Ubergraph paper presented at ICBO 2022: https://doi.org/10.5281/zenodo.7249759
- Merged set of mutually referential OBO ontologies:
- Uberon anatomy
- Cell Ontology (CL)
- Provisional Cell Ontology (PCL)
- Gene Ontology (GO)
- Biospatial Ontology (BSPO)
- Phenotype and Trait Ontology (PATO)
- Ontology of Biological Attributes (OBA)
- Human Phenotype Ontology (HPO)
- Mammalian Phenotype Ontology (MP)
- Fission Yeast Phenotype Ontology (FYPO)
- Monarch Disease Ontology (MONDO)
- Chemical Entities of Biological Interest (ChEBI)
- Relations Ontology (RO)
- Environmental conditions, treatments and exposures ontology (ECTO)
- Environment Ontology (ENVO)
- Ontology for Biomedical Investigations (OBI)
- Medical Action Ontology (MAXO)
- Evidence and Conclusion Ontology (ECO)
- NCBI Taxonomy
- Protein Ontology (PRO)
- NCIt OBO Edition (NCIT)
- Plant Ontology (PO)
- Plant Phenology Ontology (PPO)
- Plant Trait Ontology (TO)
- Plant Experimental Conditions Ontology (PECO)
- Drosophila Anatomy Ontology (FBbt)
- Food Ontology (FoodOn)
- Population and Community Ontology (PCO)
- Information Artifact Ontology (IAO)
- Measurement Method Ontology (MMO)
- Molecular Interactions Controlled Vocabulary (MI)
- Sequence Ontology (SO)
- C. elegans Gross Anatomy Ontology (WBbt)
- C. elegans development ontology (WBls)
- C. elegans phenotype (WBPhenotype)
- Zebrafish anatomy and development ontology (ZFA)
- Ascomycete phenotype ontology (APO)
- Mouse Developmental Stages (MmusDV)
- Vertebrate Breed Ontology (VBO)
- Units of Measurement Ontology (UO)
- Precomputed OWL classification (class hierarchy)
- Materialized “class relations” from existential property restrictions (in separate named graphs)
- Blazegraph RDF triplestore; SPARQL query interface
- Service:
https://ubergraph.apps.renci.org/sparql
- Query interface in YASGUI (this query retrieves the current build date): https://api.triplydb.com/s/oQgG4f0gs
Example SPARQL queries are included in the repo, and can be queried via an OpenAPI user interface at https://grlc.io/api-git/INCATools/ubergraph/subdir/sparql.
The Ubergraph triplestore is organized into several named graphs.
-
http://reasoner.renci.org/ontology
— All the merged axioms from the input ontologies (logical axioms and annotation axioms), classified usingrobot reason -r ELK
. Disjointness axioms are removed prior to reasoning. Include this graph if you want term labels.- This graph also includes:
- Generated
rdfs:isDefinedBy
triples connecting ontology terms to their OBO namespace ontology IRI. For example, there is a triple<http://purl.obolibrary.org/obo/FOO_0123456> rdfs:isDefinedBy <http://purl.obolibrary.org/obo/foo.owl>
for every term with OBO namespace "foo". Matching these triples is much faster than usingFILTER(STRSTARTS(STR(?term), "http://purl.obolibrary.org/obo/FOO_"))
. - Precomputed information content score for each ontology class, based on the count of terms related to a given term via
rdfs:subClassOf
or any existential relation. The scores arexsd:decimal
values scaled from0
to100
(e.g., a very specific term with no subclasses). Use the predicatehttp://reasoner.renci.org/vocab/normalizedInformationContent
to retrieve IC scores.
- Generated
- This graph also includes:
-
http://reasoner.renci.org/redundant
— The complete inference closure for all subclass and existential relations. This includes all transitive, reflexive subclass relations. Within this graph, all predicates with the exception ofrdfs:subClassOf
imply an OWL existential relationship. For example, the triple:CL:0000080 BFO:0000050 UBERON:0000179
(circulating cell • part of • haemolymphatic fluid)
is shorthand for the OWL axiom
- circulating cell SubClassOf (part of some haemolymphatic fluid)
-
http://reasoner.renci.org/nonredundant
— Triples in this graph are a subset of, and have the same semantics as, the "redundant" graph, pruned according to several redundancy rules. -
https://biolink.github.io/biolink-model/
— RDF rendering of the Biolink model, as well as triples connecting ontology terms to Biolink categories (derived from mappings in Biolink model), using the predicatehttps://w3id.org/biolink/vocab/category
.
You can download files pertaining to the current build.
- RDF N-Quads (RDF four column format: subject, predicate, object, graph)
- Blazegraph database file
- Redundant and nonredundant graphs as integer-based edge tables for graph analysis or machine-learning.
- Archives contain these files:
edges.tsv
: tab-separated subject, predicate, object integer IDsnode-labels.tsv
: tab-separated node ID, node IRI (nodes are subjects and objects in the edges table)edge-labels.tsv
: tab-separated relation ID, relation IRIbuild-metadata.nt
: RDF describing Ubergraph build date
- Available graphs:
- Archives contain these files:
This repository includes a CI check to provide early warning for incompatibilites between some of the most tightly coupled ontologies. In this GitHub action, the ontologies are built from the latest source and checked for unsatisfiable classes.
Development of Ubergraph has been supported by the NCATS Biomedical Data Translator project, as well as by RENCI.