Using a Centroid-based approach for a reliable identification of morels (Morchella spp.): A case study for food...

Food fraud is a problematic yet common phenomenon in the food industry. It impacts numerous sectors, including the market of edible mushrooms. Morel mushrooms are prized worldwide for their culinary and medicinal use. They represent a taxonomically complex group in which food fraud has already been reported. Among the methods to evaluate food fraud, some rely on comparisons of genetic sequences obtained from a sample to existing databases. However, the quality and usefulness of the results are limited by the type of comparison tool and the quality of the database used. The Centroid-based approach is applied by SmartGene in a proprietary artificial intelligence-based method for the generation of automatically curated reference databases that can be further expert curated. In this study, using sequences of the ribosomal internal transcribed spacer (ITS) of the genus Morchella (true morels), we compared this approach to the traditional pairwise alignment tool using two other databases: UNITE and Mycobank (MLST). The Centroid-based approach using an expert-curated database was more performant for the identification of 53 representative ITS sequences corresponding to validated species (83% accuracy, compared to 36% and 47% accuracy for UNITE and MLST, respectively). The Centroid method also revealed an inaccurate taxonomic annotation for sequences of commercial cultivars submitted to public databases. Combined with the web-based commercial software IDNS® available at Smartgene, the Centroid-based approach constitutes a valuable tool to ensure the quality of morel products on the market for actors of the food industry. PRACTICAL APPLICATION: The Centroid-based approach can be used by agri-food actors who need to identify true morels down to the species level without any prior taxonomical knowledge. These include routine laboratories of the food industry, food distributors, and public surveillance agencies. This is a reliable method that requires minimal skills and resources, therefore being particularly adapted for nonspecialists.