Search for: All records

The growth and activity of adherent cells can be enabled or enhanced through attachment to a solid surface. For food and beverage production processes, these solid supports should be food-grade, low-cost, and biocompatible with the cell of interest. Solid supports that are edible can be a part of the final product, thus simplifying downstream operations in the production of fermented beverages and lab grown meat. We provide proof of concept that edible filamentous fungal pellets can function as a solid support by assessing the attachment and growth of two model cell types: yeast, and myoblast cells. The filamentous fungus Aspergillus oryzae was cultured to produce pellets with 0.9 mm diameter. These fugal pellets were inactivated by heat or chemical methods and characterized physicochemically. Chemically inactivated pellets had the lowest dry mass and were the most hydrophobic. Scanning electron microscope images showed that both yeast and myoblast cells naturally adhered to the fungal pellets. Over 48 h of incubation, immobilized yeast increased five-fold on active pellets and six-fold on heat-inactivated pellets. Myoblast cells proliferated best on heat-treated pellets, where viable cell activity increased almost two-fold, whereas on chemically inactivated pellets myoblasts did not increase in the cell mass. These results support the use of filamentous fungi as a novel cell immobilization biomaterial for food technology applications.

 

The past decade witnessed rapid development in the measurement and monitoring technologies for food science. Among these technologies, spectroscopy has been widely used for the analysis of food quality, safety, and nutritional properties. Due to the complexity of food systems and the lack of comprehensive predictive models, rapid and simple measurements to predict complex properties in food systems are largely missing. Machine Learning (ML) has shown great potential to improve the classification and prediction of these properties. However, the barriers to collecting large datasets for ML applications still persists. In this paper, we explore different approaches of data annotation and model training to improve data efficiency for ML applications. Specifically, we leverage Active Learning (AL) and Semi-Supervised Learning (SSL) and investigate four approaches: baseline passive learning, AL, SSL, and a hybrid of AL and SSL. To evaluate these approaches, we collect two spectroscopy datasets: predicting plasma dosage and detecting foodborne pathogen. Our experimental results show that, compared to the de facto passive learning approach, advanced approaches (AL, SSL, and the hybrid) can greatly reduce the number of labeled samples, with some cases decreasing the number of labeled samples by more than half.