search

monarch-initiative / monarch-ingest

Data ingest application for Monarch Initiative knowledge graph using Koza
https://monarchinitiative.org
14 stars 1 forks source link

g2p via orthology "ingest" #425

Closed kevinschaper closed 4 months ago

kevinschaper commented 1 year ago

We yet have support for "ortholog-phenotype" associations implemented by populating our solr index with subject_ortholog_closure, before we implement that, it's worth considering whether it would make more sense to create edges with appropriate modeling.

Given

subject: ZFIN:ZDB-GENE-990415-8
predicate: biolink:has_phenotype
object: ZP:0000032

plus 

subject: HGNC:8616
biolink:orthologous_to
object: ZFIN:ZDB-GENE-990415-8

and perhaps some roll-up to an UPHENO or HP term? we don't have any direct edge for that currently that I can tell.

We would want to create an association that allows the original to be fully reconstructed

subject: HGNC:8616
[subject_qualifier? original_subject? subject_ortholog_qualifier?]: ZFIN:ZDB-GENE-990415-8
predicate: biolink:has_phenotype [too strong of an assertion? how do we say 'via orthology'?]
object: [UPHENO or HP term that is an appropriate replacement?]
[object_qualifier? original_object? subject_phenlog_qualifier?]: ZP:0000032

We should probably discuss this on a data call first, then a biolink modeling call?

cc: @putmantime @cmungall @monicacecilia @mbrush @sierra-moxon

monicacecilia commented 4 months ago

As discussed, this is being addressed differently on the UI/API front.