Computational Biology Sources

38 public databases — one loader, one command, full provenance

37
Data Sources
50M+
Claims
13M+
Entities
8
Categories
Hetionet
Heterogeneous network integrating 47K+ nodes and 2.25M edges across genes, diseases, compounds, and anatomy.
Foundation
Reactome
Curated pathway database covering 2,600+ human pathways with reaction-level detail and cross-species projection.
Foundation
DISEASES
Gene-disease associations from JensenLab. Integrates text mining, experiments, and curated knowledge.
Gene-Disease
DisGeNET
Comprehensive gene-disease and variant-disease associations from curated repositories, GWAS, and literature.
Gene-Disease
CTD Gene-Disease
Curated gene-disease relationships from the Comparative Toxicogenomics Database with evidence codes.
Gene-Disease
ClinVar
NCBI archive of genomic variant clinical significance. Pathogenicity assertions from submitting labs.
Gene-Disease
Open Targets
Systematic drug target identification platform. Aggregates genetic, genomic, and chemical evidence scores.
Gene-Disease
STRING
Functional protein association network. Combines experimental, coexpression, text mining, and genomic context.
Protein
BioGRID
Curated protein and genetic interactions from literature and high-throughput datasets across model organisms.
Protein
IntAct
Molecular interaction database from EBI. IMEx-compliant, literature-curated binary and complex interactions.
Protein
PharmGKB
Pharmacogenomics knowledge base. Curated drug-gene-variant clinical annotations and dosing guidelines.
Protein
DrugBank
Comprehensive drug database with detailed pharmacology, targets, enzymes, carriers, and transporters.
Drug-Target
ChEMBL
Bioactivity database from EBI. Millions of compound-target binding assay results from medicinal chemistry literature.
Drug-Target
CTD Chem-Gene
Curated chemical-gene interactions from CTD. Includes interaction types: increases, decreases, affects expression.
Drug-Target
STITCH
Chemical-protein interaction network. Sister database to STRING, integrating experimental and predicted binding data.
Drug-Target
DGIdb
Drug Gene Interaction Database. Aggregates drug-gene interactions and druggable gene categories from 30+ sources.
Drug-Target
BindingDB
Measured binding affinities (Ki, Kd, IC50) for drug-like molecules and their protein targets.
Drug-Target
HPO
Human Phenotype Ontology. Standardized vocabulary for phenotypic abnormalities with disease-phenotype annotations.
Disease
TISSUES
Tissue expression data from JensenLab. Integrates text mining, experiments, and curated knowledge for tissue localization.
Disease
CTD Chem-Disease
Curated chemical-disease relationships from CTD. Distinguishes therapeutic, marker/mechanism, and toxicity evidence.
Disease
SIDER
Side Effect Resource. Drug adverse reaction data mined from FDA labels with frequency classifications.
Disease
TTD
Therapeutic Target Database. Drug-target pairs with clinical development status and therapeutic classifications.
Disease
GWAS Catalog
NHGRI-EBI catalog of published genome-wide association studies. SNP-trait associations with p-values and effect sizes.
Disease
ClinicalTrials.gov
Registry of clinical studies. Extracts intervention-condition relationships with trial phase and recruitment status.
Disease
Mondo
Unified disease ontology merging OMIM, Orphanet, DOID, and others into a harmonized classification.
Ontology
Disease Ontology
Standardized disease classification with cross-references to ICD, OMIM, MeSH, and NCI Thesaurus.
Ontology
MeSH
NLM Medical Subject Headings. Controlled vocabulary for PubMed indexing with hierarchical descriptor trees.
Ontology
ChEBI
Chemical Entities of Biological Interest. Ontology of small molecules, metabolites, and natural products.
Ontology
OMIM
Online Mendelian Inheritance in Man. Authoritative catalog of human genes and genetic disorders.
Ontology
UniProt
Universal protein resource. Sequence, function, and cross-reference annotations for reviewed (Swiss-Prot) entries.
Ontology
PrimeKG
Precision Medicine Knowledge Graph. Multimodal KG linking diseases, drugs, proteins, pathways, and exposures.
Composite KG
DRKG
Drug Repurposing Knowledge Graph. 5.87M triplets across genes, compounds, diseases, side effects, and pathways.
Composite KG
RTX-KG2c
Canonicalized biomedical KG from NCATS Translator. Integrates 70+ source databases into unified Biolink model.
Composite KG
Monarch KG
Genotype-phenotype knowledge graph. Links human diseases to model organism phenotypes for rare disease research.
Composite KG
PharMeBINet
Pharmacological and biomedical network. 2.87M nodes, 15.88M edges across 66 labels and 208 edge types.
Composite KG
GOA
Gene Ontology Annotation. Maps genes to GO terms for biological process, molecular function, and cellular component.
Literature
SemMedDB
Semantic MEDLINE Database. NLP-extracted subject-predicate-object triples from all PubMed abstracts.
Literature
No sources match your search.

Build a Reference Database

Load all 37 sources into a single LMDB-backed Attest database. Every claim carries full provenance — which source, which file, which line. 

from attestdb import AttestDB
from attestdb.infrastructure.bulk_loader import BulkLoader

db = AttestDB("reference.attest")
loader = BulkLoader(db)
loader.build(
    dest_dir="/data/reference",
    sources={
        "hetionet", "string", "reactome", "drugbank",
        "chembl", "clinvar", "disgenet", "biogrid",
        "intact", "pharmgkb", "hpo", "sider",
        "primekg", "drkg", "monarch", "pharmebinet",
    }
)

Each source downloads its own data (streaming where possible), normalizes entities using the locked normalization pipeline (NFKD → lowercase → whitespace collapse → Greek letter expansion), and inserts claims at 1.3M/sec via the Rust engine.

After building, query across all sources with full provenance: 

frame = db.query("BRCA1", depth=2)
for rel in frame.focal_entity.direct_relationships:
    print(f"{rel.predicate}: {rel.target} (source: {rel.source_id})")