het.io

What Is It

Hetionet is a heterogeneous information network focused on the biomedical domain. It encodes millions of biomedical research relationships accumulated over the past 50 years into a unified resource, integrating data from 29 public databases. It contains 47,031 nodes across 11 types, such as genes, compounds, and diseases, as well as 2,250,197 edges across 24 types.

Core Dimension Analysis

• Features and Use Cases: Its core purpose is to help scientists and biologists generate new hypotheses and predictions by connecting existing biomedical data across multiple levels and entity types. It was originally used in Project Rephetio to predict new uses for existing drugs, i.e. drug repurposing, and has since expanded to broader applications such as prioritizing disease-associated genes.
• Open Source and Self-Hosting: Hetionet is fully open source and free to use, subject to upstream data licenses, and supports self-hosting. The project is developed publicly on GitHub (@hetio, @dhimmel) and accepts community feedback.
• Integrations and Ecosystem: It deeply integrates authoritative databases such as Entrez Gene, DrugBank, Disease Ontology, and Gene Ontology. Nodes are mapped using standardized vocabularies, while edges are cleaned and deduplicated to ensure data accuracy and consistency.
• Supported Languages/Frameworks and API/SDK: The text does not explicitly mention support for specific programming languages or an official API/SDK. Developers need to build on the GitHub repository themselves.

Pricing

Completely free and open source.

Pros and Cons

Pros: Highly integrated data, reducing the inconvenience and errors that come from using multiple separate databases; the heterogeneous network model can precisely distinguish different entities and mechanisms, such as binding versus upregulation; strong academic foundation, supported by peer-reviewed publications including eLife.
Cons: High technical barrier, requiring knowledge of graph networks and bioinformatics; data updates depend on upstream public databases; limited to biomedical use cases.

Who Is It For

Suitable for computational biologists, drug discovery professionals, biomedical researchers, and developers working on applications of graph neural networks in healthcare.

Access in China

The text does not mention any network access restrictions. As an academic open-source project, its code is hosted on GitHub, which usually requires a proxy for access from mainland China. However, the dataset may be available through other academic mirrors. The specific network accessibility is unknown. There is no payment barrier for the project.