Overview

BioCypher development was initially centred around a Neo4j graph database output due to the migration of OmniPath to a Neo4j backend. Importantly, we understand BioCypher as an abstraction of the build process of a biomedical knowledge graph, and thus are open towards any output format for the knowledge representation.

The used output format is specified via the dbms parameter in the biocypher_config.yaml (see the Configuration for an example). Currently supported are:

• neo4j
• arangodb
• rdf
• owl
• postgres
• sqlite
• tabular
• csv
• pandas
• networkx

Furthermore, you can specify whether to use the offline or online mode.

• For the online mode set offline: false. The behavior of the online mode depends on the specified dbms. If the specified dbms is an in-memory database (e.g. csv, networkx), the in-memory Knowledge Graph can directly be accessed from the BioCypher instance. If the specified dbms is a database (e.g. neo4j), the online mode requires a running database instance and BioCypher will connect to this instance and directly writes the output to the database.

• For the offline mode set offline: true. BioCypher will write the knowledge graph to files in a designated output folder (standard being biocypher-out/ and the current datetime). Furthermore, you can generate a bash script to insert the knowledge graph files into the specified dbms by running bc.write_import_call().

Warning

The online mode is currently only supported for neo4j, tabular, csv, pandas, and networkx.

Details about the usage of the online and offline mode and the different supported output formats are described on individual pages for each output format.

Available Output formats

Method	Offline mode	Online mode (in-memory)
Neo4j
ArangoDB		(pending)
RDF / OWL		(pending)
PostgreSQL		(pending)
SQLite		(pending)
Tabular
CSV
Pandas
NetworkX