ikraph

is a Knowledge Graph.

A large-scale biomedical knowledge graph assembled from PubMed abstracts, containing over 22 million entities and 120 million relations.

Domains

health, biomedical, drug discovery, translational, genomics

License

CC-BY-4.0

Homepage

ikraph

Repository

GitHub

Infores ID

Unknown

FAIRsharing ID

Unknown

Product Summary

Contacts

Products

From this Resource
ID Name URL Category Format Description
ikraph.site BioKDE biokde.insilicom.com GraphicalInterface http Biomedical Knowledge Discovery Engine...
ikraph.code iKraph Code iKraph ProcessProduct Code for named entity recognition, re...
ikraph.graph iKraph graph metadata data.tar.gz?download=1 (58.3 MB) GraphProduct json Graph metadata for iKraph, including ...
ikraph.graphdata iKraph graph data iKraph_full.tar.gz?download=1 (1.3 GB) GraphProduct Complete graph data for iKraph with a...

KG-Registry Curators

Details

iKraph Knowledge Graph

iKraph is a comprehensive large-scale biomedical knowledge graph developed by Insilicom for AI-powered data-driven biomedical research. It represents one of the largest structured biomedical knowledge resources assembled from literature mining.

Overview

iKraph was constructed by applying advanced natural language processing and relation extraction techniques to the entire corpus of PubMed abstracts. The knowledge graph contains over 22 million biomedical entities and 120 million relations, covering a wide range of biomedical concepts including genes, proteins, diseases, drugs, pathways, and phenotypes.

The primary goal of iKraph is to enable knowledge discovery and hypothesis generation for biomedical research and drug development. It integrates information across multiple domains to support various applications:

  • Drug repurposing and discovery
  • Target identification
  • Biomarker discovery
  • Disease mechanism exploration
  • Precision medicine approaches

Technical Details

iKraph employs a sophisticated named entity recognition and relation extraction pipeline to process biomedical literature at scale. The pipeline includes:

  1. Named entity recognition for biomedical concepts
  2. Relation extraction between identified entities
  3. Knowledge graph construction and normalization
  4. Integration with other biomedical resources
  5. Quality assurance and validation

The knowledge graph is accessible through the BioKDE (Biomedical Knowledge Discovery Engine) web interface, which provides search, visualization, and exploration capabilities for researchers.

Applications

iKraph has been successfully applied to several biomedical research areas:

  • Identification of novel drug candidates for various diseases
  • Discovery of previously unknown gene-disease associations
  • Understanding of complex disease mechanisms
  • Prediction of drug-drug interactions
  • Personalized treatment recommendations

Citation

If you use iKraph in your research, please cite:

Zhang Y, Sui X, Pan F, et al. A comprehensive large-scale biomedical knowledge graph for AI-powered data-driven biomedical research. Nature Machine Intelligence. 2025. https://doi.org/10.1038/s42256-025-01014-w

Automated Evaluation

Is this information incorrect or incomplete? Request an update.

Created: July 22, 2025 | Last modified: July 22, 2025