Predictive Food Microbiology: A Model to Achieving Shelf-Life Extension and Product Maximization

INTRODUCTION

Methods of food preservation, including salting, drying, and fermentation, have been practiced for thousands of years, showcasing an empirical method for managing microbial populations in food. These techniques are still utilized today by indigenous communities and likely in traditional products from more ‘advanced’ societies. Notable early instances of applying scientific principles to food preservation include Pasteur’s research on the specificity of unwanted fermentations in wine and Hansen’s provision of lactic starter cultures in Denmark at the close of the 19th century. Nevertheless, while a significant portion of the fermentation sector has embraced quantitative methods—likely due to large-scale production and the impact of chemical engineers—many areas of food microbiology have largely remained qualitative or, at best, semi-quantitative. Consequently, ‘shake and plate’ methods can enumerate to within F 0.5 log, with detection limits reaching as high as 100 cfu/g. Additionally, most probable number techniques often exhibit very broad confidence intervals, and enrichment methods can indicate the presence (but not necessarily the absence) of a specific organism in a sample. Naturally, the sample may be wholly insufficient to accurately reflect the prevalence (likelihood of occurrence) of the organism in the product lot, much less provide a numerical estimate of its density [1].

Worldwide, food safety protocols have advanced considerably to safeguard consumers on a broad scale. The 19th century saw the acknowledgment of the germ theory of disease, which initiated a focus on sanitation and animal welfare, resulting in the creation of improved safety regulations for dairy products. These improvements swiftly spread to other areas of the food sector [2]. The emergence of industrialization and large-scale food processing highlighted the necessity of public health initiatives designed to ensure the safety of food for consumers. Currently, stringent screening processes are implemented to reduce microbial risks (including pathogenic microbes, metabolites, and mycotoxins), chemical threats (such as heavy metals, pesticides, and carcinogens), as well as allergens and physical dangers like sharp objects [3]. Among these, biological hazards represent the most serious risk, frequently resulting in severe outbreaks [4]. A review by Lee and Yoon [5] indicates that norovirus is linked to the highest annual incidence of foodborne illnesses worldwide. Following it are Campylobacter, Salmonella, and Listeria monocytogenes. Additionally, a WHO report revealed that foodborne bacterial infections were found to be more common than viral and parasitic diseases [6].

Strategies for addressing these threats include both immediate actions, where resources are swiftly deployed to fill urgent knowledge gaps, and more extended strategic initiatives. This long-term research is essential for enhancing our capacity to respond promptly to emerging microbial threats and for helping us to be more proactive in anticipating and preventing their occurrence [7]. Key components of a proactive strategy involve gathering quantitative data on microbial behavior in food (predictive microbiology) and deepening our understanding of microbial physiology [8].

Predictive microbiology can be viewed as the application of research focused on the quantitative microbial ecology of food products. This field is founded on the idea that the reactions of microorganism populations to environmental influences are consistent. By defining environments based on these key factors that impact microbial growth and survival, it becomes feasible to use historical data to forecast how microorganisms will behave in similar settings. The phrase “quantitative microbial ecology” has been proposed as a substitute for “predictive microbiology” [9].

After a significant period of development, predictive microbiology (which encompasses the quantitative microbial ecology of foods) has emerged as a vital component of contemporary food microbiology. This discussion will explore the evolution of predictive microbiology, particularly in relation to interfaces. The term interface carries multiple meanings, describing not only the boundaries of scientific inquiry, such as the growth/no growth interface, but also the intersections between different disciplines that have resulted in important conceptual and technological progress.

 The idea of predictive microbiology is relatively new in its application, although it has existed for some time. Esty and Meyer [10] utilized mathematical methods to assess the survival rates of microorganisms. Additionally, modeling the growth of microbes has been a practice in industrial microbiology since Monod’s work in 1949. Nonetheless, it is important for food microbiology to develop its own set of models rather than replicating those from industrial microbiology, as their goals differ [1].

Predictive microbiology focuses on understanding how microbial growth responds to various environmental factors, which is encapsulated in equations or mathematical models. A database could be created to house raw data and models, allowing for the retrieval of information that can help interpret how processing and transportation methods impact microbial growth [7]. The development, validation, and application of predictive microbiology have been thoroughly examined over the past few decades [11].

