Model on the microbial quality change of seasoned soybean sprouts for on-line shelf life prediction

Bootstrapping for Estimating the Conservative Kill Ratio of the Surrogate to the Pathogen for Use in Thermal Process Validation at the Industrial Scale

A surrogate is commonly used for process validations. The industry often uses the target log cycle reduction for the test (LCRTest) microorganism (surrogate) to be equal to the desired log cycle reduction for the target (LCRTarget) microorganism (pathogen). When the surrogate is too conservative with far greater resistance than the pathogen, the food may be overprocessed with quality and cost consequences. In aseptic processing, the Institute for Thermal Processing Specialists recommends using relative resistance (DTarget)/(DTest) to calculate LCRTest (product of LCRTarget and relative resistance). This method uses the mean values of DTarget and DTest and does not consider the estimating variability. We defined kill ratio (KR) as the inverse of relative resistance.The industry uses an extremely conservative KR of 1 in the validation of food processes for low-moisture foods, which ensures an adequate reduction of LCRTest, but can result in quality degradation. This study suggests an approach based on bootstrap sampling to determine conservative KR, leading to practical recommendations considering experimental and biological variability in food matrices. Previously collected thermal inactivation kinetics data of Salmonella spp. (target organism) and Enterococcus faecium (test organism) in Non-Fat Dried Milk (NFDM) and Whole Milk Powder (WMP) at 85, 90, and 95°C were used to calculate the mean KR. Bootstrapping was performed on mean inactivation rates to get a distribution of 1000 bootstrap KR values for each of the treatments. Based on minimum temperatures used in the industrial process and acceptable level of risk (e.g., 1, 5, or 10% of samples that would not achieve LCRTest), a conservative KR value can be estimated. Consistently, KR increased with temperature and KR for WMP was higher than NFDM. Food industries may use this framework based on the minimum processing temperature and acceptable level of risk for process validations to minimize quality degradation.

Application of deep neural techniques in predictive modelling for the estimation of Escherichia coli growth rate

AIMS

To develop a predictive model for Escherichia coli using deep neural networks.

METHODS AND RESULTS

Batch experiments are conducted at different temperatures closer to optimum value (36·5°C, 37°C, 37·5°C, 38°C and 38·5°C) to obtain the growth curves of E .coli K-12. Two primary models namely modified Gompertz and new logistic are chosen. Three secondary models namely Gaussian, nonlinear autoregressive eXogenous (NARX) model and long short-term memory (LSTM) are developed. The novelty in this paper is the development of secondary models using artificial neural network (ANN) and deep network. The performance measures chosen to compare the developed primary and secondary models are correlation coefficient (R² ), root-mean-square error (RMSE) and accuracy factor (A_f ). Results show that modified Gompertz model has better R² (0·99) and RMSE (0·019) when compared to new logistic model. Also, the deep network model outperforms other secondary models. Based on the primary and novel secondary model, a predictive model (tertiary model) is developed with improved accuracy and is validated.

CONCLUSIONS

The proposed predictive model exhibit good validation results in terms of RMSE and R² values and can be applied for determining the growth rate of E. coli at a particular temperature value.

SIGNIFICANCE AND IMPACT OF THE STUDY

The proposed model can be used in food processing industries during enzyme production such as Chymosin, to predict the growth rate of E. coli as a function of temperature. Also, the developed LSTM and NARX models can be used to predict maximum specific growth rate of other microbial strains with proper training.

