Spain's germplasm banks, searchable as one.
FLAIR-GG connects plant germplasm banks and related biodiversity resources into a single FAIR network. Each bank keeps its own data and its own systems. Together, the network can be searched, queried and analysed as a whole, by researchers and by AI assistants.
One network for genetic resources, with each bank's data kept at its source.
FLAIR-GG began by making the César Gómez Campo Banco de Germoplasma Vegetal (BGV-UPM) FAIR. It now links germplasm collections across Spain with the wider ecosystem of taxonomic, conservation and molecular resources. Data never has to be pooled centrally: every member publishes its own standards-based description, and the network does the rest.
Findable by design
Every bank describes itself, its collections and its services in DCAT, published from its own FAIR Data Point. A central index harvests these descriptions, so a new member becomes searchable as soon as it registers.
A shared data model
Passport data is expressed with an ontology based on the FAO Multi-Crop Passport Descriptors. Banks export plain CSV from whatever system they already use, and the toolkit turns it into linked data.
Federated, not centralised
Because every bank uses the same model, one query can go to all sites at once. Each bank stays the authoritative, up-to-date source for its own records and decides what it shares.
Cámara Ballesteros, A., Aguayo Jara, E., Verykaki, E. S., Pastor del Olmo, G., Moreno Vázquez, S., Torres, E. & Wilkinson, M. D. The FLAIR-GG federated network of FAIR germplasm data resources. Scientific Data 11, 1386 (2024). doi:10.1038/s41597-024-04243-7
A growing national network of germplasm banks.
Ten germplasm banks have joined FLAIR-GG, about one in five of Spain's plant germplasm banks. Six already publish live FAIR endpoints, and four more are setting theirs up now.
From a spreadsheet export to a federated query, in four steps.
No bank needs a database expert or a new system. The pieces below are packaged together as FLAIR-in-a-Box, built on the open-source Sextans Suite.
Describe
Publish who you are and what you hold as DCAT metadata in your own FAIR Data Point (Sextans Sight). Or simply put a DCAT file on your website and let our proxy import it.
MetadataPublish
Export admin.csv, germplasm.csv and location.csv from your
existing system. Sextans Fix maps them onto the FLAIR-GG model as RDF.
Index
The FLAIR-GG FDP Index harvests every member's metadata and keeps track of which data services each one offers, as members come, go or move.
AutomaticQuery & analyse
The Virtual Platform searches the whole network, runs data services against every provider at once, and passes the results to ready-made analysis notebooks.
Virtual PlatformAsk the whole network a question, then analyse the answers.
The FLAIR-GG Virtual Platform (VP) is the shared portal on top of the network. It works only from the live index, so anything a member publishes appears without central curation.
-
Network-wide searchKeyword and ontology-term search across every bank's catalogues, datasets and services.
-
Federated data servicesRun one service type, for example collecting-site coordinates by species or IUCN categorisation, against every provider in a single call.
-
Results straight into analysisService results open in the FLAIR-GG Analytics Jupyter notebooks, which run in the browser with nothing to install.
-
Built for scaleTested with 3,700 datasets, more than the number of germplasm banks on Earth.
AI assistants can query the network directly.
The Virtual Platform runs a Model Context Protocol (MCP) server, the open standard that lets AI assistants such as Claude call external tools. A researcher can ask a question in plain language. The assistant then queries the live FLAIR-GG network itself and answers from authoritative, up-to-date records, not from what it remembers.
Finds catalogues, datasets and services anywhere in the network whose metadata mentions a term, using the same search that drives the VP's own interface.
Read-only SPARQL over the FDP Index for questions a keyword can't answer: counting, grouping, filtering, and "who curates this collection, and how do I contact them?"
Runs the IUCN Red List categorisation service against every live provider and returns each provider's results, so the assistant can compare conservation status across banks.
Questions researchers can simply ask
Designed to grow with the network
- One tool per service type. Every provider of a given service type implements the same interface, so one tool covers them all, and adding a provider requires no change to the tool.
- Providers are discovered live. Tools find their providers through the VP's own service discovery each time they run, so new and relocated banks are picked up automatically.
- Grounded in curated analyses. Each service-type tool points the assistant to the reference Jupyter notebook for that data, so it reproduces the established analysis instead of inventing its own.
- Native to the platform. The protocol is implemented directly in the VP's own codebase, with each tool defined in a single file. Adding a tool is a small, reviewable change.
Choose how far you want to go. We'll help with every step.
Joining never requires changes to your database or existing infrastructure. Start with a simple description of your bank and add more whenever you're ready. The onboarding guides are available in English and Spanish.
Describe your bank
Fill in our metadata template, or publish a DCAT description on your existing website. The FLAIR-GG proxy imports it and your bank becomes discoverable across the network.
Run your own node
Install FLAIR-in-a-Box: Sextans Sight serves your metadata, and Sextans Fix (optional) publishes your germplasm data from CSV exports using the FLAIR-GG data model.
Let us host it
Not ready to run infrastructure? We can host your FAIR endpoint temporarily, so you can see the benefits of membership before you commit any effort.
Talk to us. We'll walk you through it. It's easier than it looks.
Everything is open.
All FLAIR-GG code, models and configuration are open source on GitHub.
Install & onboard
Data model
Common questions from germplasm banks.
Read all the answers on the FAQ page.