Initially, modeling studies primarily focused on the thermal inactivation of pathogenic bacteria. However, subsequent studies shifted their attention to understanding how various constraints affect microbial growth (as opposed to survival or death). These studies often employed a kinetic model approach instead of probability modeling, frequently highlighting temperature as either the sole controlling factor or one among several.

For instance, the temperature dependence model for the growth of Clostridium botulinum showed a strong correlation with the data. Nevertheless, the authors cautioned that “care must be taken at extremes of growth, as no growth may be registered in a situation where growth is indeed possible but has a low probability” [12].

Fundamentals of Predictive Microbiology

The fundamental concept of predictive microbiology is based on the understanding that we can simulate microbial growth and survival using mathematical models. Various criteria influence the effectiveness of these mathematical growth models. By minimizing the number of parameters being analyzed, we can achieve better resolution and predictive relevance. Additionally, employing a mechanistic approach that focuses on a limited number of significant parameters enables us to encompass more stages of microbial growth, thereby enhancing the accuracy of the model [13].

1. Microbial Growth Models

The growth curve of most foodborne bacteria generally shows four distinct phases: an initial lag phase with little to no detectable growth, followed by an exponential (log) phase where rapid cell division occurs, then a plateau phase where growth levels off, and finally, a mortality phase where conditions become unsuitable for further growth [14].

Microbial models are categorized based on various criteria. They can be divided into kinetic and probabilistic models according to the anticipated microbial response; empirical and mechanistic models based on the analysis method; and primary, secondary, and tertiary models depending on the assessed dependent variable [15,16].

 i. Primary Models

Primary models evaluate how bacteria respond over time to certain conditions. Generally, these models strive to depict microbial growth using the least number of parameters possible [17]. They can be empirical, rate growth models, inactivation or survival models, or a mix of these types. These models assess a limited number of intuitive parameters, including the relative growth rate, initial population size, and asymptotic population size.

ii.  Secondary Models

Secondary models investigate the elements that affect the kinetic parameters identified by primary models. They define the connection between primary model parameters and both intrinsic and extrinsic factors like temperature and pH ]18]. While primary models concentrate on estimating changes in microbial populations over time and observing specific responses such as growth rate and lag phase, secondary models evaluate how intrinsic and extrinsic factors of food influence these responses (Perez-Rodriguez et al., 2013). Numerous secondary models have been created to analyze the lag phase and growth rate in relation to one or more environmental or physicochemical factors [19].

iii.  Tertiary Models

Tertiary models are software packages that operate through algorithms, integrating primary and secondary models with a graphical user Interface (GUI) to enhance usability for novice modelers. These models are mainly utilized in the food industry and research to bring together insights from both primary and secondary data. An example of a tertiary model is the Unified Growth Prediction Model (UGPM) software, which is founded on Baranyi Roberts’s primary model and temperature-sensitive secondary models [20].

b. Empirical and Mechanistic Models

 1. Empirical Models

Empirical models are practical and structured frameworks where the relationships among various parameters are expressed through mathematical equations, typically in the form of first or second-degree polynomials [7]. In these models, predictions are generated without taking into account factors like physicochemical parameters that could affect the prediction outcomes. A notable example is the quadratic response surface utilized by Gibson et al. [21].

2. Mechanistic or Deterministic Models

Mechanistic or deterministic models are based on established theories and facilitate the interpretation of responses through known phenomena and processes. These models generally involve fewer parameters, provide a better fit for the data, and offer a more accurate description of the response. They are also recognized for their superior extrapolation capabilities, making them more favorable than empirical models [18].

C. Kinetic and Probabilistic Models

1. Kinetic Models

Kinetic models are designed to ascertain the anticipated response rates, whether it be growth or decline. They are employed to forecast concentration levels linked to specific microbial strains, thereby assessing associated risks (such as infection or intoxication risks) (Stavropoulou and Bezirtzoglou, 2019). Examples include the Gompertz and square root models, which illustrate response rates like lag time, specific growth rate, and maximum population density, as well as inactivation or survival models that depict destruction or survival over time [22].

