FAIR is the exit standard. Anything that leaves your personal Green Room into a shared or published zone must be Findable, Accessible, Interoperable, and Reusable. This SOP gives the concrete VRC action for each principle — not the theory. Tick the matching items in FAIR_self_check.md before each gate (G1 leaving the Green Room, G2 publishing).
Findable — concrete VRC steps
- Apply the metadata template at upload time, not later. Templates live in the cRDM Nextcloud share and at
github.com/CoreUnitRDM/MetaDataTemplates; attach the filled Excel/CSV to the project’sdata/metadata/folder. - Use one identifier everywhere: the same
subject_id/sample_idappears in the filename (SOP file-naming), the metadata table, and the Labfolder entry. - At publication, get a persistent identifier (DOI) via WueData (JMU institutional repository,
<FILL IN URL>) or a domain repository — never rely on a Nextcloud link as the citable handle.
Accessible — concrete VRC steps
- Store shared data in a Nextcloud group folder with role-based access (requested from the cRDM helpdesk), not in personal shares or public links.
- Authentication is via your JMU/Informatikcloud identity (SSO); MFA is mandatory for sensitive zones.
- Keep metadata discoverable even when data is restricted: deposit the descriptive record openly and route the data itself through the correct controlled channel (EGA / NFDI4Health Study Hub for human-derived; UKW for patient-identifying — see the GDPR SOP).
Interoperable — concrete VRC steps
- Convert to open standard formats as your primary data: OME-TIFF/OME-Zarr (imaging), FASTQ→BAM/CRAM/VCF (omics), CSV/Parquet over XLSX, JSON/MD over Word. Proprietary originals stay in
data/raw/untouched. - Use controlled vocabularies, not free text: MIxS/ENVO (microbiota), OME-XML/REMBI (imaging), SNOMED CT / LOINC / ICD (clinical), ChEBI/RRID (wetlab). Look up terms via the EBI OLS.
- Code missing values explicitly with standard tokens (see Metadata Template Suggestions) — never leave a cell blank.
Reusable — concrete VRC steps
- Declare a licence before publishing: data → CC BY 4.0 or CC0; code → MIT / Apache-2.0 / GPL-3.0.
- Make the pipeline reproducible:
raw → primary → processedmust regenerate fromsrc/alone, versioned in VRC GitLab (public code mirrored togithub.com/CoreUnitRDMor Zenodo after publication). - Ship documentation alongside data: README (the map), the data dictionary, and the DMP travel with the project (see the project bootstrap).
VRC tool map at a glance
| FAIR need | VRC tool |
|—|—|
| Storage + access control | cRDM Nextcloud group folders (coreunitrdm.biozentrum.uni-wuerzburg.de) |
| Documentation | Labfolder ELN (via RZ), README, data dictionary |
| Metadata templates | CoreUnitRDM MetaDataTemplates + Nextcloud metadata app |
| Code / provenance | VRC GitLab (+ GitHub/Zenodo mirror at publication) |
| Compute | JupyterHub + HPC Julia I/II / HPC Bioinformatics |
| Preservation | RZ Archive Server (10-year retention) |
| Publication + DOI | WueData / Zenodo / domain repository |
| DMP | RDMO (JMU) or DataPLAN |
Related: SOP — RDM Lifecycle · FAIR self-check · GDPR & sensitive data SOP · Metadata Template Suggestions.