Graph Schema

OptimusKG is a Labeled Property Graph where every node has the schema (id, label, properties) and every edge has the schema (from, to, label, relation, undirected, properties).

All node and edge properties include a sources struct with direct: List(String) and indirect: List(String) fields for provenance tracking. This is omitted from the individual property tables below for brevity.

Node Types

Node Type	Label	Primary Sources
Gene	`GEN`	OpenTargets, Bgee, DisGeNET
Disease	`DIS`	OpenTargets, DrugCentral, MONDO
Drug	`DRG`	OpenTargets, DrugBank, DrugCentral, OnSIDES
Phenotype	`PHE`	OpenTargets, HPO, MedDRA
Biological Process	`BPO`	GO
Anatomy	`ANA`	UBERON
Molecular Function	`MFN`	GO
Cellular Component	`CCO`	GO
Pathway	`PWY`	Reactome, OpenTargets
Exposure	`EXP`	CTD, MeSH

Node IDs use standard ontology prefixes (e.g., MONDO:, ENSG, DrugBank:, HP:, GO:, UBERON:, R-HSA-).

Node Schema

{
    "id": "MONDO:0005015",       # Unique identifier with ontology prefix
    "label": "DIS",              # 3-letter node type code
    "properties": {
        "sources": {
            "direct": ["opentargets"],
            "indirect": ["mondo"]
        },
        "name": "diabetes mellitus",
        "description": "A metabolic disease ...",
        # ... type-specific properties
    }
}

Edge Types

Edge Type	Label	Relation Types	Dir.	Primary Sources
Disease - Protein	`DIS-PRO`	ASSOCIATED_WITH	undirected	OpenTargets, DisGeNET
Anatomy - Protein	`ANA-PRO`	EXPRESSION_PRESENT, EXPRESSION_ABSENT	undirected	Bgee
Drug - Drug	`DRG-DRG`	SYNERGISTIC_INTERACTION, PARENT	directed	DrugBank, OpenTargets
Phenotype - Protein	`PHE-PRO`	ASSOCIATED_WITH	directed	OpenTargets, DisGeNET
Protein - Protein	`PRO-PRO`	INTERACTS_WITH	directed	PrimeKG
Biological Process - Protein	`BPO-PRO`	INTERACTS_WITH	undirected	OpenTargets
Disease - Phenotype	`DIS-PHE`	PHENOTYPE_PRESENT, PHENOTYPE_ABSENT	undirected	OpenTargets
Cellular Component - Protein	`CCO-PRO`	INTERACTS_WITH	undirected	OpenTargets
Molecular Function - Protein	`MFN-PRO`	INTERACTS_WITH	undirected	OpenTargets
Drug - Disease	`DRG-DIS`	INDICATION, OFF_LABEL_USE, CONTRAINDICATION	undirected	OpenTargets, DrugCentral
Drug - Protein	`DRG-PRO`	TARGET, ENZYME, INHIBITOR, AGONIST, + more	directed	DrugBank, OpenTargets
Drug - Phenotype	`DRG-PHE`	ADVERSE_DRUG_REACTION, INDICATION, CONTRAINDICATION	undirected	OnSIDES, OpenTargets, DrugCentral
Pathway - Protein	`PWY-PRO`	INTERACTS_WITH	undirected	OpenTargets, Reactome
Disease - Disease	`DIS-DIS`	PARENT	directed	OpenTargets
Anatomy - Anatomy	`ANA-ANA`	PARENT	directed	UBERON
Biological Process - Biological Process	`BPO-BPO`	IS_A	directed	GO
Cellular Component - Cellular Component	`CCO-CCO`	IS_A	directed	GO
Molecular Function - Molecular Function	`MFN-MFN`	IS_A	directed	GO
Phenotype - Phenotype	`PHE-PHE`	PARENT	directed	HPO
Pathway - Pathway	`PWY-PWY`	PARENT	directed	Reactome
Protein - Protein	`PRO-PRO`	INTERACTS_WITH	directed	PrimeKG
Exposure - Protein	`EXP-PRO`	INTERACTS_WITH	directed	CTD, OpenTargets
Exposure - Disease	`EXP-DIS`	LINKED_TO	directed	CTD, MONDO
Exposure - Exposure	`EXP-EXP`	PARENT	directed	CTD
Exposure - Biological Process	`EXP-BPO`	INTERACTS_WITH	undirected	CTD, GO
Exposure - Molecular Function	`EXP-MFN`	INTERACTS_WITH	undirected	CTD, GO
Exposure - Cellular Component	`EXP-CCO`	INTERACTS_WITH	undirected	CTD, GO

Edge Schema

{
    "from": "DrugBank:DB00945",
    "to": "MONDO:0005015",
    "label": "DRG-DIS",              # Source-Target type code
    "relation": "indication",        # Standardized relation type
    "undirected": False,
    "properties": { ... }
}

Relation Taxonomy

Relations are standardized across all data sources using a controlled vocabulary:

Category	Relations
Hierarchy	`PARENT`, `IS_A`
Association	`INTERACTS_WITH`, `ASSOCIATED_WITH`, `LINKED_TO`
Expression	`EXPRESSION_PRESENT`, `EXPRESSION_ABSENT`
Phenotype	`PHENOTYPE_PRESENT`, `PHENOTYPE_ABSENT`
Drug - Disease	`INDICATION`, `OFF_LABEL_USE`, `CONTRAINDICATION`
Drug - Phenotype	`ADVERSE_DRUG_REACTION`
Drug - Drug	`SYNERGISTIC_INTERACTION`
Drug - Protein (roles)	`TARGET`, `ENZYME`, `TRANSPORTER`, `CARRIER`
Drug - Protein (actions)	`ACTIVATOR`, `AGONIST`, `ANTAGONIST`, `INHIBITOR`, `MODULATOR`, `SUBSTRATE`, `BLOCKER`, `INVERSE_AGONIST`, `PARTIAL_AGONIST`, `DEGRADER`, and more

When multiple relations exist for the same entity pair, a priority system resolves to the most specific relation.