Skip to main content
ARS Home » Northeast Area » Wyndmoor, Pennsylvania » Eastern Regional Research Center » Residue Chemistry and Predictive Microbiology Research » Research » Research Project #430595

Research Project: Development of Predictive Microbial Models for Food Safety using Alternate Approaches

Location: Residue Chemistry and Predictive Microbiology Research

2018 Annual Report


1a. Objectives (from AD-416):
The main goal of this project is to develop and validate more accurate and robust mathematical models and computational algorithms for predicting the growth of human pathogens in processed foods exposed to complex processing, distribution, and storage conditions. This project focuses on applying and improving a new one-step dynamic kinetic analysis methodology and optimization method to generate kinetic models. We will develop models for foods, pathogens, and environmental factors that are not in duplication of the existing models in the USDA PMP or products of other research institutions. We will continue to optimize a new dynamic approach aiming at more accurate and rapid estimation of kinetic parameters by direct construction of predictive models for foodborne pathogens. This direct approach, recently applied in experiments for determining the growth kinetics of Clostridium perfringens, was very accurate in predicting its growth in cooked beef during cooling in the validation studies. We also experimented with a probabilistic approach to predict the growth of C. perfringens. In the next five years, we will continue to optimize the methodology, experimental design, and computational algorithms for determining the growth kinetics of other high-priority pathogens, such as Listeria monocytogenes, Salmonella spp., pathogenic Escherichia coli, C. perfringens, and Bacillus cereus, in various types of food products. We will continue to examine and expand the application of probabilistic simulation methods for process risk assessment, real-time food safety decision-making, and quality control. Furthermore, we will continue to support the scientists in the predictive microbiology community by providing more user-friendly, comprehensive, and robust interactive tools for data analysis for application in research and education. This research will fill the gap between the immense need in the nation for predictive modeling and the availability of highly accurate dynamic and probabilistic modeling methods and tools. Therefore, the specific objectives of this project include: 1: Development and validation of predictive models for growth of high priority pathogens in processed foods. 2: Dynamic simulation and probabilistic modeling of growth of foodborne pathogens in foods. 3: Develop an advanced decision support system and software for predictive microbiology and food safety regulations. 4: Further, expand where necessary the ARS curve-fitting (modeling) program also known as the “Integrated Pathogen Modeling Program (IPMP)”.


1b. Approach (from AD-416):
A new dynamic approach will be developed and optimized to simulate and predict the growth and survival of major foodborne pathogens in meat and poultry products exposed to complex changes in the environmental conditions during heating, cooling, and storage. The research will utilize an advanced computational framework and probabilistic Monte Carlo simulation to analyze the dynamic changes in the population of foodborne pathogens, and will develop an expert decision support system to assist the food industry and regulatory agencies in making scientifically sound food safety decisions for products of concern. This project will continue to improve and upgrade the USDA Integrated Pathogen Modeling Program (IPMP) for data analysis, and develop a new data analysis tool, IPMP Global Fit, that minimizes the global residual errors for curve-fitting of growth and survival curves.


