Research Data Management at France BioImaging

Guillaume Gay

Sophia Antipolis, April 4 2024

France BioImaging

  • 7 Nodes 10+1 Nodes
  • 18+ Imaging Core Facilities
  • 30 associated R&D teams
  • 300+ staff
  • 350+ set-ups
  • 20-25 M€ of budget/year (full cost)
  • more than 7000 users

FAIR Microscopy data management

Core issues

  • Data are tricky:
    • Too many file formats
    • Increasing data volumes
  • Metadata too:
    • Multiple instruments, techniques, methods
    • No established standard

Metadata recommendations

What is “good enough” metadata for Biological Images?

Recommended Metadata for BioImages:

REMBI metadata categories

Public Archives

RDM architecture at France BioImaging

Objectives

  • Bring microscopy data to the ☁️
  • Together with bioimage analysis 💡
  • Switch to an Open Science culture 🌍

Constrains

Infra

  • IaaS
  • Pluggable at the facility
  • Minimal local IT involvement

Data

  • FAIR
  • Close to compute

User

  • Keep raw data access
  • Opt in
  • Ease collaborations
  • Don’t repeat user inputs
  • Ease analysis worflows

BioImage Town

J. Moore et al.

iRODS, OMERO, Madbot

OMERO

madbot

  • manage ISA architecture
  • manage metadata and data links
  • listen to them in 10’ 😉

iRODS

Services architecture

  • Core services in the datacenter (Montpellier or Nantes)
  • Facility buffer / regional storage

omero-quay

Objectives

  • Interact with data management tools APIs
  • Trigger actions:
    • Create / Read / Update / Delete
    • Transport / Import
  • Synch databases ⚠️

Example route:

User experience

0. At training

  • 📖 learn
  • 👤 create account
  • 🔑 authenticate

1. Instrument ➡️ User space

  • 🔬 acquire
  • 🔑 mount network share
  • 🚌 transfer to user layer
  • 🔍 pre-analysis & curation
  • ⬇️ user download (user can get out here)

2. User space ➡️ Cooperative

  • 🎫 plan data “metro ride”
  • 🔗 link to context
  • 🚋 import into coop

3. Cooperative ↔︎️ Cooperative

  • 🔄 share
  • 👁️ Visualise
  • 🖊️ Annotate
  • 📊 Analyse (in place)

4. Cooperative ➡️ Commons

  • Easy if we agree on metadata 😉

Thank you!

France BioImaging Data Team:

  • Perrine Paul-Gilloteaux
  • Emmanuel Faure
  • Dorian Kauffmann
  • Jean-François Guillaume
  • Raphaël Braud-Mussi
  • Guillaume Maucort
  • Théo Barnouin
  • Marc Mongy

And the always supportive BioImaging & Open Science Communities 😄