cdcdb

is a Data Source.

CDCDB (Continuous Drug Combination Database) is an automatically curated and continuously updated database of drug combinations, developed at Ben-Gurion University of the Negev as a successor to the static DCDB (last released in 2014). It collects combinations of two or more drugs from clinical trial design groups in ClinicalTrials.gov (via the AACT database), approved combination products in the FDA Orange Book, and drug patents from Clarivate Integrity, using named entity recognition against UMLS to separate drugs from free-text intervention names and mapping each drug to DrugBank and PubChem identifiers. The 2022 publication reported 40,795 drug combinations, of which 17,107 are unique combinations of 4,129 individual drugs, along with trial conditions, MeSH terms and references, Orange Book product and patent details, and patent metadata. New versions were built weekly and offered as CSV and SQLite downloads, intended for training and evaluating drug synergy prediction models; the most recent version on the website is dated 18 June 2024.

License

CC0 1.0

Homepage

cdcdb

Repository

GitHub

External IDs

No external IDs available

Product Summary

Products

From this Resource
ID Name URL Category Format Description
cdcdb.portal CDCDB Website CDCDB Graphical Interface http CDCDB website, with a preview table o...
cdcdb.csv CDCDB CSV Tables 18.06.2024.zip (67.2 MB) Product csv Latest CDCDB release (version of 18 J...
cdcdb.sqlite CDCDB SQLite Database 18.06.2024.sqlite.zip (71.5 MB) Product sqlite Latest CDCDB release (version of 18 J...
cdcdb.figshare CDCDB Figshare Collection (12 April 2022) m9.figshare.c.5375228.v1 Product mixed Figshare collection accompanying the ...
cdcdb.code CDCDB Generation Code CDCDB Process http Source code used to generate CDCDB fr...

Details

Continuous Drug Combination Database

Overview

CDCDB (Continuous Drug Combination Database) is a database of drug combinations developed at Ben-Gurion University of the Negev and described in Scientific Data (2022). It was built as a successor to DCDB, whose last version (DCDB 2.0) dates from 2014: rather than manual curation, CDCDB is generated automatically and was meant to be rebuilt weekly so it stays current.

Combinations come from three sources:

  • clinical trial design groups in ClinicalTrials.gov, via the AACT database;
  • approved combination products in the FDA Orange Book;
  • drug patents from Clarivate Integrity, with metadata from the Derwent Innovations Index.

Named entity recognition against UMLS separates drug names from free-text intervention descriptions, and each drug is mapped to DrugBank and PubChem identifiers. The 2022 paper reported 40,795 drug combinations (17,107 unique) involving 4,129 individual drugs. The data are intended for training and evaluating drug synergy prediction models.

Access

  • Website, with a preview table and downloads of current and past versions: https://icc.ise.bgu.ac.il/medical_ai/CDCDB/
  • Figshare snapshot of 12 April 2022 (CSV and SQLite, CC0): https://doi.org/10.6084/m9.figshare.c.5375228.v1
  • Generation code: https://github.com/Omer-N/CDCDB

Status

The website and downloads were reachable on 2026-11-12, but the most recent weekly version listed is dated 18 June 2024, so the database no longer appears to be updated.

Is this information incorrect or incomplete? Request an update.

Created: November 12, 2026 | Last modified: November 12, 2026