Skip to main content

HuggingFace datasets

Selected GWAS Catalog datasets are available on HuggingFace for bulk analysis, machine learning and data-processing workflows.

View GWAS Catalog datasets on HuggingFace

The datasets are converted to Parquet, can be explored using HuggingFace Data Studio, and can be loaded using the standard HuggingFace datasets library.

The current release includes:

  • Associations
  • Studies
  • Samples

Pilot service​

The HuggingFace datasets are currently a pilot service. The initial datasets were released in 2026, with future releases planned periodically while usage is evaluated.

Check the HuggingFace dataset pages for the date of the latest release.

If you use these datasets or have feedback, contact gwas-info@ebi.ac.uk.