2. Probability Models

Probability models, on the contrary, are associated only with the probability of growth or toxin production and do not predict the rate at which this occurs [23]. They are used to show the absolute limit of microbial growth within specific environments and demonstrate stress threshold levels, which may limit growth but ultimately permit it [24].

Applications of Predictive Microbiology in Food Safety

1. Quality Control of Food Products.

Predictive microbiology can be utilized to confirm the success of microbial inactivation methods like drying, heat treatment, and refrigeration. The food processing sectors that produce yogurt, milk, wine, and sous vide items must strictly follow specific refrigeration temperatures and heat treatment protocols [25]. Unfortunately, these treatment methods are frequently insufficient or poorly managed, resulting in microbial proliferation. For example, in the dairy sector, the inadequacy of heat treatments as the sole control measure against fungal spore germination has been emphasized; this is attributed to the capacity of certain spore-forming fungal pathogens to endure pasteurization or flourish at lower temperatures. Organisms such as Bacillus sporothermophilus are recognized for their ability to withstand the extremely high temperatures achieved during heat treatments [26]. Numerous research investigations into cold chains have also uncovered significant noncompliance with regulations during distribution, retail, and storage phases, affecting the final consumer. This situation underscores the urgent need for quality control checkpoints throughout the food supply chain.

 A study conducted by Gougouli et al. [27] revealed that the growth of microbes in yogurt, indicated by signs such as mycelial formation, is greatly affected by factors like warehousing duration, storage temperature, and the specific microbial strain involved. As a result, predictive models that can forecast fungal growth by examining these variables are extremely beneficial in the dairy sector. These models can improve quality control processes by incorporating them into the final testing of yogurt products for fungal contamination before their release into the market. In these scenarios, predictive models assist in identifying the best conditions for the growth of different microbial species during end-product challenge testing.

It is crucial to understand that various microbial contaminants thrive under different conditions. Thus, the choice of environmental conditions during end-challenge tests must be grounded in solid scientific evidence, as arbitrary selections can result in false negatives and the unintended sale of contaminated products. Additionally, other predictive models within the dairy industry have assessed the impact of factors such as temperature, pH, water activity, and inoculum size on the proliferation of Listeria monocytogenes in milk, as well as the growth of Yersinia enterocolitica in Camembert-type cheese [28].

2. Risk Assessment and Management.

 Risk assessment is a structured and scientific approach aimed at evaluating the human risk linked to exposure to foodborne hazards [29]. It consists of four sequential phases: hazard identification, exposure assessment, hazard characterization, and risk characterization.    Data from MRA play a crucial role in shaping policies and legislation concerning the most significant foodborne pathogens [30].

Traditional methods for assessing risks associated with food products and additives are deterministic [31]. These methods operate under the premise that estimated parameters remain constant, even though these parameters are actually variable and their measurements carry uncertainty. As a result, the differences observed between challenge tests and laboratory assays can lead to flawed decisions when solely relying on deterministic methods for decision-making. Therefore, it is essential to adopt more sensitive techniques, such as predictive modeling. By implementing predictive models within the food industry, it becomes possible to estimate microbial hazards from production through to final consumption (Membre and Guillou, 2016). The insights gained from these models can then inform decisions regarding acceptable levels of microbial exposure and the necessary measures to reduce risk for the end consumer [33].

3. HACCP (Hazard Analysis and Critical Control Point) Systems.

The HACCP is a food safety framework aimed at identifying and preventing potential issues throughout the stages of food production, distribution, and consumption.

It utilizes a systematic method to pinpoint pathogens in raw materials and processing entry points, apply suitable techniques for their removal, and identify possible problems with the final product due to improper handling. Predictive food microbiology is vital in the execution of the HACCP concept.

The use of Quantitative Microbial Risk Assessment (QMRA) is crucial in hazard analysis as it assesses potential microbial risks within the food chain and identifies critical control points (CCPs) and critical limits (CL). Predictive models can also be created to assess systems set up to monitor CCPs and confirm the effectiveness of the HACCP framework. Testing models can forecast CCPs by establishing levels for various parameters that allow microbial growth. Furthermore, they can estimate different levels of microbial behavior to recommend acceptable levels and thresholds for critical limits. Therefore, the integration of HACCP with predictive models holds significant promise in decision-making [34]

4. Shelf-Life Determination.

 Traditional microbiological techniques for assessing shelf life can be quite labor-intensive, necessitating significant bacterial cell growth before any visible spoilage reactions are noticed. On the other hand, newer methodologies might demand sophisticated and costly equipment.

