The microbial spoilage of meat and seafood products with short shelf lives is responsible for a significant amount of food waste. Food spoilage is a very heterogeneous process, involving the growth of various, poorly characterized bacterial communities. In this study, we conducted 16S ribosomal RNA gene pyrosequencing on 160 samples of fresh and spoiled foods to comparatively explore the bacterial communities associated with four meat products and four seafood products that are among the most consumed food items in Europe. We show that fresh products are contaminated in part by a microbiota similar to that found on the skin and in the gut of animals. However, this animal-derived microbiota was less prevalent and less abundant than a core microbiota, psychrotrophic in nature, mainly originated from the environment (water reservoirs). We clearly show that this core community found on meat and seafood products is the main reservoir of spoilage bacteria. We also show that storage conditions exert strong selective pressure on the initial microbiota: alpha diversity in fresh samples was 189±58 operational taxonomic units (OTUs) but dropped to 27 ± 12 OTUs in spoiled samples. The OTU assemblage associated with spoilage was shaped by low storage temperatures, packaging and the nutritional value of the food matrix itself. These factors presumably act in tandem without any hierarchical pattern. Most notably, we were also able to identify putative new clades of dominant, previously undescribed bacteria occurring on spoiled seafood, a finding that emphasizes the importance of using culture-independent methods when studying food microbiota.

In order to characterise the spoilage related to microbiota of raw salmon, a combination of culture-dependent and -independent methods, including PCR-TTGE, was used to analyse 3 raw salmon batches stored for 3 days at chilled temperature in modified atmosphere packaging (MAP) (50% CO₂/50% N₂) or under vacuum. Sensory evaluation, microbiological enumeration and chemical analysis were performed after 3, 7 and 10 days of storage. At the onset of spoilage, 65 bacterial isolates were picked from the plates. Thus, 13 different genera or species were identified by phenotypic and molecular tests: Serratia spp., Photobacterium phosphoreum, Yersinia intermedia, Hafnia alvei, Buttiauxella gaviniae, Pseudomonas sp., Carnobacterium maltaromaticum, Carnobacterium divergens, Lactococcus piscium, Lactobacillus fuchuensis, Vagococcus carniphilus, Leuconostoc gasicomitatum and Brochothrix thermosphacta. The PCR-TTGE profiles and band identification enabled a shift of the dominant populations during the storage to be visualised for all the batches, probably due to the temperature change and the packaging. At the beginning of storage, Pseudomonas sp. dominated the raw salmon microbiota while in the following days (7 and 10), P. phosphoreum and L. piscium were identified as the main bacterial groups. This study enhances the knowledge of MAP and vacuum-packed raw salmon spoilage microbiota.