A demonstration of building a unified knowledge graph of OBO Foundry ontologies
building on ROBOT for parsing ontologies,
bioontologies
for wrapping
ROBOT in Python and providing an object model of
OBO Graph JSON, and the
Bioregistry for standardization.
Warning Despite the commitment of the OBO community to adopting share semantics, it's clear from the high-level summary of relations that many relationships remain unstandardized and from the example triples that many entities remain unstandardized, so this graph should not be used directly without the generation and application of additional rules to fix its content.
graph.tsv.gz
The graph is a gzipped tab-separated values file that contains the following columns:
subject
- the subject entity as a CURIE when possible to normalizeprediate
- the predicate entity as a CURIE when possible to normalizeobject
- the object entity as a CURIE when possible to normalizesources
- a pipe-delimited list of ontologies in which the triple was foundThe graph can be re-built with the following code:
$ git clone https://github.com/cthoyt/obo-foundry-graph.git
$ cd obo-foundry-graph
$ pip install tox
$ tox