Nevertheless, insights gained from predictive modeling of microbial behavior in food offer a solid basis for creating devices that can monitor food shelf life during warehousing, transportation, distribution, and retail.

Effective shelf-life prediction models must meet several criteria, including the identification of spoilage reactions (SRs) such as slime formation, color alterations, and unpleasant odors.

They should also pinpoint the specific microorganisms (SSOs) that cause these spoilage reactions and examine the spoilage domain (SD), which refers to the environmental conditions that allow a particular SSO to thrive and function. The process of developing and validating a shelf-life prediction model involves conducting experiments that demonstrate the spoilage organisms, their reactions, and the spoilage domain. Following this, modeling microbial behavior within this domain is essential for establishing the ‘minimum spoilage level,’ which is the concentration of SSO needed to lead to product rejection. Importantly, the metabolic activities of spoilage organisms, rather than just their numbers, significantly influence spoilage (Dalgaard and Henrik Huss, 2020).

A variety of predictive models have been created for shelf-life research in the food sector. For instance, there are models for forecasting shelf life in yogurt by Mataragas et al. [36]; minced beef by Limbo et al. [37]; and Nutri cereal baby food by Rasane [38]. Furthermore, research has concentrated on the effects of specific bacteria on spoilage in certain foods, such as Pseudomonas in pork and poultry by Bruckner et al. [39] and lactic acid bacteria in cooked ham by Kreyenschmidt [40]. These predictive models improve food quality and safety by offering valuable insights into the elements that affect shelf life and the potential mechanisms of spoilage.

Major Limitations of Predictive Models.

 Models offer numerous advantages in the decision-making process. However, it is important to remember that they are, at best, simplified representations of intricate biological processes. As a result, predictions derived from model outcomes should be approached with caution, taking into account past experiences and other microbial ecological factors that the models may not fully encompass. For example, it is vital to recognize that models can only be extrapolated for values within the experimental ranges for which they were designed, particularly concerning parameters like temperature or water activity. This limitation occurs because models, especially empirical ones, are created by fitting observed data and may only partially replicate actual microbial behavior. Research conducted by Fakruddin et al. [41] also indicates that certain models may forecast faster growth rates than what is observed in real-life situations. This inconsistency has been linked to the fact that most models are developed using laboratory media, which diminishes their predictive relevance in the food industry, even though they are validated in food contexts [42]. Additionally, professionals in the food industry have pointed out the mismatch of models created under stable environmental conditions when assessing real-life products that undergo varying conditions of temperature, pH, and water activity. Furthermore, predictive models cannot account for all the variables influencing food spoilage and microbial growth [9]. Most models focus on just one or a few parameters related to food spoilage [43].

Challenges and Future Directions:  Most microbial models tend to be simplistic, mainly focusing on observable changes in response to environmental dynamics.

While these models have shown effectiveness in predicting parameters like growth or inactivation rates within optimal temperature ranges, their capacity to model intricate phenomena such as lag phases and adaptive responses is still uncertain.

As a result, there is an increasing demand for more adaptable and mechanistic next-generation models that incorporate cellular information to decode complex microbial behavior [44,45].

New modeling strategies should focus on understanding the behavior of individual cells to gain a more accurate grasp of the dynamics of entire populations. Recent research has highlighted the importance of individual cell heterogeneity as a key source of variability [46].

Considering that food contamination can happen with very few pathogenic bacteria, grasping the mechanisms of individual cell behavior is essential for effective microbial risk assessment [47].

To tackle these issues, research is currently being conducted on stochastic individual cell modeling grounded in systems biology approaches. These investigations have uncovered the variability of bacterial cells within a clone and the link between gene expression and phenotypic expression [48].

Next-generation models also need to examine complex interactions, such as intraspecies diversity present in spore-forming bacteria and their significance in risk assessment. This aspect has not been sufficiently modeled in existing approaches [49].

The combination of Predictive Microbiology with various other technologies.

