One bridge, 14 ELN/LIMS systems — plus the entire open-source materials stack around it
OCO does not bridge each ELN/LIMS vendor system individually but bridges the ELN-Filetype standard — the RO-Crate-based cross-vendor export currently adopted by 14 ELN/LIMS systems. One bridge covers the whole pool. In addition, the L0 layer connects OCO directly to the open-source materials stack that is the de-facto backbone of the MSE community: PMDco, EMMO-Crystallography, BattINFO, CHMO, OBI, schema.org, RO-Crate, Materials Project, OPTIMADE, Croissant.
What you get
- Cross-vendor bridge instead of 14 individual mappings: connection via the ELN-Filetype standard (RO-Crate-based) — whatever you export today from eLabFTW, Chemotion, or Kadi4Mat lands as ABox-ready material in OCO.
- L1 as directly importable LIMS/ELN baseline: sample, equipment, measurement, identifier, provenance — the material-agnostic layer of the architecture, released under CC-BY-SA 4.0, no more class double-modelling.
- SIPOC fragment ABox model: OCO does not expect every measurement position to be complete. Real lab data and literature extraction are never complete — the SIPOC granularity plus the three-tier identifier hierarchy (mandatory / recommended / optional) accepts this rather than rejecting incomplete data.
- Reuse-before-invention discipline: in the most recent interop wave, 18 candidate properties were reduced to 0 and 15 candidate classes to 5 because they already existed in the open stack. You do not pull anything new into the stack that already lives somewhere else.
The 14-system bridge via the ELN-Filetype standard
The ELN-Filetype standard (maintained by The ELN Consortium) is an RO-Crate-based export format establishing cross-vendor interoperability for electronic lab notebooks and lab information management systems. OCO bridges to this format once — and thereby covers the entire adopters pool without maintaining a per-vendor mapping file.
As of 2026: 14 productive ELN/LIMS systems adopt the standard. Notable adopters:
| System | Origin / focus | Adopter status |
|---|---|---|
| eLabFTW | Open-source ELN, Pasteur Institute / widely deployed in EU research | RO-Crate export productive |
| Chemotion ELN | NFDI4Chem backbone, KIT, chemistry-focused | RO-Crate export productive |
| openBIS | ETH Zurich, broadly deployed in life sciences and materials | RO-Crate export productive |
| Kadi4Mat | KIT, materials-engineering-focused, part of NFDI-MatWerk | RO-Crate export productive |
| PASTA-ELN | FZJ, materials-science-oriented | RO-Crate export productive |
| Herbie | RWTH Aachen / IEHK, steel / materials data | RO-Crate export productive |
| LabIMotion | NFDI4Chem, ELN add-on | RO-Crate export productive |
| Sample DB | PTB, measurement-centric ELN variant | RO-Crate export productive |
| … and others in the current ELN-Filetype consortium (total: 14 systems productive). | ||
Instead of 14 vendor mappings, OCO has one bridge, plus a thin application profile that bundles the typically-needed modules (sample, equipment, measurement, process, identifier, provenance, format), without the maintenance load of an extracted subset distribution.
L1 as the LIMS/ELN baseline
The material-agnostic L1 layer is exactly the vocabulary that already exists in every LIMS/ELN installation — only this time as a shared definition rather than a per-project rebuilt copy. As of v0.94:
| L1 module | Content | Bridge targets |
|---|---|---|
| oco-sample | Sample, sampling, sample state, sample treatment | PMDco, OBI |
| oco-equipment | Lab equipment, measurement devices, facilities, calibration status | PMDco, CHMO, OBI, schema.org |
| oco-measurement | Measurement, measurement workflow, parameters, result | PMDco, QUDT (units), PROV-O |
| oco-identifier | Identifier hierarchy (mandatory / recommended / optional), external IDs | schema.org, ROR, ORCID |
| oco-investigation | Investigation / Study / Assay structure, actors, roles | PROV-O, PMDco, OBI |
| oco-process | Process steps, process inputs/outputs, process provenance | PMDco, PROV-O |
These L1 modules are publicly available under CC-BY-SA 4.0 (see Ontology Architecture for the licence matrix). They can be imported directly into a LIMS/ELN workflow without pulling the material-specific L2 depth.
