Downloads

Versioned snapshots of the Preclinical Database

About the bundles

Each release bundles the primary-source TSVs (pcdb_complete.tsv, the three entity catalogs, and unmapped extractions), a spreadsheet view (preclinical_database.xlsx), plus a README.md with loading examples for pandas, polars, DuckDB, and R, and a CHANGELOG.md. Verify integrity with the SHA256 checksum listed below.

v1.0

released 2026-07-23 · 134,565 articles · 186,628 records

Initial public release of the Preclinical Database, comprising 134,565 PubMed Central full-text articles and covering 2,602 diseases, 11,753 drug compounds, and 100 animal species. AI-powered curation combined high-confidence screening for in vivo preclinical studies with LLM-based extraction and ontology-guided standardization of disease, drug, and animal entities using MeSH, Disease Ontology, PubChem, DrugBank, and UMLS.

preclinical-db-v1.0.zip
ZIP containing TSVs (pcdb_complete, entity catalogs, unmapped extractions), xlsx, README, CHANGELOG44.53 MBsha256 b9dfdefde0f07f64…
Download

Citation & license

Provided for academic research. If a snapshot supports your work, please acknowledge the Preclinical Database — a formal citation will appear here once our paper is published.