3. Progress Report:
Within this Fiscal Year, progress was made to address the objectives of NP108 Food Safety Component 1 Foodborne Contaminants, Problem F, Predictive Microbiology/Modeling: Data Acquisition and Storage. Experiments were conducted with an emphasis on one-step dynamic modeling to more efficiently and effectively develop predictive models for foodborne pathogens in meat and poultry products. New software codes, based on open-source technologies, were developed for dynamic modeling to achieve significantly improved accuracy for prediction. Detailed progress to achieve the overall objectives is listed below. Growth/No Growth Boundary of Clostridium perfringens in cooked meat (Objs. 1 & 3). The effect of common food ingredients, including sodium chloride (NaCl), sodium lactate (NaL), sodium diacetate, sodium nitrite, and sodium tripolyphosphate (STPP), on the growth and no-growth boundary of C. perfringens in cooked beef was investigated. Different combinations of these ingredients were used to investigate the probability of growth of C. perfringens under optimum temperature conditions. Results showed that proper combinations of NaCl, NaL, and STPP could inhibit the growth of C. perfringens by killing the bacteria, extending the lag phase, or reducing the growth rate. A Growth/No Growth boundary model was developed to calculate the probability of growth of C. perfringens. This model can be used to formulate meat products that may prevent of growth of C. perfringens even under the optimum temperature condition, and thus may be used to control the growth of this pathogen in cooked meats during extended cooling, potentially eliminating the need for cooling. Dynamic modeling of growth of C. perfringens in cooked meats and development of a composite predictive model (Objs. 1, 2, & 3). The growth kinetics of C. perfringens in cooked meats, including ground chicken, roasted chicken, and braised beef, was investigated under dynamic cooling and storage conditions using the one-step dynamic analysis method recently developed at ERRC. This study was conducted in part (roasted chicken and braised beef) by collaboration with Henan Agricultural University in China. One-step dynamic analysis was shown to be more efficient and more accurate than the traditional method, and was tested for the first time with the Baranyi model. In addition, the dynamic method was used to develop a composite model from the data collected in beef, turkey, and chicken meats. This model may be used to predict the growth of C. perfringens in different meat products in the event of cooling deviation. Growth Models of Salmonella in Cooked Meat (Objs. 1 & 2). A 5-strain mixture of Salmonella spp. was inoculated on slices of cooked ham. The sample were stored at 8, 12, and 16°C or pre-expose to one temperature for its lag phase duration and then moved to another temperature, e.g., 12 or 16°C to 8°C, 8 or 16 °C to 12 °C, and 8 or 12°C to 16 °C. The growth curves of Salmonella were used to develop linear growth rate model for each temperature exposure. Growth of a non-toxigenic Clostridium botulinum mutant in cooked beef (Objs. 1 & 3). The growth of a non-toxigenic C. botulinum LNT01 mutant was investigated in cooked ground beef at differential temperature conditions to develop kinetic models and estimate kinetic parameters using the USDA IPMP-Global Fit, a new data analysis software tool developed at ERRC. This new data analysis tool was used to determine the minimum, optimum, maximum growth temperatures, and optimum growth rate of this microorganism in cooked beef. Its growth kinetics was compared with C. sporogenes and C. perfringens in cooked beef during cooling. Dynamic simulation was performed to evaluate the effect of different cooling profiles on the growth of this microorganism. Gaseous chlorine dioxide for decontamination of foodborne pathogens on produce and low-moisture foods (Obj. 1). A study was conducted to evaluate gaseous ClO2 in a simulated pilot scale for decontaminating tomato, blueberry, baby-cut carrot, peppercorn, and almond. Different gas concentrations and treatment times under 70, 85, or 95% relative humidity were tested to determine the most suitable decontamination conditions for individual products. In-situ generation of chlorine dioxide for decontamination of sprout seeds and cantaloupes (Obj. 1). A study was conducted to examine the sequential application of sodium chlorite and hydrochloric acid for decontaminating Salmonella spp., Listeria monocytogenes, and Shiga toxin-producing Escherichia coli on sprout seeds and cantaloupes. The experiments showed that the sequential treatment was more effective in decontaminating surface-bound pathogens. Growth models for Staphylococcus aureus in lettuce (Obj. 1). In collaboration with National Taiwan University, a study was conducted to develop mathematical models for predicting the growth of Staphylococcus aureus in raw lettuce at different storage temperatures (4-40°C) for application in quantitative microbial risk assessment of S. aureus in lettuce salad in Taiwan. Models for growth rate and maximum population of S. aureus in lettuce were developed. A quantitative microbiological risk assessment of the safety of this pathogen in lettuce salad was performed in Taiwan after the models were developed.


