Developed at the MRC Integrative Epidemiology Unit (IEU) at the University of Bristol, this resource is a manually curated collection of complete GWAS summary datasets made available as open source files for download, or by querying a database of the complete data.

For full details, including dataset contributors and researchers who have developed the resource see our website.

Dataset history

This project began as the underlying database for the MR-Base (decommissioned) and LD Hub projects. These data now serve as an input source to a wider number of analytical tools that implement methods such as Mendelian randomization, fine mapping, colocalisation, GWAS visualisation etc. Please see the API page for a list of R and Python packages that will connect to the data.

Dataset access

Access to the resource is currently free of charge for academic researchers, subject to fair use limits to avoid overloading our services. Users can register and receive an API token directly on the website.

The database comprises mainly publicly available datasets, but also includes a number of private datesets whose access is controlled through OAuth2.0 authentication. Please contact us if you have datasets that you would like to add, whether public or private.

We have made all the public data available for download. We are using the GWAS VCF format to store the GWAS summary data to ensure alignment with the hg19 reference sequence, and to enable very fast querying. More information is available in the GWAS-VCF specification and on biorxiv.

A list of all datasets, along with their meta-data and IDs, is also available through the API, R or Python packages, so it should be straightforward to create a subset of datasets to be downloaded through a loop. Note that there will be a relatively higher cost (deducted from your periodic allowance, usually given free of charge) to do this, as we pay money to the cloud service provider for the storage, requests handling and outbound bandwidth.