Oversampling methods for machine learning model data training to improve model capabilities to predict the presence of Escherichia coli MG1655 in spinach wash water.

We assessed the efficacy of oversampling techniques to enhance machine learning model performance in predicting Escherichia coli MG1655 presence in spinach wash water. Three oversampling methods were applied to balance two datasets, forming the basis for training random forest (RF), support vector machines (SVMs), and binomial logistic regression (BLR) models. Data underwent method-specific centering and standardization, with outliers replaced by feature-specific means in training datasets. Testing occurred without these preprocessing steps. Model hyperparameters were optimized using a subset of testing data via 10-fold cross-validation. Models were trained on full datasets and tested on newly acquired spinach wash water samples. Synthetic Minority Oversampling Technique (SMOTE) and Adaptive Synthetic Sampling approach (ADASYN) achieved strong results, with SMOTE RF reaching an accuracy of 90.0%, sensitivity of 93.8%, specificity of 87.5%, and an area under the curve (AUC) of 98.2% (without data preprocessing) and ADASYN achieving 86.55% accuracy, 87.5% sensitivity, 83.3% specificity, and a 92.4% AUC. SMOTE and ADASYN significantly improved (p < 0.05) SVM and RF models, compared to their non-oversampled counterparts without preprocessing. Data preprocessing had a mixed impact, improving (p < 0.05) the accuracy and specificity of the BLR model but decreasing the accuracy and specificity (p < 0.05) of the SVM and RF models. The most influential physiochemical feature for E. coli detection in wash water was water conductivity, ranging from 7.9 to 196.2 µS. Following closely was water turbidity, ranging from 2.97 to 72.35 NTU within this study.

Effect of conditioner and moisture content on flowability of yellow cornmeal.

Flow characterization of ground materials based on standard physical properties is not always accurate and must be complemented with other properties that help characterize their flowability. The flow properties of yellow cornmeal as a function of moisture content (10.0%, 13.5% 17.0%, and 20.0% wet basis) and concentrations of added conditioner (calcium stearate, a caking agent classified at GRAS, at 0.50%, 0.75%, and 1.00% wt/wt) were measured. The optimum flow behavior characteristics of the cornmeal were achieved at 0.50% wt/wt calcium stearate and 10.0% (w.b.) moisture content based on the flow function test. Overall, the material's flowability decreased with increased moisture content based on very high values for Hausner's ratio, Carr's Index, and angle of repose. The flow index (ffc) obtained by the flow function test decreased from 6.47 to 3.82 as moisture increased, indicating increased cohesivity. Calcium stearate increased the flowability of the material at 0.50% wt/wt, beyond which the flowability was not affected (p > .05). Samples were very hygroscopic and classified as Class IV. The Hausner's ratio, Carr's Index, and angle of repose showed a strong linear relationship with the flow index with correlation values of 0.91, 0.88, and 0.88, respectively. Isotherms together with physical properties should be used to determine the flow characteristics of granulated materials such as yellow cornmeal to establish the best storage and processing conditions.

Effect of Condensed Tannin Profile on Wheat Flour Dough Rheology.

Proanthocyanidins (PA) cross-link proteins and could expand wheat gluten functionality; however, how the PA MW or gluten profile affect these interactions is unknown. Effect of PA MW profile (sorghum versus grape seed PA) on dough rheology of high versus low insoluble polymeric protein (IPP) wheat flour was evaluated using mixograph, large (TA.XT2i) and small (HAAKE Rheostress 6000) deformation rheometry. Sorghum PA (93% polymeric) more effectively (p < 0.05) strengthened both glutens than grape seed PA (45% polymeric), without reducing gluten extensibility. These effects were higher in low IPP (weak gluten) flour, e.g., sorghum PA doubled IPP, increased mix time by 75%, dough elasticity by 82%, and peak angle by 17° versus control. Grape seed PA increased IPP by 75% and elasticity by 36%, but reduced peak angle by 15°, indicating reduced mixing tolerance. Sorghum PA, but not grape seed PA, increased (p < 0.05) all above parameters in high IPP dough. Polymeric PA more effectively strengthened gluten than oligomeric PA, likely via more efficient protein cross-linking to overcome strong antioxidant effect of PA. High MW PA may be useful natural gluten strengtheners for diverse applications.