The open-source materials stack — what’s wired in
“One bridge, 14 systems” is the ELN side. The tooling and data world around it is wired in as an L0 bridge topology too. Two layers, both productive:
PMD consortium and materials ontologies
| Standard / tool | Domain | Bridge status in OCO |
|---|---|---|
| PMDco | Platform MaterialDigital Core Ontology (mid-level, BFO-aligned) | full section, primary L0 anchor |
| EMMO-Crystallography | European Materials & Modelling Ontology, crystallography sub-module | own bridge section, sub-module versionable |
| BattINFO | Battery materials ontology | bridge to all classes that recur in the ceramics domain |
| EMMO Chemistry / Materials | EMMO sub-modules Chemistry, Materials | own bridge sections per sub-module |
| EMMO-ISQ | ISO/IEC 80000 Quantity vocabulary (93 mappings) | fully mapped |
| CHMO | Chemical Methods Ontology | equipment anchor bridge |
| ChEBI | Chemical Entities of Biological Interest | substance anchor bridge |
| OBI | Ontology for Biomedical Investigations | investigation anchor bridge |
| MADO / MWO | Materials Acquisition Description Ontology / Materials Workflow Ontology | own bridge sections |
| KnowNow | PMD predecessor ontology | 14 process / property mappings (LTCC specifics deliberately excluded) |
| SmaDi | Smart-Materials-Discovery ontology | 15 mappings, piezo-ceramics subset (shape memory deliberately excluded) |
| Mieller-Ferrit | Ferrite-specific PMD ontology | bridge preparation for ferrite pilot |
FAIR / data-infrastructure stack
| Standard / tool | Role | OCO use |
|---|---|---|
| RO-Crate | Research Object Crate format (FAIR data packaging) | ELN-Filetype is based on RO-Crate; anchor in oco-investigation |
| Materials Project | DFT data corpus, LBNL | external cache of ~155,000 entries (layer 2 Energy/DFT) |
| OPTIMADE | Cross-database API for material structure search | lookup bridge for structure search API fields |
| Croissant (MLCommons) | ML-ready dataset metadata standard | bridge for ML application view over materials data |
| QUDT | Quantity/Unit/Dimension/Type ontology | L0 anchor for all physical units and measurement parameters |
| PROV-O | W3C Provenance Ontology | L0 anchor for any provenance annotation |
| schema.org | structured-data web standard | equipment, identifier, organisation bridge |
| ROR / ORCID | Research Organisation Registry / Open Researcher ID | actor identifier bridge |
| NFDIcore | NFDI Common Ontology | cross-NFDI consortium bridge |
829 explicit cross-ontology mappings across 40 sections are maintained in bridge_mappings.yaml and mwo_mappings.yaml as single sources of truth. Each bridge section is independently versionable — if PMDco bumps a version, exactly one bridge file changes.
SIPOC fragment ABox model — reality tolerance instead of schema tyranny
Real experiment data is never complete. With own experiments completeness can at least be aspired to; with published literature and patents, incompleteness is often deliberate (competitive or patent strategy). A schema that mandates complete fields would either reject most real-world data or force placeholder fabrication — both unacceptable.
OCO ingests ABox data as SIPOC fragments (Suppliers / Inputs / Process / Outputs / Customers). The SIPOC granularity combined with the three-tier identifier hierarchy (mandatory / recommended / optional) accommodates incomplete data without breaking the schema. This makes OCO practically usable for the two dominant data sources — own lab data and literature extraction.
Concretely: a sintering experiment batch does not need to have all 47 possible parameters filled in. A SIPOC fragment with a process step + two output identifiers is enough for a valid ABox position. The missing fields are not imputed, not hallucinated, but annotated as “not recorded”.
Reuse-before-invention — the discipline in numbers
Before any new class or property is modelled into OCO, a checklist runs:
- Is there already a fitting OCO class?
- External identifier? → sub-class of
oco-identifier:Identifier. - Provenance? → PROV-O.
- Unit? → QUDT
QuantityValue.
In the most recent interop wave this checklist reduced 18 candidate properties to 0 and 15 candidate classes to 5. This is not austerity — this is avoidance of pseudo-innovation that would later cause cross-mapping work.
In practice that means: if you pull OCO into a LIMS/ELN pipeline, the likelihood is high that every class you already know (sample, sensor, measurement, person) already has its OCO bridge to the standard you work with.
The PMD consortium context — why OCO ends the duplication
In the PMD consortium (Platform MaterialDigital) more than a dozen projects are active, each with its own material-specific ontology: KupferDigital, GlasDigital, StahlDigital, DiProMag, iBain, KnowNow, Mieller-Ferrit, SmaDi and others. The problem: only a fraction of the modelled content is actually material-specific (the material class hierarchy). The largest part — workflow provenance, equipment, methods, identifier schemes — is material-agnostic and shareable.
OCO is built to end this duplication structurally: L0+L1 is the shared vocabulary, L2 is the material-specific layer that sister projects can replace (polymers, metallurgy, batteries) without re-modelling L0+L1. A polymer L2 replaces the ceramics L2 — the equipment, sample, provenance, identifier classes stay unchanged and shared.
Example competency questions
Six of the 163 published CQs that are particularly relevant for LIMS/ELN integrators. Tag column = OCO module that answers the question.
Which ELN-Filetype fields can be mapped automatically to OCO-L1, and which need to be enriched per vendor?
oco_eln_profile · bridge/eln_filetypeWhat is the minimal-valid SIPOC fragment template for a sintering experiment (stage 1: mandatory only)?
oco-process · oco-identifierWhich PMDco class corresponds to “sample” in our Chemotion ELN export?
oco-sample · PMDco bridgeWhich provenance chain documents batch X from powder delivery to sintered specimen?
oco-investigation · PROV-O · executable SPARQLWhich equipment anchors apply to a differential scanning calorimeter from CHMO and PMDco?
oco-equipment · CHMO bridge · PMDco bridgeWhich QUDT unit applies to the measurement position “Curie temperature”, and how is the uncertainty annotated?
oco-measurement · QUDT bridgeRelation to the OCO distribution
The LIMS/ELN baseline — L1 skeletons (sample, equipment, measurement, identifier, process, investigation) plus tensor roots, role individuals, and cross-axioms — is publicly available under CC-BY-SA 4.0. The L0 bridges to PMDco, CHMO, OBI, schema.org, QUDT, PROV-O, ELN-Filetype, RO-Crate, Croissant, OPTIMADE are freely available under CC-BY 4.0. oco-supplier, the material-detail L2, and the compliance modules are proprietary — but you can run the full LIMS/ELN stack without them. The bundle “LIMS/ELN-Ready” (L0+L1 without supplier) is downloadable with registration. → Distribution & licence architecture