These scenarios show how LAMBDA-BER metadata, RO-Crate packaging, local query stores, and agent-assisted workflows can turn facility outputs into reusable, federated research objects without forcing every site into one storage system.
The shared pattern
The use cases are not six unrelated demos. They trace one practical architecture: facilities expose typed metadata and stable file pointers; labs and compute environments ingest those records into local tables; agents use the shared schema to query, route processing, integrate results, package depositions, and decide what deserves reanalysis.
RO-Crate gives the portable graph and bundle. LAMBDA-BER gives the structural biology projection that is easy to validate, query, and join.
Use-case catalogue
Each page is a concrete workflow sketch: who is doing the work, what metadata crosses the boundary, what the agent does, and what durable record is left behind.
Register beamtime from facility storage as searchable local records: file pointers, checksums, samples, instruments, and experiment runs.
02Turn natural-language questions into joins across your local holdings and public structural biology references such as the PDB.
03Use typed technique metadata to route collected runs to the right processing engines and record each step as workflow provenance.
04Bring cryo-EM, SAXS, and X-ray results from different facilities into one integrative structural biology workspace.
05Package finished results as self-describing crates, deposit public artifacts, and land metadata in a lakehouse for reuse.
06Find exactly which old results are affected by a new tool, reference, or method, then rerun only the impacted records.
End-to-end path
Pull a crate from a facility and record stable identifiers, sample metadata, and file locations.
Project the graph into local tables so scientists and agents can ask operational questions.
Route by technique and capture every computational step as explicit workflow provenance.
Join outputs across techniques, facilities, archives, and local experiments.
Deposit, index, and revisit results when upstream tools or references improve.
Who benefits
Expose enough structured context for outside users to find, validate, and act on data without copying whole storage systems.
Build local, queryable workspaces that preserve provenance while letting agents assemble the right files and workflows.
Land interoperable metadata in BRIDGE, a lakehouse, or a repository while keeping raw data federated by design.