Physical Stability of Octenyl Succinate-Modified Polysaccharides and Whey Proteins for Potential Use as Bioactive Carriers in Food Systems.

The high cost and potential toxicity of biodegradable polymers like poly(lactic-co-glycolic)acid (PLGA) has increased the interest in natural and modified biopolymers as bioactive carriers. This study characterized the physical stability (water sorption and state transition behavior) of selected starch and proteins: octenyl succinate-modified depolymerized waxy corn starch (DWxCn), waxy rice starch (DWxRc), phytoglycogen, whey protein concentrate (80%, WPC), whey protein isolate (WPI), and α-lactalbumin (α-L) to determine their potential as carriers of bioactive compounds under different environmental conditions. After enzyme modification and particle size characterization, glass transition temperature and moisture isotherms were used to characterize the systems. DWxCn and DWxRc had increased water sorption compared to native starch. The level of octenyl succinate anhydrate (OSA) modification (3% and 7%) did not reduce the water sorption of the DWxCn and phytoglycogen samples. The Guggenheim-Andersen-de Boer model indicated that native waxy corn had significantly (P < 0.05) higher water monolayer capacity followed by 3%-OSA-modified DWxCn, WPI, 3%-OSA-modified DWxRc, α-L, and native phytoglycogen. WPC had significantly lower water monolayer capacity. All Tg values matched with the solid-like appearance of the biopolymers. Native polysaccharides and whey proteins had higher glass transition temperature (Tg) values. On the other hand, depolymerized waxy starches at 7%-OSA modification had a "melted" appearance when exposed to environments with high relative humidity (above 70%) after 10 days at 23 °C. The use of depolymerized and OSA-modified polysaccharides blended with proteins created more stable blends of biopolymers. Hence, this biopolymer would be suitable for materials exposed to high humidity environments in food applications.

Factors affecting radiation D-values (D10) of an Escherichia coli cocktail and Salmonella Typhimurium LT2 inoculated in fresh produce.

UNLABELLED: This study evaluated the effect of produce type, resuspension medium, dose uniformity ratio (DUR), and sample preparation conditions (tissue exposure, MAP, anoxia) on the D10 -value of an Escherichia coli cocktail (BAA-1427, BAA-1428, and BAA-1430) and Salmonella Typhimurium LT2 inoculated on the surfaces of tomato, cantaloupe, romaine lettuce, and baby spinach. Produce at room temperature were irradiated using a 1.35 MeV Van de Graaf electron beam accelerator at 0.2 to 0.9 kGy. The D10-values for E. coli and Salmonella were 0.20 ± 0.01 kGy and 0.14 ± 0.01 kGy, respectively. Bacterial inactivation was not affected by produce type as long as the samples were irradiated in unsealed bags, the bacteria were suspended in broth, and the sample tissue was exposed. Sample location in front of the e-beam source during exposure is crucial. A 20% increase in DUR yielded a 53% change in the D10- values. Variations in sample preparation, microbiological methods and irradiation set-up, result in variable D10-values for different microorganisms on fresh produce. PRACTICAL APPLICATIONS: Most irradiation studies disregard the effect of sample handling and processing parameters on the determination of the D10-value of different microorganisms in fresh and fresh-cut produce. This study shows the importance of exposure of sample, resuspension medium, available oxygen, and dose uniformity ratio. D10-values can differ by 35% to 53% based on these factors, leading to considerable under- or over-estimation of the irradiation treatment. Results from this study will help to lay firm groundwork for future studies on D10-values determination for different pathogens on fruits and vegetables.

