queries.txt

# Get mappings distance 2 filtering by either source or datsource target
MATCH path= shortestPath( (ft:Term)-[m:MAPPING*1..2]-(tt:Term))
WHERE ft.curie = 'EFO:0004230'
WITH tt,path, extract ( r in m |  r.sourcePrefix) as source
MATCH (tt)-[HAS_SOURCE]-(td)
WHERE 'KEGG' in td.prefix and 'efo' in source
UNWIND source as source1
RETURN tt.curie, tt.label,collect (distinct td.prefix), collect (distinct source1), length(path) as dist
ORDER BY dist


# old experimental stuff below

CREATE CONSTRAINT ON (i:Identifier) ASSERT i.term IS UNIQUE
CREATE CONSTRAINT ON (i:Uri) ASSERT i.term IS UNIQUE
CREATE CONSTRAINT ON (i:Cui) ASSERT i.term IS UNIQUE
CREATE CONSTRAINT ON (i:PrefixedCui) ASSERT i.term IS UNIQUE
CREATE CONSTRAINT ON (i:Datasource) ASSERT i.prefix IS UNIQUE

CREATE INDEX ON :Datasource(source)
CREATE INDEX ON :MAPPING(sourceName)
CREATE INDEX ON :MAPPING(scope)


Queries

Get direct xrefs for a term (show sources)


Match (u:Uri)-[r1:XREF]->(x:PrefixedCui)<-[r2:XREF]-(u1:Uri)
Where u.term = 'http://www.ebi.ac.uk/efo/EFO_0000349'
Return u1.term, collect (distinct r1.sourcePrefix) + collect(distinct r2.sourcePrefix)


Get derived xrefs (1 hop) - high score

term -> xref -> term -> xref

Get derived xrefs (4 hops) - low score

Get lexical mappings
	- search zooma, get best hits, get high score mappings for those


Mapping project

1. Take HP, hit Zooma with each label, get ORDO hits. 
2. Take HP, hit Zooma, with each id hit, get mappings to ordo. 


MATCH (c:Identifier)-[*1..5]-(m:Uri) 
WHERE c.term = 'HP:0000819'
WITH m as mappedUri
MATCH (mappedUri)-[:DATASOURCE]->(d:Datasource { prefix : 'ordo'})
WITH mappedUri as ordoUri
MATCH (ordoUri)-[ATLID]->(ordoP:PrefixedCui)
RETURN distinct ordoP.term


MATCH (source:Uri)-[DATASOURCE]->(d:Datasource { prefix : 'efo'})
WITH source
MATCH path = allShortestPaths ( (c:Identifier)-[:ALTID|:XREF*..10]-(source) )
WHERE c.term = 'DOID_4467'
RETURN distinct source.term , length(path) as dist
ORDER BY dist


// find a Uri 

MATCH (i:Identifier)-[*0..1]-(u:Uri)
WHERE i.term = 'EFO_0000001'
RETURN u.term



// Use case

Given set of ids

1. Get all mappings (don't map to own datasource)

x1 mapped to y1 (1 sources)
x1 mapped to y2 (2 sources)
x1 mapped to z3 (2 sources) (inferred) (show)

2. Get mappings to ontology (e..g HP)

x1 mapped to y1 (1 sources)
x1 mapped to y2 (2 sources)
x1 mapped to z3 (2 sources) (inferred) (show)

3. Need ability to validate

4. Ability to add new mapping