 Recent advancements in whole genome sequencing (WGS) technologies, along with developments in genomics and metagenomics, are transforming the way we detect, characterize, and identify pathogens in food safety. This has led to the emergence of a new field called foodomics [50]. Additionally, proteomics and metabolomics techniques are now being utilized to identify bacterial toxins and mycotoxins in food products, making omics-based tools essential for risk assessment and food safety monitoring in the 21st century. The incorporation of WGS into surveillance programs has proven effective in addressing outbreaks of Listeriosis and Salmonellosis that might have gone undetected previously [51]. By leveraging WGS technologies in predictive microbiology, researchers can obtain critical information such as serotype, virulence factors, antimicrobial resistance genes, and genetic variations like single-nucleotide polymorphisms (SNPs) [52]. This facilitates a more accurate risk assessment that effectively correlates microbial genotype with varying clinical outcomes (Njage et al., 2019). For example, Shiga toxin-producing Escherichia coli is linked to a range of clinical outcomes, from diarrhea to hemolytic uremic syndrome (HUS) and other long-term effects. Utilizing WGS technologies in risk assessment allows for precise connections between specific hazards and their associated clinical outcomes [53]. This approach was exemplified in a study by Pielaat et al. [54], where molecular data from WGS were used to characterize hazards associated with Escherichia coli O157:H7 [55].

 Furthermore, metagenomics serves as a valuable approach to explore the interactions among microorganisms within a community [56]. Microbial communities are intricate systems, where the actions of one microbe can significantly affect the behaviors of others. By employing metagenomics to investigate these interactions, we can create more advanced predictive models that reflect the dynamics of microbial communities.

Additionally, a powerful tool that holds great promise for enhancing current predictive modeling systems is artificial intelligence and machine learning (ML) technology. The development of a predictive model produces extensive datasets that are often too complex for traditional statistical methods to analyze. Artificial intelligence and machine learning can effectively process these large datasets, uncovering hidden patterns and correlations. Thus, artificial intelligence enables the creation of more refined and precise models that account for various factors influencing microbial growth and behavior.

Conclusion

The review highlighted the extensive use of predictive models in food microbiology to analyze the growth of microorganisms in food items. However, these models face challenges when it comes to accurately representing complex microbial interactions, especially in foods with diverse bacterial populations.

Nevertheless, the incorporation of cutting-edge technologies like whole genome sequencing (WGS), metagenomics, artificial intelligence, and machine learning, along with devices utilizing robotics, the Internet of Things, and time-temperature indicators, has significantly enhanced the efficiency and precision of these models.

In particular, the application of machine learning models has proven to be extremely advantageous in creating more sophisticated models. Therefore, ongoing research and development in these areas are crucial for enhancing food safety and minimizing the risk of microbial contamination in food products.

Microbial food safety is a significant concern for the industry, government, and the general public, with each group expecting a safe and wholesome food supply as a fundamental principle of a developed society. It is evident that predictive microbiology plays a crucial role in fulfilling this expectation, and it has already become a vital component of contemporary food microbiology. This recognition has been achieved through the continuous enhancement of our understanding of the quantitative microbial ecology of foods and by creating connections with other fields to utilize the insights from predictive models. Its effectiveness will be further improved when predictive microbiology is acknowledged as a rapid and efficient method of food processing, preservation, and product optimization.

DECLARATION OF COMPETING INTEREST

I hereby declare that there is no conflict of interest in this review

Acknowledgement

I thank God Almighty who gave us life to be strong and be able to do this work. I also acknowledge my Head of Department, Prof. H.C. Okereke, for his contribution and guidance. 

