All-Food-Sequencing: Identification and quantification of food ingredients by whole-genome metagenomics
Loading...
Date issued
Authors
Editors
Journal Title
Journal ISSN
Volume Title
Publisher
Reuse License
Abstract
All-food sequencing (AFS) is a non-targeted, whole-genome DNA-based screening method for simultaneous qualitative and quantitative species diagnosis in complex foods. No a priori knowledge or primers are required, and a single analysis can potentially detect animal and plant components as well as fungi, bacteria, and other accompanying microbiota.
Quantitative performance of AFS was evaluated using controlled calibration samples, including defined meat and fish mixtures. In direct comparison with established assays, AFS delivered quantification that outperformed conventional qPCR and matched ddPCR, while retaining a key practical advantage: broad, primer-independent screening and quantification of multiple ingredients within a single workflow, rather than many separate target specific assays. The calibration series also identified two systematic factors that influence quantitative accuracy. First, differences in genome size can shift read proportions, but this bias was shown to be correctable by genome size normalization. Second, matrix effects of the food composition can alter DNA extraction yield between ingredients, so the highest quantitative agreement is achieved with matrix-specific calibration - a constraint shared by all DNA-based methods.
Application for real food products showed how AFS behaves under practical conditions with regard to heterogeneity and incomplete or inaccurate labelling of species components. Doner kebab samples showed significant deviations and several cases in which the predominant meat type did not correspond to the advertised expectation. In addition, other ingredients were detected that suggest unintentional admixture or deliberate substitution. In the case of seafood and surimi products, the declared main ingredients were generally confirmed. In addition, the analysis revealed undeclared taxa, including additional seafood and plant components in mixed dishes, some with potential allergen relevance. Beyond the main ingredient composition, AFS provided early warning signals, including allergen-relevant plant admixtures and microbial patterns that indicate incipient spoilage.
The analysed samples also revealed practical limitations of AFS that are crucial for regulatory surveillance: ambiguous classification within closely related species, dependence on reference genomes, and the need to interpret low-level findings conservatively and verify them with targeted follow-up measures. Specificity of read classification was improved by both algorithm choice and sequencing technology. K-mer-based classification and database partitioning enabled screening against significantly larger genome reference collections. In a comparison between Illumina short reads and Oxford Nanopore long reads on calibration sausages, long-read sequencing improved quantification accuracy and reduced the number of false-positives despite higher error rates in long read, as longer reads provide more discriminative information for resolving conserved regions that otherwise drive ambiguous classification.
AFS shows the potential to provide reliable formulation information and cross-domain early warning signals as a universal, primer-free WGS screening method, provided that reference dependency and taxonomic ambiguity are recognized as current limitations and integrated into the decision on verification or follow-up measures. Therefore, it is promising for routine screening in official food monitoring in the future.