Development of antifungal ingredients for dairy products: From in vitro screening to pilot scale application.

Biopreservation represents a complementary approach to traditional hurdle technologies for reducing microbial contaminants (pathogens and spoilers) in food. In the dairy industry that is concerned by fungal spoilage, biopreservation can also be an alternative to preservatives currently used (e.g. natamycin, potassium sorbate). The aim of this study was to develop antifungal fermentates derived from two dairy substrates using a sequential approach including an in vitro screening followed by an in situ validation. The in vitro screening of the antifungal activity of fermentates derivating from 430 lactic acid bacteria (LAB) (23 species), 70 propionibacteria (4 species) and 198 fungi (87 species) was performed against four major spoilage fungi (Penicillium commune, Mucor racemosus, Galactomyces geotrichum and Yarrowia lipolytica) using a cheese-mimicking model. The most active fermentates were obtained from Lactobacillus brevis, Lactobacillus buchneri, Lactobacillus casei/paracasei and Lactobacillus plantarum among the tested LAB, Propionibacterium jensenii among propionibacteria, and Mucor lanceolatus among the tested fungi. Then, for the 11 most active fermentates, culture conditions were optimized by varying incubation time and temperature in order to enhance their antifungal activity. Finally, the antifungal activity of 3 fermentates of interest obtained from Lactobacillus rhamnosus CIRM-BIA1952, P. jensenii CIRM-BIA1774 and M. lanceolatus UBOCC-A-109193 were evaluated in real dairy products (sour cream and semi-hard cheese) at a pilot-scale using challenge and durability tests. In parallel, the impact of these ingredients on organoleptic properties of the obtained products was also assessed. In semi-hard cheese, application of the selected fermentates on the cheese surface delayed the growth of spoilage molds for up to 21 days, without any effect on organoleptic properties, P. jensenii CIRM-BIA1774 fermentate being the most active. In sour cream, incorporation of the latter fermentate at 2 or 5% yielded a high antifungal activity but was detrimental to the product organoleptic properties. Determination of the concentration limit, compatible with product acceptability, showed that incorporation of this fermentate at 0.4% prevented growth of fungal contaminants in durability tests but had a more limited effect against M. racemosus and P. commune in challenge tests. To our knowledge, this is the first time that the workflow followed in this study, from in vitro screening using dairy matrix to scale-up in cheese and sour cream, is applied for production of natural ingredients relying on a large microbial diversity in terms of species and strains. This approach allowed obtaining several antifungal fermentates which are promising candidates for dairy products biopreservation.

Technical note: High-throughput method for antifungal activity screening in a cheese-mimicking model.

In this study, we developed a high-throughput antifungal activity screening method using a cheese-mimicking matrix distributed in 24-well plates. This method allowed rapid screening of a large variety of antifungal agent candidates: bacterial fermented ingredients, bacterial isolates, and preservatives. Using the proposed method, we characterized the antifungal activity of 44 lactic acid bacteria (LAB) fermented milk-based ingredients and 23 LAB isolates used as protective cultures against 4 fungal targets (Mucor racemosus, Penicillium commune, Galactomyces geotrichum, and Yarrowia lipolytica). We also used this method to determine the minimum inhibitory concentration of a preservative, natamycin, against 9 fungal targets. The results underlined the strain-dependency of LAB antifungal activity, the strong effect of fermentation substrate on this activity, and the effect of the screening medium on natamycin minimum inhibitory concentration. Our method could achieved a screening rate of 1,600 assays per week and can be implemented to evaluate antifungal activity of microorganisms, fermentation products, or purified compounds compatible with dairy technology.

Improving motion-mask segmentation in thoracic CT with multiplanar U-nets.

PURPOSE: Motion-mask segmentation from thoracic computed tomography (CT) images is the process of extracting the region that encompasses lungs and viscera, where large displacements occur during breathing. It has been shown to help image registration between different respiratory phases. This registration step is, for example, useful for radiotherapy planning or calculating local lung ventilation. Knowing the location of motion discontinuity, that is, sliding motion near the pleura, allows a better control of the registration preventing unrealistic estimates. Nevertheless, existing methods for motion-mask segmentation are not robust enough to be used in clinical routine. This article shows that it is feasible to overcome this lack of robustness by using a lightweight deep-learning approach usable on a standard computer, and this even without data augmentation or advanced model design. METHODS: A convolutional neural-network architecture with three 2D U-nets for the three main orientations (sagittal, coronal, axial) was proposed. Predictions generated by the three U-nets were combined by majority voting to provide a single 3D segmentation of the motion mask. The networks were trained on a database of nonsmall cell lung cancer 4D CT images of 43 patients. Training and evaluation were done with a K-fold cross-validation strategy. Evaluation was based on a visual grading by two experts according to the appropriateness of the segmented motion mask for the registration task, and on a comparison with motion masks obtained by a baseline method using level sets. A second database (76 CT images of patients with early-stage COVID-19), unseen during training, was used to assess the generalizability of the trained neural network. RESULTS: The proposed approach outperformed the baseline method in terms of quality and robustness: the success rate increased from 53 % to 79 % without producing any failure. It also achieved a speed-up factor of 60 with GPU, or 17 with CPU. The memory footprint was low: less than 5 GB GPU RAM for training and less than 1 GB GPU RAM for inference. When evaluated on a dataset with images differing by several characteristics (CT device, pathology, and field of view), the proposed method improved the success rate from 53 % to 83 % . CONCLUSION: With 5-s processing time on a mid-range GPU and success rates around 80 % , the proposed approach seems fast and robust enough to be routinely used in clinical practice. The success rate can be further improved by incorporating more diversity in training data via data augmentation and additional annotated images from different scanners and diseases. The code and trained model are publicly available.