REFERENCES

1. Baranyi, J., & Roberts, T.A., (2014). A dynamic approach to predicting bacterial growth in food. Int. J. Food Microbiol. 23, 277 – 294
2. Osterholm, M. T. and Potter, M. E. (2017) “Irradiation pasteurization of solid foods: taking food safety to the next level,” Emerging Infectious Diseases, vol. 3, no. 4, p. 575,
3. Kobets, T. Smith, B. P. and Williams, G. M.  (2022) “Food-borne chemical carcinogens and the evidence for human cancer risk,” Foods, vol. 11, no. 18, p. 2828, 2022.
4. Lawley, R. Curtis, L. and Davis, J. (2012) The Food Safety hazard Guidebook, Royal Society of Chemistry, London, UK,
5. Lee H. and Yoon, Y.  (2021) “Etiological agents implicated in foodborne illness world wide,” Food science of animal resources, vol. 41, no. 1, p. 1,
6. World Health Organization (2015) “WHO estimates of the global burden of foodborne diseases,” 2015,https://apps.who.int/iris/bitstream/handle/10665/199350/9789241565165_eng.pdf?sequence=1.
7. Buchanan, R.L., (2017). Identifying and controlling emerging foodborne pathogens. Emerg. Infect. Dis. 3, 517 – 522.
8. McMeekin, T.A., Brown, J.L., Krist, K., Miles, D., Nuemeyer, K., Nichols, D.S., Olley, J., Presser, K., Ratkowsky, D.A., Ross, T., Salter, M. and Soontranon, S., (2017). Quantitative microbiology: a basis for food safety. Emerging Infect. Dis. 3, 541 – 549
9. Ross, T.  and McMeekin, T. A.  (2013) “Modeling microbial growth within food safety risk assessments,” Risk Analysis: International Journal, vol. 23, no. 1, pp. 179–197.
10. Esty, J.R., Meyer, K.F., (2022). The heat resistance of the spore of B. botulinium and allied anaerobes XI. J. Infect. Dis. 31, 650 – 663.
11. McMeekin, T.A., Olley, J., Ross, T., and Ratkowsky, D.A., (2013). Predictive Microbiology: Theory and Application. Research Studies Press, Taunton, UK
12. Graham, A, and Lund, B.M. (2013) The effect of temperature on the growth of non proteolytic type B Clostridium botulinum. Letters in Applied Microbiology; 16:158-60
13. Esser, D. S. Leveau, J. H. and Meyer, K. M.  (2015) “Modeling microbial growth and dynamics,” Applied Microbiology and. Biotechnology, vol. 99, pp. 8831–8846,
14. Peleg, M.  and Corradini, M. G. (2011) “Microbial growth curves: what the models tell us and what they cannot,” Critical Reviews in Food Science and Nutrition, vol. 51, no. 10, pp. 917–945.
15. Baranyi, I. and da Silva, N. B. (2017)“The use of predictive models to optimize risk of decisions,” International Journal of Food Microbiology, vol. 240, pp. 19–23, 2017.
16. Jiang, M. P. Zheng, S. Y.  and Wang, H. (2019) “Predictive model of Aflatoxin contamination risk associated with granary-stored corn with versicolorin A monitoring and logistic regression,” Food Additives & Contaminants: Part A, vol. 36, no. 2, pp. 308–319, 2019.
17. Stavropoulou E. and Bezirtzoglou, E. (2019) “Predictive modeling of microbial behavior in food,” Foods, vol. 8, no. 12, p. 654,
18. Li, H. Xie, G. and Edmondson, A. S. (2018) “Review of secondary mathematical models of predictive microbiology,” Journal of Food Products Marketing, vol. 14, no. 2, pp. 57–74,
19. Valero, A. Carrasco, E. and Garcıa-Gimeno, R. M. (2012) “Principles and methodologies for the determination of shelf-life in foods,” Trends in vital food and control engineering, vol. 1, pp. 3–42.
20. Psomas, A. (2012) “Development of mathematical models for spoilage prediction,” Agricultural University of Athens, Athens, Greece,  Ph.D. thesis.
21. Gibson, A. M. Bratchell, N. and Roberts, T. A. (2017) “The effect of sodium chloride and temperature on the rate and extent of growth of Clostridium botulinum type A in pasteurized pork slurry,” Journal of Applied Microbiology, vol. 62, no. 6, pp. 479–490,
