through yet) the Wed version was hitting the old URL, which gave results, but these results were buggy.

The "correct" result is that there were no results coming back.

We're working on pulling this information from various KPs, documented in our ticket https://github.com/ranking-agent/strider/issues/107 and KP tickets linked from there.

[ehinderer]

I suspect that Explanatory Agent is not returning results because we do not yet have a use-case that includes gene-pathway associations and as such, do not have a workflow in place to return a response including that predicate. We have use cases for other predicates in MolePro. The new requirements for full integration dictates that we account for all predicates in supported KPs, so we will be looking into this. @pg427 @tziomics

[sstemann]

I suspect that Explanatory Agent is not returning results because we do not yet have a use-case that includes gene-pathway associations and as such, do not have a workflow in place to return a response including that predicate. We have use cases for other predicates in MolePro. The new requirements for full integration dictates that we account for all predicates in supported KPs, so we will be looking into this. @pg427 @tziomics

@ehinderer do you all track this in github or somewhere we can link to?

[cbizon]

With the BTE fix, we're now returning results for this query.

[ehinderer]

@sstemann I've commented to check additional 0-results errors in issue 18 https://github.com/NCATSTranslator/Explanatory-Agent/issues/18.

[sstemann]

as of 4/20/21 (PK: 605fc2ef-58bd-4176-8661-9a5a09f26c83):

[balhoff]

CAM-KP has data for this gene, but not related to any concepts mapped to biolink:Pathway (i.e., GO:0007165). Generalizing to biolink:BiologicalProcess returns GO:0030851 'granulocyte differentiation'.

[sstemann]

@cbizon previously ARAGORN responded to this query. The query has been updated for v1.1 The results are here: 3eb5b2de-c28f-4cec-a1df-3584fbc25d3d

ARAGORN now has 0.

[sstemann]

@cbizon sorry for my confusion - i think during the evolution of this query - we changed it from Pathway to BiologicalProcess. So using BiologicalProcess instead of Pathway - ARAGORN does respond. 87aff34f-c138-4618-92f7-f8b1e8287f1d

However, now Textmining responds to pathway but not biologicalprocess @bill-baumgartner so i'm not sure which

[sstemann]

TRAPI v1.1 Queries as of 7/20:

1. Biological Process
2. Pathway

[colleenXu]

I ran the updated query below through the ARS, PK: f4818fbf-9293-4f33-bc3e-b184ea8dc193

The updated query should be something like this (it's missing the workflow section but the rest is TRAPI v1.2):

{
    "message": {
        "query_graph": {
            "nodes": {
                "n0": {
                    "ids": ["NCBIGene:1017"],
                    "categories":["biolink:Gene"]
                },
                "n1": {
                    "categories": ["biolink:Pathway"]
                }
            },
            "edges": {
                "e0": {
                    "subject": "n0",
                    "object": "n1"
                }
            }
        }
    }
}