Neo4j and Cypher Notes

Creating Drug/Chemical Co-Occurrence Graph

• Data on occurance of drugs/chemicals in pubmed abstracts imported
• generated in section detailed in elasticsearch document

Python neo4j-driver used to create a graph data base with the schema:

Querying Reactome in Cypher

Finding gene products associated with cellular response to oxidative stress

MATCH (gt:GO_Term{accession: "0034599"})-[:goBiologicalProcess]-(p:Pathway)-[]-(r:Reaction)-[]-(ge:GenomeEncodedEntity)
WHERE ge.speciesName = "Homo sapiens"
RETURN DISTINCT ge.displayName
LIMIT 10

Results

"2xHC-TXN2 [mitochondrial matrix]"
"TXN [cytosol]"
"P4HB [endoplasmic reticulum lumen]"
"2xHC-TXN [cytosol]"
"HC53,56-P4HB [endoplasmic reticulum lumen]"
"TXN2 [mitochondrial matrix]"

Finding gene products associated with cardiovascular system disease

MATCH (d:Disease)-[]-(p:PhysicalEntity)
WHERE toLower(toString(d.displayName)) CONTAINS 'card'
OR toLower(toString(d.displayName)) CONTAINS 'vasc'
RETURN d.displayName, size((d)--(:Drug)) as Drugs, size((d)--(:GenomeEncodedEntity)) as GeneEntities,  count(p) AS PhysicalEntityCount, size((d)--(:PhysicalEntity)--(:Disease)) as PEAssocDisease, size((d)--(:PhysicalEntity)--(:ReactionLikeEvent)) as PEAssocReactLike, size((d)--(:PhysicalEntity)--(:GenomeEncodedEntity)) as PEAssocGeneEntities
ORDER BY PhysicalEntityCount DESC

Results

d.displayName	Drugs	GeneEntities	PhysicalEntityCount
"cardiofaciocutaneous syndrome"	0	23	34
"myocardial infarction"	20	0	20
"basal ganglia cerebrovascular disease"	0	6	7
"cerebrovascular disease"	5	0	5