The DOSSIER_DOCS sheet in EFSA OpenFoodTox 3.0 (Zenodo record 19388272, DOI 10.5281/zenodo.19388272) is the relationship table linking dossiers to their constituent documents (substance records, endpoint study records, endpoint summaries, flexible summaries, flexible records). 106409 rows × 5 cols (DOSSIER UUID, DOCUMENT TYPE, DOCUMENT SUBTYPE, DOCUMENT UUID).
Document type distribution
| Document Type | Count | Share |
|---|---|---|
| ENDPOINT_STUDY_RECORD | 50556 | 47.5% |
| ENDPOINT_SUMMARY | 23670 | 22.2% |
| FLEXIBLE_SUMMARY | 20468 | 19.2% |
| SUBSTANCE | 11613 | 10.9% |
| FLEXIBLE_RECORD | 101 | 0.1% |
| Total document links | 106409 | 100% |
SUBSTANCE linkage key statistics
| Statistic | Count |
|---|---|
| Unique substances in SUB sheet | 7880 |
| Substances linked to a dossier (via DOSSIER_DOCS type=SUBSTANCE) | 5620 |
| Orphan substances (in SUB sheet, not linked to any dossier) | 2260 |
| Link rate (substances with dossiers) | 71.3% |
| Unique dossiers | 11613 |
| Substance-to-dossier cardinality | 1-to-1 strict |
Key finding: Each substance is associated with exactly one dossier (1-to-1 strict join). No substance has multiple dossiers; no dossier has multiple substances. This simplifies substance-level queries: a substance's full dossier metadata can be retrieved with a single join.
The 2260 orphan substances (28.7%) are chemical entities registered in the SUB sheet but not yet linked to a completed assessment dossier. These are typically:
- Pre-registration entries: substances filed for future EFSA review
- Reference substances only: used as analytical comparators in
other dossiers (CAS-number, InChIKey, SMILES) but not directly assessed
- Incomplete legacy entries: substances carried over from OFT 2.0
where the dossier was retired or split
Orphan substance composition distribution
| Composition | Orphan count |
|---|---|
| (no composition recorded) | 2134 |
| multi-constituent substance | 74 |
| mono-constituent substance | 26 |
| microorganism or toxin produced by a microorganism | 25 |
| polymer | 1 |
2134 of 2260 orphans (94.4%) carry no composition type. Most orphans are "reference-only" entries with a CAS number or InChIKey but not yet characterized into the IUCLID composition taxonomy.
Orphan substance origin distribution
| Origin | Orphan count |
|---|---|
| organic | 1858 |
| (no origin recorded) | 241 |
| inorganic | 110 |
other: | 26 |
| organometallic | 25 |
1858 orphans are organic (82.2% of orphans) — consistent with the overall dataset's 78.5% organic share, suggesting the orphans are not biased by chemistry type.
Full Composition × Origin cross-tab (7880 substances)
| Composition | Origin | Count | Share |
|---|---|---|---|
| mono-constituent substance | organic | 3835 | 48.7% |
| (no composition) | organic | 1973 | 25.0% |
| multi-constituent substance | (no origin) | 613 | 7.8% |
| (no composition) | (no origin) | 313 | 4.0% |
| multi-constituent substance | organic | 277 | 3.5% |
| mono-constituent substance | (no origin) | 190 | 2.4% |
| mono-constituent substance | inorganic | 158 | 2.0% |
| microorganism or toxin | (no origin) | 124 | 1.6% |
| (no composition) | inorganic | 103 | 1.3% |
| polymer | organic | 89 | 1.1% |
| mono-constituent substance | organometallic | 53 | 0.7% |
| multi-constituent substance | other: | 41 | 0.5% |
| (no composition) | organometallic | 27 | 0.3% |
| (no composition) | other: | 22 | 0.3% |
| multi-constituent substance | inorganic | 15 | 0.2% |
Headline categories:
- mono-constituent × organic (3835, 48.7%): standard single-
molecule organic substances — vitamins, amino acids, fatty acids, pesticide actives, pharmaceutical residues, feed additives
- (no composition) × organic (1973, 25.0%): organic substances
with no IUCLID composition assignment — often UVCB (Unknown or Variable composition, Complex reaction products or Biological materials) or natural extracts
- multi-constituent (956, 12.1%): formulations, botanical
extracts, mixtures — distributed across organic, inorganic, and undefined origins
Pet food relevance
Substance-to-dossier mapping enables several pet food analyses:
- Substance-centric queries: For any given substance
(e.g. "carrageenan"), query its dossier once, get full metadata (ExpertGroup, DateOfEvaluation, Regulation, ToxicologicalReferences) in a single join.
- Polymer + multi-constituent (1045 substances): This is the
pet-food-formulation-relevant subset — gelling agents (carrageenan, agar, alginate), emulsifiers, stabilizers, palatants, and mixed feed additives. The dossier-level data can be used to identify which polymers/mixtures have explicit pet-food safety assessments.
- Orphan substances (2260): Useful as a watchlist for
future EFSA assessments. Pet food manufacturers tracking regulatory pipelines can monitor when orphan substances gain dossier associations.
Source: toxicology/efsa/OFT3.0_export_repository.xlsx, sheet DOSSIER_DOCS (106409 rows × 5 cols), joined to SUB (7880 rows × 11 cols) via Document UUID.