This test seems to occasionally fail due to the generated graph being too large (exhausted memory). This can be fixed by replacing EMMO with a smaller test ontology. This will also make the test run faster.