Microbiological and Sensory Quality of Fresh Ready-to-Eat Artichoke Hearts Packaged under Modified Atmosphere

In recent years the sales of minimally processed vegetables have grown exponentially as a result of changes in consumer habits. The availability of artichoke buds as a ready-to-eat product would be, therefore, highly advantageous. However, minimally processed artichoke hearts are difficult to preserve because of their rapid browning and the proliferation of naturally occurring microorganisms. We developed artichoke hearts prepared as ready-to-eat products that maintain the characteristics of the fresh product. The microbiological stability, sensory qualities, and shelf life of the processed artichoke hearts were determined. During the shelf life, Salmonella, Listeria monocytogenes, and Escherichia coli counts were below the limits legally established by European regulations for minimally processed vegetables. The pH played an important role in microbial growth. Artichoke hearts had lower microbial counts in experiments conducted at pH 4.1 than in experiments conducted at pH 4.4, although the recommended threshold value for total plate count (7 log CFU/g) was not exceeded in either case. Sensory parameters were affected by the microorganisms, and artichoke products at lower pH had better sensory qualities. Vacuum impregnation techniques, modified atmosphere packaging, and low storage temperature were very effective for increasing the shelf life of minimally processed artichokes. The average shelf life was approximately 12 to 15 days.

Modeling the Growth of Epiphytic Bacteria on Kale Treated by Thermosonication Combined with Slightly Acidic Electrolyzed Water and Stored under Dynamic Temperature Conditions

The growth of epiphytic bacteria (aerobic mesophilic bacteria or Pseudomonas spp.) on kale was modeled isothermally and validated under dynamic storage temperatures. Each bacterial count on kale stored at isothermal conditions (4 to 25 °C) was recorded. The results show that maximum growth rate (μmax ) of both epiphytic bacteria increased and lag time (λ) decreased with increasing temperature (P < 0.05). The maximum population density (Nmax ) of Pseudomonas spp. was significantly greater than that of aerobic mesophilic bacteria, particularly in treated samples and/or at 4 and 10 °C (P < 0.05). The relationship between μmax of both epiphytic bacteria and temperature was linear (R(2) > 0.97), whereas lower R(2) > 0.86 and R(2) > 0.87 was observed for the λ and Nmax , respectively. The overall predictions of both epiphytic bacterial growths under nonisothermal conditions with temperature abuse of 15 °C agreed with the observed data, whereas those with temperature abuse of 25 °C were greatly overestimated. The appropriate parameter q0 (physiological state of cells), therefore, was adjusted by a trial and error to fit the model. This study demonstrates that the developed model was able to predict accurately epiphytic bacterial growth on kale stored under nonisothermal conditions particularly those with low temperature abuse of 15 °C.

Quantitative risk assessment for Listeria monocytogenes in selected categories of deli meats: impact of lactate and diacetate on listeriosis cases and deaths

Foodborne disease associated with consumption of ready-to-eat foods contaminated with Listeria monocytogenes represents a considerable pubic health concern. In a risk assessment published in 2003, the U.S. Food and Drug Administration and the U.S. Food Safety and Inspection Service estimated that about 90% of human listeriosis cases in the United States are caused by consumption of contaminated deli meats. In this risk assessment, all deli meats were grouped into one of 23 categories of ready-to-eat foods, and only the postretail growth of L. monocytogenes was considered. To provide an improved risk assessment for L. monocytogenes in deli meats, we developed a revised risk assessment that (i) models risk for three subcategories of deli meats (i.e., ham, turkey, and roast beef) and (ii) models L. monocytogenes contamination and growth from production to consumption while considering subcategory-specific growth kinetics parameters (i.e., lag phase and exponential growth rate). This model also was used to assess how reformulation of the chosen deli meat subcategories with L. monocytogenes growth inhibitors (i.e., lactate and diacetate) would impact the number of human listeriosis cases. Use of product-specific growth parameters demonstrated how certain deli meat categories differ in the relative risk of causing listeriosis; products that support more rapid growth and have reduced lag phases (e.g., turkey) represent a higher risk. Although reformulation of deli meats with growth inhibitors was estimated to reduce by about 2.5- to 7.8-fold the number of human listeriosis cases linked to a given deli meat subcategory and thus would reduce the overall risk of human listeriosis, even with reformulation deli meats would still cause a considerable number of human listeriosis cases. A combination of strategies is thus needed to provide continued reduction of these cases. Risk assessment models such as that described here will be critical for evaluation of different control approaches and to help define the combinations of control strategies that will have the greatest impact on public health.