Effect of oxygen-absorbing packaging on the shelf life of a liquid-based component of military operational rations.

Oxygen within the sealed package can reduce the quality of liquid-based food products with high oil content such as hot-filled meal-ready-to-eat (MRE) cheese spread, a component of military operational rations. The aim of this study was to test a novel oxygen absorber-containing laminate material and its ability to maintain and/or extend shelf life of a cheese-spread MRE item. An iron-based oxygen absorber (ABSO(2)RB(R)) activated by moisture was incorporated into the laminate and used to pack hot-filled cheese spread MREs. The kinetics of oxygen absorption due to humidity and temperature were characterized and peel tests performed to ensure pouch seal integrity. Accelerated shelf-life tests of ABSO(2)RB and regular MRE pouches without the O(2)-absorber were conducted for 3 mo at 51.7 degrees C (125 degrees F), and 6 mo at 37.8 degrees C (100 degrees F) by measuring oxygen concentration (Mocon O(2)-analyzer), microbiological, and physicochemical quality characteristics, including color, texture, moisture, free fatty acid (FFA), pH, water activity, and vitamins and A. Pouches stored at 26.7 degrees C (80 degrees F) for 12 mo served as calibrated controls. Consumer tests were conducted in-house and a confirmatory sensory test was conducted at Natick by a trained panel using a 9-point hedonic scale. ABSO(2)RB-laminates maintain the same seal integrity and strength as those of the control samples. The headspace oxygen concentrations in these pouches reached (P < 0.05) < 0.5% in 11 d of storage at 26.7 degrees C (80 degrees F) and remained below this level throughout the storage period (1 y). No microbial growth (aerobic, coliforms, yeast, and molds) was detected (P < 0.05) for both packages. Overall, the ABSO(2)RB-pouches indicate an improved reduction in oxygen and vitamin C retention compared with MRE controls and maintained product quality (physicochemical and organoleptic). ABSO(2)RB-laminates met the accelerated shelf-life requirement of 1 mo at 51.7 degrees C (125 degrees F), and 6 mo at 37.8 degrees C (100 degrees F). This study clearly shows the benefits of using active packaging technology on retaining nutrition and prolonging shelf life of high-fat, liquid content MRE items.

Surrogates for validation of electron beam irradiation of foods.

The aim of this study was to identify a potential surrogate to describe the radiation sensitivity of the most common pathogens encountered in fruits. Three pathogens: Escherichia coli O157:H7 933, Listeria monocytogenes ATCC 51414, and Salmonella Poona, and five non-pathogens: E. coli K-12 MG1655, Listeria innocua Seeliger 1983 (NRRL B-33003 and NRRl B-33014), Enterobacter aerogenes, and Salmonella LT2 were inoculated (populations of 10(7)-10(9) CFU/ml) into model food systems (10% w/w gelatin) and exposed to doses up to 1.0 kGy using a 2 MeV Van der Graaf linear accelerator. The non-pathogen E. coli K-12 MG1655 was highly resistant to radiation (D(10)=0.88 kGy) in comparison to the other strains while L. monocytogenes was the more radiation-resistant pathogen (D(10)=1.09 kGy). Thus, E. coli K-12 MG1655 could be a suitable surrogate for e-beam studies with L. monocytogenes as the indicator pathogen. L. innocua strains were more radiation-sensitive (D(10)=0.66, 0.72 kGy) than their pathogenic counterpart. S. Poona and E. coli O157:H7 were even more radiation-sensitive (D(10)=0.38, 0.36 kGy, respectively). S. LT2 was the least radiation-resistant pathogen with D(10)=0.12 kGy. In a later study, the radiation resistance of the pathogens and the surrogate was evaluated when inoculated in a real food (i.e., fresh cantaloupe). The D(10) values obtained in this experiment were higher than those obtained with the model foods. However, the surrogate was still more radiation-resistant and could therefore be used to indicate decontamination of the target pathogens under electron beam irradiation.