Prior drinking motives predict alcohol consumption during the COVID-19 lockdown: A cross-sectional online survey among Belgian college students.

The global outbreak of coronavirus disease 2019 (COVID-19) and the resulting lockdown measures have raised concerns regarding their effect on alcohol consumption. We investigated alcohol use during lockdown in a population of college students, usually characterized by social and heavy drinking. We also tested the predictive role of pre-lockdown drinking motives on alcohol consumption during lockdown. We collected data from 1951 French-speaking Belgian students during the lockdown period (April 1st - May 3rd, 2020) through a cross-sectional online survey. Participants self-reported their daily alcohol consumption (1) during a typical week in normal circumstances (i.e., before lockdown), and (2) since lockdown onset. We also assessed drinking motives and severity of alcohol use before lockdown. Our findings showed that 68.2% of the sample reported a lower alcohol consumption during lockdown compared to before lockdown, 17.2% conversely reporting a higher consumption. Enhancement, social and coping motives were all associated with heavy drinking before lockdown. Enhancement and social motives predicted lower alcohol consumption during lockdown among heavy drinkers. Conversely, coping motives, as well as social motives among low drinkers, predicted higher consumption during lockdown. Conformity motives, as well as enhancement motives among low and moderate drinkers, did not predict alcohol consumption before or during lockdown. Overall, several pre-lockdown drinking motives reliably predicted alcohol consumption during lockdown and could thus be used to identify at-risk populations and to tailor intervention programs on alcohol misuse during sanitary crises.

Alcohol consumption changes during the first COVID-19 lockdown: an online population survey in a convenience sample of French-speaking Belgian residents.

An online survey among French-speaking Belgians (N=7711) investigated self-reported changes in alcohol consumption during the first COVID-19-related lockdown (March 17th - May 4th 2020). Population-weighted estimates indicated that 31.37% of the population increased consumption during lockdown, 30.32% decreased consumption and 38.31% reported unchanged consumption. The magnitude of change was higher for "decreasers" than "increasers", resulting in a slight reduction in overall consumption. A multiple regression analysis revealed that age, occupational status, number of cohabitants, perceived alcohol accessibility, drinking motivations and previous consumption predicted consumption changes. The lockdown was thus associated with consumption modifications among French-speaking Belgians, without a prevailing direction of change.

The addition of a cocktail of yeast species to Cantalet cheese changes bacterial survival and enhances aroma compound formation.

Indigenous yeasts can be detected at high populations in raw milk Cantal cheese, a French Protected Denomination of Origin (PDO) hard cheese. To investigate their use as adjunct cultures to promote flavour development in Cantalet (small Cantal) cheese, three strains isolated from raw milk Cantal cheese, Kluyveromyces lactis, Yarrowia lipolytica, and Pichia fermentans were added at 3 (E3) and 5 (E5) log(10) colony-forming units (cfu)/mL to microfiltered milk at a ratio of 80/10/10 viable cells, respectively. The global microbial, compositional and biochemical changes induced by the presence of yeasts in cheese were determined. Adjunct yeasts did not grow but stayed at viable populations of approximately 4 and 6 log(10) cfu/g in E3 and E5 cheeses, respectively, throughout the ripening period. They were mainly constituted of K. lactis, while P. fermentans and Y. lipolytica were not detectable after 3 and 45 days of ripening, respectively. Several species of indigenous yeasts were also detected in E3 cheeses at the beginning of ripening only, and in the control cheeses without yeasts added. Lactoccoci survived for longer periods in the presence of yeast adjuncts, while, conversely, the viability of Streptococcus thermophilus decreased more rapidly. The addition of yeasts did not influence cheese composition and total free amino acid content. In contrast, it slightly increased lipolysis in both E3 and E5 cheeses and markedly enhanced the formation of some volatile aroma compounds. The concentrations of ethanol, ethyl esters and some branched-chain alcohols were 6 to 10 fold higher in E5 cheeses than in the control cheeses, and only slightly higher in E3 cheeses. This study shows that K. lactis has a potential as cheese adjunct culture in Cantalet cheese and that, added at populations of 4-5 log(10) cfu/g cheese, it enhances the formation of flavour compounds.

Oxytocin not only increases trust when money is at stake, but also when confidential information is in the balance.

Past studies have suggested that the neuropeptide oxytocin (OT) could play a crucial role in human trusting behavior. Specifically, people on OT would be more willing to entrust someone with their money than would people on a placebo. Because alternative explanations-which do not involve trust-exist for these studies' findings, the present study aimed to rule out confounds and test how OT influences trust behavior in a totally different context. The variable at stake was not money but confidential information. Sixty participants were randomly assigned to receive either OT or a placebo. Results showed that oxytocin does increase trust, and that its effects extend beyond money. Specifically, participants on OT were 44 times more trusting that their privacy would not be violated than participants on placebo.