Ubergraph merges 50+ mutually referential biomedical and life-science ontologies into a single RDF triplestore with precomputed inference closures for all relations — queryable over a public SPARQL endpoint.
Enter an OBO ID (CL:0000080) or full IRI to see its relationships in the nonredundant graph. Subsumers sit toward the top; subClassOf runs vertically. Click any node to re-center on it.
Data is fetched live from the Ubergraph SPARQL endpoint. High-level terms can have thousands of subclasses; results are capped per direction and the true totals are shown above the graph.
Everything is available over standard SPARQL, with example queries and an OpenAPI wrapper.
Query all named graphs directly at the public endpoint:
https://ubergraph.apps.renci.org/sparql
Curated example queries in the repo are exposed as a REST API with an interactive UI via grlc.
Include the ontology graph for rdfs:label. Match rdfs:isDefinedBy to find a term's source ontology quickly.
The triplestore is organized into named graphs — the four primary graphs below, plus a separate named graph holding each individual source ontology. All predicates other than rdfs:subClassOf in the redundant / nonredundant graphs denote OWL existential restrictions — e.g. X part_of Y means X SubClassOf (part_of some Y).
http://reasoner.renci.org/ontology
Merged logical & annotation axioms, ELK classification, rdfs:isDefinedBy links, and normalized information-content scores. Include this for term labels.
http://reasoner.renci.org/redundant
The complete inference closure for all subclass and existential relations, including transitive and reflexive edges.
http://reasoner.renci.org/nonredundant
A pruned subset of the redundant graph with the same semantics — the compact view used by the explorer above.
https://biolink.github.io/biolink-model/
The Biolink model in RDF, plus term-to-category mappings via biolink:category.
A merged set of mutually referential OBO ontologies spanning anatomy, phenotypes, diseases, chemistry, environment, and more.
Files for the current build are published at RENCI.
Integer-based edge tables for graph analysis & ML — nodes, edges, and labels.