GISAID

GISAID (the Global Initiative on Sharing All Influenza Data, later re-scoped beyond influenza) is a data-sharing initiative that hosts genetic sequence and associated epidemiological data for priority viral pathogens, run as a public-private partnership rather than a government agency or academic consortium. It is one of the field’s clearest examples of a repository that is free to access but not open access in the sense researchers usually mean the term — a distinction that matters when writing a data management plan or evaluating a repository against funder open-data expectations.

History

GISAID emerged from a 2006 dispute over avian-influenza data sharing: virologist Ilaria Capua refused to submit her H5N1 sequence data to a closed WHO-affiliated database and called instead for public deposit, and businessman Peter Bogner helped organize and fund what became GISAID’s founding in 2008. That origin — a reaction against restricted, closed pandemic-preparedness data — is part of why GISAID’s own access model is a recurring point of debate: it was built to be more open than what came before it, but it is still gated relative to fully public archives like GenBank.

What it hosts

GISAID operates several pathogen-specific databases under one platform: EpiFlu for influenza virus sequences (its original scope), EpiCoV for SARS-CoV-2 sequences (by volume, the largest COVID-19 sequence collection of the pandemic), and, added in 2022, EpiPox for mpox virus data and EpiRSV for respiratory syncytial virus sequences.

The Database Access Agreement

Data in GISAID is not released to the public domain. Every user must register and accept a Database Access Agreement before searching or downloading, and that agreement carries real, enforceable conditions: data cannot be redistributed to non-registered third parties; users must acknowledge the originating laboratory (which generated the specimen) and the submitting laboratory (which sequenced and deposited it) in any resulting publication; no patent may be filed on data obtained through GISAID; and users are expected to make a reasonable attempt to collaborate with the data-generating group before publishing analyses that rely heavily on their sequences. This is a fundamentally different governance model from an open repository’s public-domain or permissive-license default — it functions closer to a controlled data-sharing agreement than a download.

Why this is a live controversy, not settled practice

GISAID’s access model has drawn sustained criticism for being less transparent and less open than its founders’ stated mission implies. The re3data registry reclassified GISAID from “open access” to “restricted access” following account suspensions in March 2023, and researchers — including staff at the European Bioinformatics Institute — have argued the acknowledgment and no-redistribution requirements slow down exactly the kind of rapid, cross-institutional analysis a pandemic response depends on. For a research administrator, the practical takeaway is not to treat “hosted in GISAID” as equivalent to “openly available” when reviewing a data management plan’s repository choice or a funder’s open-data compliance claim.

Examples

• A genomic-epidemiology lab sequences a novel influenza strain from a surveillance sample and deposits it in EpiFlu, agreeing to the Database Access Agreement so other registered researchers can incorporate it into phylogenetic tracking tools such as Nextstrain.
• A pandemic-response researcher downloads thousands of EpiCoV SARS-CoV-2 sequences for a variant-tracking analysis, and because GISAID’s terms require it, credits every originating and submitting laboratory in the resulting publication’s acknowledgments rather than citing GISAID alone.

Counter-example

A dataset a funder’s open-data mandate requires to be released under an open license (e.g. CC0 or CC-BY, with no redistribution restriction) is not satisfied by deposit in GISAID alone, since GISAID’s Database Access Agreement imposes redistribution and attribution conditions that a fully open license does not — a research administrator reviewing compliance should check the specific mandate’s language before treating GISAID deposit as equivalent to open publication.