Try 'pruning' the terms associated with 'Pancreatic Stellate Cells'... if all leaf nodes of a parent are enriched, then drop the leaf terms. Then we can see a cleaner list of what phenotypes are associated with it

[NathanSkene]

Try 'pruning' the terms associated with 'Pancreatic Stellate Cells'... if all leaf nodes of a parent are enriched, then drop the leaf terms. Then we can see a cleaner list of what phenotypes are associated with it

[bschilder]

Took a somewhat inverted approach by checking if each term have an descendants. If it does, only include its most precise descendants. This helps remove redundant terms in the target prioritisation pipeline, while retaining maximum precision of phenoptypes.

library(KGExplorer)
dat <- data.table::data.table(hpo_id=c("HP:0000001","HP:0000002","HP:0000003"),
                             name=c("term1","term2","term3"))
ont <- get_ontology("hp")
dat2 <- prune_ancestors(dat,ont=ont)