Trait processing is parameterised by a "target ontology" which it will use for querying Zooma/OxO, as well as checking obsoleteness etc. This is specified in the header of the latest mappings file, to ensure the relevant ontology is always kept together with the current mappings.
XML annotation will use the relevant ontology as the DB attribute
Evaluation code and evidence string generation are not changed, i.e. will always assume EFO mappings (I think the evaluation code could be updated if needed)
Various cosmetic changes (e.g. changing variable names)
Added end-to-end tests for the curation Nextflow pipelines so I could ensure I wasn't breaking anything, though these are not currently run in Github because they take about 20 minutes to run
I tested the three pipelines using Mondo as the target ontology and it does work, though of course the output is less valuable as it's not backed by years of curation and feedback... One mildly interesting thing is that as HP terms are sometimes imported into Mondo, it will add this information to the ClinVar XML - e.g.
Summary of changes:
I tested the three pipelines using Mondo as the target ontology and it does work, though of course the output is less valuable as it's not backed by years of curation and feedback... One mildly interesting thing is that as HP terms are sometimes imported into Mondo, it will add this information to the ClinVar XML - e.g.