SSBD Ontology (v2025‑05)

Lead developer: Yuki Yamagata (RIKEN R-IH and RIKEN BRC)

SSBD Ontology: A Two Tier Approach for Interoperable Bioimaging Metadata Repository Layer (rapid deposition) & Added‑Value Layer (curated rich annotations)

　Bioimaging data produced by modern microscopy are expanding at an unprecedented pace, yet their scientific value is often constrained by fragmented metadata. We present the SSBD Ontology, a two-layer semantic model that reconciles rapid publication with ontology-aligned curation. A lightweight repository layer captures only the indispensable descriptors required for DOI assignment, while an added-value layer enriches the same instances with biological context—organism, strain, cell type, anatomy and GO terms—as well as imaging-method and instrument semantics.

The ontology bridges the instance-centric nature of imaging repositories and the class-centric structure of existing OBO ontologies, enabling seamless integration into knowledge graphs. A single SPARQL query retrieves, across modalities, all datasets of C57BL/6J mouse brain recorded by AMATERAS light microscopy and FIB-SEM electron microscopy, returning direct OME-Zarr and Vizarr links in sub-second time.

SSBD Ontology (OWL DL), exemplar instances and conversion scripts are released under CC-BY 4.0 on GitHub and will be mirrored on BioPortal, strengthening the FAIR data ecosystem for bioimaging research.

---

1. Ontology 

Item	URL / File
Integrated ontology (OWL/RDF/XML)	ontology/ssbd_integrated.owl
Core ontology (OWL/RDF‑XML)	ontology/ssbd_core.owl
All individuals (TTL)	instance/ssbd_biosample.ttl

1.1 Core layer:Seven key entity types

Layer	Entity class	Typical properties	Linked external vocab
Repository	SSBD_Project	has_project_name, has_dataset_output (subClassOf RO:0002234) (→ Dataset)	—
Repository	SSBD_dataset	has_biosample_information, has_ome_zarr_information (→ OME-NGFF-ZARR)	—
Added	SSBD_OME_NGFF_ZARR	has_s3_endpoint, has_vizarr_url, sizes	—
Added	SSBD_biosample_information	is_about_organism/strain/cell/anatomy/GO*	NCBITaxon, CL, UBERON, GO
Added	SSBD_imaging_method_information	has_detection_method, has_imaging_method_recorded_type	FBbi
Added	SSBD_imaging_instruments	has_component (objective, detector …)	—
Added	SSBD_dimension_data	x/y/z/t scale + unit	IAO / UO

Seven core entities form a two—tier model: project, Dataset, Biosample, Imaging Method, Instrument, Dimension, and OME-NGFF metadata. We distinguish between the data required for rapid publication in the repository tier and the added-value database tier in SSBD. Within the repository, the essential information comprises a Project, a Dataset, and bibliographic and author (person) information, thereby ensuring the minimum metadata needed for rapid publication. The added-value tier delivers deep, ontology-aligned curation while reusing external OBO vocabularies. SSBD ontology covers entities related to imaging method(e.g., imaging method, imaging device, image dimension, the storage URI of the dataset).

1.2 Example instance relationships (Project 199 – AMATERAS brain‑slice)

Project → Dataset → Biosample → OME‑Zarr relations Project 199 (Ichimura): A single Dataset, its Biosample and the associated OME-Zarr metadata are connected via RO relations; external strain and organism terms are linked for immediate cross-repository interoperability.

1.3 Quick SPARQL tutorial  🔍

The SSBD graph can be queried live at

https://knowledge.brc.riken.jp/bioresource/sparql

The endpoint supports HTML, JSON, CSV, TSV and RDF results. If you are new to SPARQL, start with the two ready‑made examples in sparql/README.md – each file contains:

• a short natural‑language question
• the corresponding .rq query
• a clickable link that opens the query in the endpoint UI

After running the query you can change Output → Download as to JSON / CSV, etc.

1.4 Sample SPARQL query

PREFIX ssbd: <http://ssbd.riken.jp/ontology/>
PREFIX rdfs: <http://www.w3.org/2000/01/rdf-schema#>
PREFIX obo:  <http://purl.obolibrary.org/obo/>

SELECT ?dataset ?bs ?title ?methodIRI ?methodLabel ?zarr ?vizarr
WHERE {
  # --- strain filter (C57BL/6J) ---
  ?bs ssbd:is_about_strain <https://www.jax.org/strain/000664> .

  # --- Dataset ↔ Biosample ---
  ?dataset ssbd:has_biosample_information ?bs ;
           ssbd:has_dataset_title         ?title .
  
    # --- Imaging method  ---
  ?dataset ssbd:has_imaging_method_total_info  ?imNode .
  ?imNode  ssbd:has_imaging_method_recorded_type ?methodIRI .
  OPTIONAL { ?methodIRI rdfs:label ?methodLabel }
  
   # --- NGFF ---
  OPTIONAL {
    ?dataset ssbd:has_ome_zarr_information ?ngff .
    OPTIONAL { ?ngff ssbd:has_s3_endpoint ?zarr }
    OPTIONAL { ?ngff ssbd:has_vizarr_url  ?vizarr }
  }

}
ORDER BY ?methodIRI ?dataset

A single query retrieves all strain C57BL/6J datasets, together with the imaging-method sub-hierarchical tree, OME-Zarr URLs, and Vizarr viewer.

1.5 Related resources and gap analysis of ontology usage

• foundingGIDE Deliverable D6.1 – Comparative analysis of ontology usage in SSBD, IDR, BIA and other imaging repositories, plus a field‑by‑field mapping from REMBI to SSBD metadata. • Zenodo DOI: 10.5281/zenodo.15553217 • Tables 1 (REMBI → SSBD) and 2 (Ontology coverage) are the authoritative reference; we therefore link them here rather than duplicating the full spreadsheets in this repository.