22. Yi, Z. and Xie, J.  (2021)“Prediction in the dynamics and spoilage of Shewanella putrefaciens in bigeye tuna (Tunnus obesus) by gas sensors stored at different refrigeration temperatures,” Foods, vol. 10, no. 9, p. 2132,
23. Horowitz, M. Normand, D. Corradini, M. G. and Peleg, M. (2010) “Probabilistic model of microbial cell growth, division, and mortality,” Applied and Environmental Microbiology, vol. 76, no. 1, pp. 230–242.
24. Koutsoumanis, K. Pavlis, A. Nychas, G. J. E. and Xanthiakos, K.  (2010) “Probabilistic model for Listeria monocytogenes growth during distribution, retail storage, and domestic storage of pasteurized milk,” Applied and Environmental Microbiology, vol. 76, no. 7, pp. 2181–2191.
25. Granados, L. R. Perez, F. and Valero, A. (2018) “Food quality management systems in the dairy industry: a case study on the application of predictive microbiology in the microbial quality of milk,” in Technological Approaches for Novel Applications in Dairy Processing, IntechOpen, London, UK.
26. Flint, S. Bremer, P. and Brooks, J (2020) “Bacterial fouling in dairy processing,” International Dairy Journal, vol. 101, Article ID 104593,
27. Gougouli, M. Kalantzi, K. Beletsiotis, E. and Koutsoumanis, K. P. (2021) “Development and application of predictive models for fungal growth as tools to improve quality control in yogurt production,” Food Microbiology, vol. 28, no. 8, pp. 1453–1462,
28. Kowalik J. and Lobacz, A.  (2015) “Development of a predictive model describing the growth of Yersinia enterocolitica in Camembert-type cheese,” International Journal of Food Science and Technology, vol. 50, no. 3, pp. 811–818
29. Codex Alimentarius Commission (2014), Principles and Guidelines for the Conduct of Microbiological Risk Assessment. CAC/GL 30-1999, Food and Agriculture Organization of the United Nations/World Health Organization, Rome/Geneva, Italy, 2014
30. World Health Organization/Food and Agriculture Organization, Risk Assessment of Listeria Monocytogenes in Ready- To-Eat Foods, World Health Organization, Rome, Italy, 2016, http://www.who.int/foodsafety/publications/mra_5/en/.
31. Garcıa-Moraleja, A. Font, G. Mañes, J. and Ferrer, E. (2015)“Analysis of mycotoxins in cofee and risk assessment in Spanish adolescents and adults,” Food and Chemical Toxicology, vol. 86, pp. 225–233.
32. Membre J. M. and Guillou, S. (2016) “Latest developments in foodborne pathogen risk assessment,” Current Opinion in Food Science, vol. 8, pp. 120–126,
33. Gonzalez, S. C. Possas, A. Carrasco E. (2019) “MicroHibro: A software tool for predictive microbiology and microbial risk assessment in foods,” International Journal of Food Micro– biology, vol. 290, pp. 226–236.
34. Benito, S. (2019) “The management of compounds that influence human health in modern winemaking from an HACCP point of view,” Fermentation, vol. 5, no. 2, p. 33,
35. Dalgaard, I. and Henrik  Huss, H. (2020)“Mathematical modeling used for evaluation and prediction of microbial fish spoilage,” in Seafood Safety, Processing, and Biotechnology, pp. 73–89, CRC Press, Boca Raton, FL, USA,.
36. Mataragas, M. Dimitriou, V. Skandamis, P. N. and Drosinos, E. H. “Quantifying the spoilage and shelf-life of yoghurt with fruits,” Food Microbiology, vol. 28, no. 3, pp. 611–616,
37. Limbo, S. Torri, L. Sinelli, N. Franzetti, L. and Casiraghi, E. (2010) “Evaluation and predictive modeling of shelf life of minced beef stored in high-oxygen modified atmosphere packaging at different temperatures,” Meat Science, vol. 84, no. 1, pp. 129–136,
38. Rasane, P. Jha, A. and Sharma, N.  (2015) “Predictive modeling for shelf-life determination of nutricereal-based fermented baby food,” Journal of Food Science and Technology, vol. 52, pp. 5003–5011.
39. Bruckner, I. Albrecht, A. Petersen, B. and Kreyenschmidt, J. (2013) “A predictive shelf life model as a tool for the improvement of quality management in pork and poultry chains,” Food Control, vol. 29, no. 2, pp. 451–460,
