QA of AGR Ingest

[mbrush]

QA Resources:

• KGX Summary report: https://docs.google.com/spreadsheets/d/1mYfU-YNMFQ03TxHmIASKvBQU9OYTi7nWAn3qcce3vBg/edit?gid=1668228334#gid=1668228334
• RIG: https://github.com/NCATSTranslator/translator-ingests/blob/main/src/translator_ingest/ingests/alliance/alliance_rig.yaml
• RIG updates PR: https://github.com/NCATSTranslator/translator-ingests/blob/alliance-qc-fixes/src/translator_ingest/ingests/alliance/alliance_rig.yaml

---

Issues to consider / address for next update release:

• stage qualifiers specified in the RIG are missing on all expressed_in edges - add them
• The RIG indicates that a generic biolink:qualifiers property will be used to capture the type of assay/method supporting a given expressed_in edge. Consider using the biolink:supporting study method types edge property instead of making this a qualifier. Update RIG accordingly if we go with this approach.
• In the original RIG we indicated that experimental conditions under which a phenotype was observed would be captured as a qualifier on has_phentype edges. This is not in the latest RIG - what was the barrier for this? If we plan to include later, can we add as a future_consideration in the RIG?
  • MHB: Kevin indicated this data is not particularly relevant for Translator use cases, so we can not worry about it for now.

---

Issues to consider longer term:

• Consider how these expressed_in edges differ from those we get from BGee, and make sure EPC metadata and UI explanations clearly explain methods/provenance, and allow assessment of differences and utility. Consider if there are modeling changes that could explicitly encode this.

---

Questions:

• I see monarch initiative as an aggregator knowledge source on these edges . . . is this because we ingest from Monarch's ingest of AGR?