search

monarch-initiative / monarch-cypher-queries

4 stars 3 forks source link

Check logic of phenotype-pathway query #10

Open kshefchek opened 7 years ago

kshefchek commented 7 years ago

We have a query that infers phenotype to pathway associations through gene/variant to phenotype associations, optionally through orthologs. I'm not sure if it's biologically accurate to make these inferences. I'm also not sure if these are configured to be shown on phenotype pages in the monarch app.

See https://github.com/monarch-initiative/monarch-cypher-queries/blob/master/src/main/cypher/golr-loader/pathway-phenotype.yaml

When I think "phenotypes of pathways" I assume we mean pathways linked to diseases, and the phenotypes associated with those diseases, but this query is not included.

kshefchek commented 7 years ago

Re-reading https://github.com/monarch-initiative/monarch-analysis/issues/2, looks like this type of query is necessary for pathway analysis.

kshefchek commented 7 years ago

Reopening for review

mbrush commented 7 years ago

Summarizing the logic Kent implement in these cypher queries to infer pathway-phenotype associations. This file is a union of three separate queries. All of them essentially propagate phenotypes asserted on a genotype or variant to the affected gene(s) and then to any pathway in which this gene (or homologs of the gene) participates.

The logic Kent implemented here seems sound - just want to confirm with @mellybelly, others who requested these inferred associations that these inference paths are what you had in mind for the use case described here.