40. Kreyenschmidt, J. Hubner, A. Beierle, E. Chonsch, L. Scherer, A. and Petersen, B. (2010) “Determination of the shelf life of sliced cooked ham based on the growth of lactic acid bacteria in different steps of the chain,” Journal of Applied Microbiology, vol. 108, no. 2, pp. 510–520,
41. Fakruddin, M. D. Mazumder, R. M. and Mannan, K. S. B. (2011) “Predictive microbiology: modeling microbial responses in food,” Ceylon Journal of Science, vol. 40, no. 2, pp. 121–131,
42. Ross, T.  Baranyi, J. and McMeekin, T. A. (2020) “Predictive microbiology and food safety,” in Encyclopaedia of Food Microbiology, R. Robinson, C. A. Batt, and P. Patel, Eds., pp. 1699–1710, Academic Press, London, UK.
43. Dalgaard, P. (2015) “Qualitative and quantitative characterization of spoilage bacteria from packed fish,” International Journal of Food Microbiology, vol. 26, no. 3, pp. 319–333, 1995
44. Pin, C. Metris, A. and Baranyi, J.  (2013) “Next generation of predictive models,” in Advances in Microbial Food Safety, J. Sofos, Ed., pp. 498–515, Woodhead Publ, Cambridge, UK,
45. Brul, S. Mensonides, F. I. Hellingwerf, K. J. and de Mattos, M. J. T. (2018) “Microbial systems biology: new frontiers open to predictive microbiology,” International Journal of Food Microbiology, vol. 128, no. 1, pp. 16–21,
46. Papagianeli, S. D. Lianou, A. Aspridou, Z. Stathas, L. and Koutsoumanis, K. (2022) “The magnitude of heterogeneity in individual-cell growth dynamics is an inherent characteristic of Salmonella enterica ser. Typhimurium strains,” Food Research International, vol. 162, Article ID 111991.
47. Tegtmeyer, N. Rohde, M. and Backert, S. (2021) “Clinical presentations and pathogenicity mechanisms of bacterial foodborne infections,” in Microbial Food Safety: An Introduction, pp. 13–31, Springer New York, New York, NY, USA.
48. Ackermann, M.  (2015) “A functional perspective on phenotypic heterogeneity in microorganisms,” Nature Reviews Microbiology, vol. 13, no. 8, pp. 497–508, 2015
49. Augustin, . C. (2021) “Challenges in risk assessment and predictive microbiology of food-borne spore-forming bacteria,” Food Microbiology, vol. 28, no. 2, pp. 209–213,.
50. Giacometti, J. and Josic, D.  (2013) “Foodomics in microbial safety,”TrAC, Trends in Analytical Chemistry, vol. 52, pp. 16–22,
51. Hassena, B. Haendiges, J. and Zormati, S. (2021) “Virulence and resistance genes profiles and clonal relationships of non-typhoidal food-borne Salmonella strains isolated in Tunisia by whole genome sequencing,” International Journal of Food Microbiology, vol. 337, Article ID 108941,
52. Kovac, I., den Bakker, H. Carroll, L. M. and Wiedmann, M. (2017) “Precision food safety: a systems approach to food safety facilitated by genomics tools,” TrAC, Trends in Analytical Chemistry, vol. 96, pp. 52–61,
53. Spinale, M. Ruebner, R. L. Copelovitch, L. and Kaplan, B. S. (2013) “Long-term outcomes of Shiga toxin hemolytic uremic syndrome,” Pediatric Nephrology, vol. 28, pp. 2097–2105,
54. Pielaat, A. Boer, M. P. and Wijnands, L. M. (2015) “First step in using molecular data for microbial food safety risk assessment; hazard identifcation of Escherichia coli O157: H7 by coupling genomic data with in vitro adherence to human epithelial cells,” International Journal of Food Microbiology, vol. 213, pp. 130–138,
55. Carriço, A. Sabat, A. J. Friedrich, A. W. Ramirez, M. and on behalf of the Escmid Study Group for Epidemiological Markers C, (2013) “Bioinformatics in bacterial molecular epidemiology and public health: databases, tools and the next- generation sequencing revolution,” Euro Surveillance, vol. 18, no. 4.
56. Faust, K. (2015) “Metagenomics meets time series analysis: unravelling microbial community dynamics,” Current Opinion in Microbiology, vol. 25, pp. 56–66.