En savoir plus

Deux doctorants de SECALIM y ont présenté leur poster.

Rodney Feliciano	Antoine Gaillac

P2 Poster Session 2 – Beverages and Acid/Acidified Foods, Dairy, General Microbiology, Laboratory and Detection Methods, Low-Water Activity Foods, Meat, Poultry and Eggs, Microbial Food Spoilage, Packaging, Sanitation and Hygiene, and Seafood

P2-03 - Raw Milk Quality Under Hot Weather Conditions: Statistical Analysis of a Large Dataset

Mr. Rodney J Feliciano¹, Géraldine Boué¹, Fahad Mohssin², Muhammad Mustafa Hussaini² and Dr. Jeanne-Marie Membré¹,

(1)Secalim, INRAE, ONIRIS- Ecole Nationale Vétérinaire, Agroalimentaire et de l'alimentation de Nantes-Atlantique, Nantes, France, (2)AlSafi Danone, Al-Kharj, Saudi Arabia

Abstract

Introduction: Raw milk quality, such as microbial counts and physicochemical properties, impacts the quality of derived dairy products including the payment of milk to the producer. Moreover, conditions of dairy production under hot weather conditions (temperature, relative humidity, cow feeding, barn organisation, etc) may have negative impacts on these qualities.
Purpose: This study aimed to decipher the structure of a raw milk dataset, based on microbial counts and physicochemical properties, produced under hot weather conditions. In particular, the presence of a seasonal pattern was sought.
Methods: One-year raw milk quality control data, composed of microbial counts (Total aerobic counts thermophilic and E. coli counts) and physicochemical data (e.g. protein, fat, total solids, solids not fat) were obtained from a large-scale dairy farm (having several locations) operating under hot weather conditions. Daily temperature and relative humidity were collected and added to the dataset as “environmental variables”. Finally, a data set of 473 rows and 18 columns was obtained. Correlation analysis, Principal Component Analysis (PCA) and Partial Least-Squares (PLS) regression were performed to analyse the trends and dataset structure.
Results: Positive correlations were observed between microbial counts, physicochemical properties, and environmental variables. However, no high correlations between these three blocks of variables were observed. Likewise, PCA did not indicate a high structure within the dataset but interestingly revealed a seasonal pattern (more or less pronounced according to the dairy farm locations). Finally, PLS enabled to predict percent fat and protein, two variables associated with the payment of milk to the producer, by a subset of physicochemical properties and environmental variables. On the opposite, microbial counts could not be predicted by physicochemical properties or environmental variables.
Significance: This study highlights that combinations of statistical techniques help in deciphering the safety and quality of raw milk production as shown by quality control data sets and consequently aid decision-making processes.

P2-28 - Biofilm Formation Diversity of Brochothrix thermosphacta a Major Food Spoiler

Antoine Gaillac¹, Julien Deschamps², Romain Briandet², Evelyne Vigneau³, Philippe Courcoux³, Emmanuel Jaffrès¹ and Herve Prevost¹,

(1)INRAE-Oniris, Nantes, France, (2)INRAE, Paris, France, (3)Oniris, Nantes, France

Abstract

Introduction: Biofilms play an important role in residence and persistence of microorganisms in food industry. Brochothrix thermosphacta is considered as a major food spoilage bacteria (Illikoud et al., 2019). This bacteria has been identified in biofilms on multiple surfaces of the food processing environment (Wagner et al., 2021, 2020).
Purpose: The study of B. thermosphacta ability to form biofilm is necessary to better understand the residence and persistence of this bacteria in food processing environment and finally to better control food contamination from the industrial surfaces.
Methods: Biofilms formation diversity at 25°C of a B. thermosphacta strains collection (i.e. 31 strains) was characterized by using the biofilm ring test (cBRT), the crystal violet staining and the laser scanning confocal microscopy (LSCM), with the images analysis by BiofilmQ (Hartmann et al., 2021).
Results: Using cBRT, the 31 strains were classified in 4 groups, according to biofilm formation. Fourteen were classified as poor biofilm producer and represent 65% of the strains. Four strains as weak (12.5%), five as medium (15.5%), and two as high biofilm producer (6%). This shown a predominance of poor and weak early biofilm stage producer strains. The ability to form a mature biofilm was analysed with crystal violet staining. A statistical analysis shows seven significant different group of mature biofilm formation (p<0.05). The five higher biofilm producer groups are composed by the same 6 strains already classified as high and medium biofilm early stage producer by cBRT.
The analysis of LSCM images by BiofilmQ identified 10 parameters that influence the diversity. Biofilm are flat with high cell density, biofilm volume, biovolume, number of cells, and substrate coverage.
Significance: This study is to our knowledge the first describing biofilm formation by B. thermosphacta showing 7 high biofilm producing strains.