OMOP_to_Graph

Transform the OMOP CDM to graph model
Manuscript published on 2024/12 - Evolution of a Graph Model for the OMOP Common Data Model

1. Draft Schemas

1. Schema Draft 1
2. Schema Draft 2 [ Exclude ]
3. Schema Draft 3 [ To review ]
4. Schema Draft 4 [Finalized version for AMIA submission]

2. Multiple ways to load data and build the database in seconds to minutes -

2.1 Create the neo4j database directly from a relational database instructions

(This method is only recommended for small to medium sized database, as the JDBC connection can get timeout when loading too much data in one transaction - in general it is a good practice to preprocess the data and then load into Neo4j)

1. Create a local or cloud Neo4j database - more information here
2. Use neo4j drivers and embedd the cypher code into your scripts, either Python, Java, JavaScript .etc
   OR
   Use Cypher shell to run the code directly in the terminal - more information here
   Steps:
   • Go to the local bin folder
   • ./cypher-shell -a [your-server-link]:[your-port-number] -u [your-username]
   • Copy paste the code you want to load or use :source import/your_cypher_file.cypher that contains the loading code

2.2 Build database with neo4j-admin import

(fastest import method, only works for empty database,datasets in CSV format)

1. Prepare datasets with the sql code from OMOP relational database and save as CSV files as below
   cond_visit.csv
   cond.csv
   drug_visit.csv
   drug.csv
   eth.csv
   gender.csv
   has_cond.csv
   has_drug.csv
   has_eth.csv
   has_gender.csv
   has_meas.csv
   has_obs_period.csv
   has_obs.csv
   has_proc.csv
   has_race.csv
   has_visit.csv
   meas_visit.csv
   meas.csv
   obs_period.csv
   obs_visit.csv
   obs.csv
   person.csv
   proc_visit.csv
   proc.csv
   race.csv
   visit.csv
2. run the command below in the Neo4j database shell
   ./bin/neo4j-admin import --database [database_name] --nodes=Person=import/person.csv --nodes=Measurement=import/meas.csv --nodes=VisitOccurrence=import/visit.csv --nodes=ConditionOccurrence=import/cond.csv --nodes=DrugExposure=import/drug.csv --nodes=Observation=import/obs.csv --nodes=ObservationPeriod=import/obs_period.csv --nodes=ProcedureOccurrence=import/proc.csv --nodes=Race=import/race.csv --nodes=Gender=import/gender.csv --nodes=Ethnicity=import/eth.csv --relationships=HAS_MEASUREMENT=import/has_meas.csv --relationships=ASSOCIATED_DURING_VISIT=import/meas_visit.csv, import/cond_visit.csv, import/drug_visit.csv, import/proc_visit.csv, import/obs_visit.csv --relationships=HAS_CONDITION_OCCURRENCE=import/has_cond.csv --relationships=HAS_DRUG_EXPOSURE=import/has_drug.csv --relationships=HAS_PROCEDURE_OCCURRENCE=import/has_proc.csv --relationships=HAS_VISIT=import/has_visit.csv --relationships=HAS_OBSERVATION=import/has_obs.csv --relationships=HAS_OBSERVATION_PERIOD=import/has_obs_period.csv --relationships=HAS_Gender=import/has_gender.csv --relationships=HAS_race=import/has_race.csv --relationships=HAS_ETHNICITY=import/has_eth.csv --ignore-empty-strings