Is process active in a tissue or cell

[Chris-Evelo]

For a process (WikiPathways pathway, GO term, or pathway ontology term) find all proteins involved and see how many are expressed in a specific tissue or cell type (using NextPot as the source for expression)

[Chris-Evelo]

Tissue/cell combo's to try: liver/hepatocyte, fat(adipose tissue)/adipocyte

[Chris-Evelo]

For this one we decided to try simple SPARQL queries on the relevant resources and combine them in Python.

[AlasdairGray]

Query for discovering the cells in which a given protein is highly expressed
https://github.com/openphacts/FederatedPhacts/blob/master/nextprot-proteinHighlyExpressedInCell.rq
Available as a REST interface taking the nextprot identifier for the protein as an argument
http://grlc.io/api/openphacts/FederatedPhacts#/nextprot/get_nextprot_proteinHighlyExpressedInCell

Query for discovering the tissues in which a given protein is highly expressed
https://github.com/openphacts/FederatedPhacts/blob/master/nextprot-highlyExpressed.rq
Available as a REST interface taking the nextprot identifier for the protein as an argument
http://grlc.io/api/openphacts/FederatedPhacts#/nextprot/get_nextprot_highlyExpressed

[AlasdairGray]

In the query to discover tissues in which a given protein is highly expressed, we do not filter for things that are declared to be of type nextprot:Tissue since this is not used consistently in the nextprot data and eliminates too much data.

The query would then have been

select distinct ?tissue ?tisslab where {
  ?_entry_iri :isoform ?iso.
  # get all expression
  ?iso :expression ?e1.
  # highly expressed
  ?e1 :evidence/:expressionLevel :High.
  ?e1 :term ?tissue .
  ?tissue rdfs:label ?tisslab .
  #Return things that are tissues
  ?tissue :childOf cv:TS-2090
  #Eliminate things that are cells
  filter not exists {?tissue :childOf cv:TS-2035}
}