Section 3 generates the largest variety of additional data requests. What EFSA expects, what most applicants get wrong, and how to avoid the common gaps.
Among the ten mandatory sections of a novel food dossier, compositional data generates the widest variety of additional data requests. Ververis et al. (2020) noted that compositional characterisation produces the "largest variety of requests" across published EFSA opinions — more diverse than any other section.
In our own analysis of 152 published opinions, the compositional data section averaged 8,900 characters — the second-longest section after toxicological data. Length alone does not equal quality, but it reflects the depth of evidence EFSA requires.
The 2024 scientific guidance sets out the requirements clearly:
Five independent batches. Compositional data must be generated from a minimum of five independent, representative production batches. "Independent" means produced at different times using the routine manufacturing process — not five aliquots from the same run.
Six mandatory categories. Every novel food must be characterised for carbohydrates, fat, protein, minerals, vitamins, and moisture. This framework comes from EFSA's NORA extraction methodology and applies regardless of the novel food type.
Contaminant screening. Heavy metals, mycotoxins, pesticide residues, polycyclic aromatic hydrocarbons, and microbiological criteria must be analysed. The specific contaminants depend on the source material and production process.
Validated methods. Analytical methods must be validated and fit for purpose. Where standard methods exist (AOAC, ISO, CEN), EFSA expects their use unless a justified alternative is provided. Limits of detection and quantification must be stated.
Mass balance. The sum of all characterised components should account for the full composition of the novel food. A significant uncharacterised fraction raises questions about what the product actually contains.
Based on published opinions and ADR patterns, the most frequent compositional data gaps are:
Insufficient batch data. Submitting three batches when five are required. This triggers a near-automatic ADR and delays the assessment by the time it takes to produce and analyse two more batches — which can be months if the production process has a long cycle time.
Missing contaminant analyses. Applicants sometimes omit contaminants they consider irrelevant to their product. EFSA takes a different view: the contaminant panel must be comprehensive, and the absence of data is not evidence of absence.
Unvalidated or inappropriate methods. Using a method developed for a different matrix, or failing to report the method's performance characteristics. EFSA cannot assess the reliability of data without knowing the method's limits.
Incomplete mass balance. Reporting macronutrients and a few micronutrients but leaving 15–20% of the composition unaccounted for. This is particularly common with complex biological matrices like algal biomass or fermentation products.
Compositional data does not exist in isolation. The specifications in Section 1 must be supported by the analytical data in Section 3. The production process in Section 2 must produce a product consistent with the composition reported. The stability data in Section 4 must track the compositional parameters over time.
Cross-section inconsistencies involving compositional data are among the most consequential findings in EFSA's assessment. They suggest disconnected data generation rather than a coherent characterisation programme.
- Ververis, E. et al. (2020). Novel foods in the European Union. *Food Research International*. - EFSA (2024). Scientific Guidance for the preparation of applications for authorisation of novel foods. *EFSA Journal*. - Foley & Nicová (2024). NORA Standard Operating Procedure. *EFSA Supporting Publication*.
Paste a section and Borgh will flag the gaps that matter — with references to the relevant EFSA guidance.
Try a section →