Visibility: Public
BioBTree is a unified biomedical knowledge graph that integrates more than 70 primary data sources, including Ensembl, UniProt, PDB, ChEMBL, Reactome, and ClinVar, into a single queryable graph of cross-reference edges spanning genes, proteins, compounds, diseases, pathways, and clinical data. It lets users search, map, and traverse identifiers across databases with a chain-query syntax (for example, mapping a gene in Ensembl to UniProt proteins and on to PDB structures), and exposes the graph through a REST API, a Model Context Protocol (MCP) server, and a question-answering web interface for grounding LLM responses in structured biomedical data.
Graph at a glance
40,160,474
Nodes
132,075,627
Edges
26
Node categories
32
Predicate types
Products & downloads
Node categories 26
BiologicalProcessCellCellLineCellularComponentCodingSequenceDiseaseDiseaseOrPhenotypicFeatureDrugExonGeneGrossAnatomicalStructureMacromolecularComplexMolecularActivityNoncodingRNAProductNucleicAcidSequenceMotifOrganismTaxonPathwayPhenotypicFeatureProteinProteinDomainProteinFamilyPublicationRegulatoryRegionSequenceVariantSmallMoleculeTranscript
Edge / predicate types 32
actively_involved_inaffectsassociated_withclose_matchderives_fromdirectly_physically_interacts_withenablesexpressed_ingene_associated_with_conditiongene_product_ofhas_adverse_eventhas_gene_producthas_parthas_participanthas_phenotypein_clinical_trials_forin_taxoninteracts_withis_sequence_variant_oflocated_inmember_ofmentionsorthologous_toparalogous_to+8 more