4. Accomplishments
1. Growth/No Growth boundary of Clostridium perfringens in cooked meat. C. perfringens is one of the most rapidly-growing foodborne pathogens commonly found in processed meat products and can produce an enterotoxin that causes acute abdominal cramps and diarrhea in consumers. Rapid cooling after cooking is essential to control the growth of this pathogen in meats. ARS scientists at Wyndmoor, Pennsylvania, evaluated the effect of common food ingredients, including sodium chloride (NaCl), sodium lactate (NaL), and sodium tripolyphosphate (STPP), on the Growth and No Growth boundary of C. perfringens in cooked beef. Results showed that proper combinations of NaCl, NaL, and STPP could effectively inhibit the growth of C. perfringens. A Growth/No Growth boundary model was developed to calculate the probability of growth of C. perfringens. This model can be used to formulate meat products that may prevent the growth of C. perfringens even under the optimum temperature condition, and thus may be used to control the growth of this pathogen in cooked meats during extended cooling and to enhance the safety of cooked meat products.

2. Mathematical modeling of growth of Clostridium botulinum in cooked beef using a nontoxigenic strain. C. botulinum is a spore-forming anaerobe that can produce potent neurotoxins during growth. It is difficult to study the growth kinetics of this pathogen in a regular BSL-2 laboratory due to its potential risks. C. botulinum LNT01 is a nontoxigenic mutant of one of the most toxic proteolytic strains (C. botulinum 62A). ARS scientists at Wyndmoor, Pennsylvania, studied the growth kinetics of this mutant as a surrogate of C. botulinum in cooked beef. A dynamic predictive model was developed and compared with the predictive models for C. sporogenes (another surrogate) and C. perfringens. The results of computer simulation showed that, while prolific growth of C. perfringens may occur in ground beef during cooling, no growth of C. botulinum LNT01 or C. sporogenes would occur under the same cooling conditions. The models developed in this study may be used for prediction of the growth and risk assessments of proteolytic C. botulinum in cooked meats during cooling and storage.

3. Dynamic prediction of growth of Salmonella Enteritidis in liquid egg whites. S. Entertidis (SE) is a major foodborne pathogen associated with eggs and egg products. Temperature abuse during refrigeration of contaminated eggs and egg products may allow this pathogen to grow and cause foodborne infections. Applying a new one-step dynamic analysis method, ARS scientists at Wyndmoor, Pennsylvania, investigated the growth and survival of SE in liquid egg whites (LEW) in a range of temperature conditions and demonstrated that this approach is an accurate and efficient method for direct construction of predictive models and estimation of the associated kinetic parameters. Since the mathematical model has been validated, it can be used to predict the growth and survival of SE in LEW during storage and distribution and for conducting risk assessments of this microorganism.

4. Growth models for Staphylococcus aureus in lettuce. S. aureus is a toxin-producing foodborne pathogen found in leafy greens such as lettuce. In collaboration with National Taiwan University, ARS scientists at Wyndmoor, Pennsylvania studied the growth kinetics of S. aureus using USDA-IPMP 2013, and then conducted a quantitative risk assessment (QMRA) of S. aureus in lettuce salad as affected by temperature abuse in Taiwan. The results showed that the risk of poisoning caused by S. aureus in products produced under the Taiwan Certified Agricultural Standards (CAS) are almost 80 times lower that the products without the certification. The results of this study may help the growers and producers to produce safer salad products in Taiwan.


Review Publications
Huang, L. 2017. IPMP Global Fit – A one-step direct data analysis tool for predictive microbiology. International Journal of Food Microbiology. 263:38-48.

Huang, L., Li, C., Hwang, C. 2017. Growth and no-growth boundary of Clostridium perfringens in cooked meat: a probabilistic anaylysis. International Journal of Food Microbiology. 107:248-256.

Huang, L. 2018. Growth of non-toxigenic clostridium botulinum mutant LNT01 in cooked beef: one-step kinetic analysis and comparison with C. sporogenes and C. perfringens. Food Research International. 107:248-256.

Huang, Y., Hwang, C., Huang, L., Wu, V.C., Hsiao, H. 2017. The risk of Vibrio parahaemolyticus infections associated with consumption of raw oysters as affected by processing and distribution conditions in Taiwan. Food Control. 86:101-109.