1. EFSA’s Chemical Hazards Database
OpenFoodTox:
EFSA’s Chemical Hazards Database
Jean Lou Dorne
Senior Scientific Officer
Scientific Committee and
Emerging Risks Unit
Datathon 15/06/2018

2. Four pillars of risk assessment
Hazard identification and characterisation
Exposure assessment
Levels in food, feed, water,
environmental media, dietary
exposure, food consumption,
relevant food groups, time trends
Deterministic vs probabilistic
ADME, acute/sub-chronic/chronic toxicity,
human data, genotox, reprotox, mode of action,
NOECs (ecological)
NOELs, LOAELs (animal, human)
mathematical modelling (BMD),
Health based guidance (TDI…)
OpenFoodTox
The risk-characterization step integrates exposure to a chemical in a given population with toxicological effects on which a health-based guidance value is based, and concludes with the likelihood of adverse effects.
Risk characterisation
Relate exposure to environmental standards
(ecological), margin of safety (animal)
Margin of Exposure for genotoxic carcinogens
(Human)

3. -Why EFSA’s Chemical Hazards Database ?-
Catalogue of EFSA’s chemical risk assessment since its creation
-Contaminants (Human and Animal health)
-Vitamins and minerals (Human health) (NDA),
-Food additives and Nutrient Sources, Food contact materials, Flavourings and processing aids ,enzymes (Human Health)
-Feed Additives (Human and Animal Health, Ecotoxicology)
-Pesticides (Human and Animal health, Ecotoxicology)
Easy Reference and Crisis
One reference Database on Chemical Hazards: Search easily and efficiently
Crisis: Quick and Easy access to all EFSA’s Hazard Data
International Harmonisation
International efforts to Harmonise Templates for Hazard Data (ECHA/OECD)
Data Model based on templates compatible with IUCLID/ ECHA-OECD QSAR toolbox

4. -Qualitative map of the data model-
Chemical components
Substances
EFSA documents
Hazard identification/ Characterisation
Endpoint Study (e.g., critical studies)
Human health (including genotoxicity)
Animal health (target and non-target species)
Eco-toxicological data
Hazard Characterisation/Risk Characterisation
Health-based guidance value (threshold effects)
Margin of exposure (MOE) values (non-threshold effects)
Exposure
(for MOE)

5. -Overview of the Data Model -
Chemical Information
Information on chemical nomenclature (EU nomenclature, IUPAC, CAS... ), trade name, chemical group/panel (i.e. pesticide), chemical use (i.e fungicide), chemical structure (i.e triazoles, organophosphates….).
Document descriptors
Information on EFSA’s opinion for the specific chemical or group of chemicals. Info from EFSA ‘s RAW system (question number, mandate, number), link to the document
Toxicity Endpoint/ Hazard identification
Information on critical toxicity study using OECD picklists when possible (species, dose, target organ…)
Critical study to demonstrate genotoxicity status
Providing essential information of critical genotoxicity study when assessed
Hazard /Risk characterisation
Information for health based guidance values (ADI/TDI), margin of exposures, safety factors…

6. Relationships among data
Toxicity data: Endpoint Study
(e.g., critical studies)
Human health(including genotoxicity on positive substances)
Animal health
(target and non-target animals)
Ecotoxicity
Hazard/Risk characterisation
(ADI, TDI, MOE, MOS)
Toxicity data
(NOAEL, effects, toxicity)
Bibliographic details
(publication date, titles, authors)
Chemical identification
(substance and components)
Chemical components
Substances
Hazard characterisation/
Risk characterisation
Health-based guidance value (threshold effects)
Margin of exposure (non-threshold effects)
Margin of safety
Environmental standards
Documents

8. CONTENT
Microstrategy Tool:
Full Download Knowledge junction:
1,650 Scientific outputs (metadata + DOI)
4,400 Substances (chemical identifiers including SMILES)
12,000 risk assessment summaries
10,000 Toxicological endpoint studies
140 Positive genotoxicity studies

10. Main title

11. Openfoodtox and in silico tools
Future Mixture RA EFSA
Openfoodtox and in silico tools
Case studies to develop in silico tools
QSAR model on pesticide Toxicity in bees (OpenFoodTox, US-EPA, DEMETRA DB) : Classifier
QSAR model to predict LC50 in rainbow trout (OpenFoodTox) : Continuous model
-Physico-chem properties, structure, toxicity : R2 >0.75
QSAR model to predict NOAEL in rats (OpenFoodTox, Fraunhofer) : Continuous model
Scientific report Summer 2017
QSAR model to predict NOAEL for liver toxicity in rats (OpenFoodTox, Fraunhofer) : Continuous model

12. Main title

13. Main title

14. Main title

15. THe Future of Openfoodtox Sep 2018-2022
Collect new properties
Physico-chemical properties, TK data, bioaccumulation etc
Summary exposure estimates
Intermediate effects (mechanistic data)
New templates (OECD harmonised Templates)
Design template for WoE, biological relevance and uncertainty
Update TK template
Link OpenFoodTox with QSAR platforms
e.g VEGA
Published EFSA values and Predicted values

16. Many Thanks
Questions ?