Dynamic model for predicting growth of Salmonella spp. in ground sterile pork

A predictive model for Salmonella spp. growth in ground pork was developed and validated using kinetic growth data. Salmonella spp. kinetic growth data in ground pork were collected at several isothermal conditions (between 10 and 45 °C) and Baranyi model was fitted to describe the growth at each temperature, separately. The maximum growth rates (μ_max) estimated from the Baranyi model were modeled as a function of temperature using a modified Ratkowsky equation. To estimate bacterial growth under dynamic temperature conditions, the differential form of the Baranyi model, in combination with the modified Ratkowsky equation for rate constants, was solved numerically using fourth order Runge-Kutta method. The dynamic model was validated using five different dynamic temperature profiles (linear cooling, exponential cooling, linear heating, exponential heating, and sinusoidal). Performance measures, root mean squared error, accuracy factor, and bias factor were used to evaluate the model performance, and were observed to be satisfactory. The dynamic model can estimate the growth of Salmonella spp. in pork within a 0.5 log accuracy under both linear and exponential cooling profiles, although the model may overestimate or underestimate at some data points, which were generally < 1 log. Under sinusoidal temperature profiles, the estimates from the dynamic model were also within 0.5 log of the observed values. However, underestimation could occur if the bacteria were exposed to temperatures below the minimum growth temperature of Salmonella spp., since low temperature conditions could alter the cell physiology. To obtain an accurate estimate of Salmonella spp. growth using the models reported in this work, it is suggested that the models be used at temperatures above 7 °C, the minimum growth temperature for Salmonella spp. in pork.     

低温条件下冷却猪肉中假单胞菌生长模型的比较分析

为了确定拟合冷却猪肉中假单胞菌低温下生长的最适模型,分别对低温(0、5、10℃)条件下托盘和真空包装冷却猪肉中假单胞菌的生长特点进行分析,应用修正的Gompertz、Baranyi及Huang模型对其进行拟合,通过残差和拟合度(RSS、AIC、RSE)等统计指标比较3种模型的拟合能力,分析不同模型拟合假单胞菌生长的差别。结果表明：低温托盘和真空包装条件下假单胞菌在延滞期出现了明显的菌数下降现象,随后呈现“S”形生长;0℃条件下Baranyi模型拟合出最小的RSS、AIC、RSE值,分别是5.2933、－54.0428、0.1708;而修正的Gompertz模型和Huang模型分别在5℃和10℃条件下拟合出最小的RSS、AIC、RSE值,分别是17.7372、－18.9098、0.5068和13.0410、－22.4848、0.4207。拟合冷却猪肉中假单胞菌生长的最适模型0℃是Baranyi模型,5℃是修正的Gompertz模型,10℃是Huang模型。因此,在冷却猪肉腐败菌预测时,不同温度条件下应该选择最适合的模型而不是单一的模型来预测假单胞菌的生长。     

Predictive model for growth of Clostridium perfringens during cooling of cooked uncured meat and poultry

Comparison of Clostridium perfringens spore germination and outgrowth in cooked uncured products during cooling for different meat species is presented. Cooked, uncured product was inoculated with C. perfringens spores and vacuum packaged. For the isothermal experiments, all samples were incubated in a water bath stabilized at selected temperatures between 10 and 51 °C and sampled periodically. For dynamic experiments, the samples were cooled from 54.4 to 27 °C and subsequently from 27 to 4 °C for different time periods, designated as x and y hours, respectively. The growth models used were based on a model developed by Baranyi and Roberts (1994. A dynamic approach to predicting bacterial growth in food. Int. J. Food Micro. 23, 277-294), which incorporates a constant, referred to as the physiological state constant, q₀. The value of this constant captures the cells’ history before the cooling begins. To estimate specific growth rates, data from isothermal experiments were used, from which a secondary model was developed, based on a form of Ratkowsky’s 4-parameter equation. The estimated growth kinetics associated with pork and chicken were similar, but growth appeared to be slightly greater in beef; for beef, the maximum specific growth rates estimated from the Ratkowsky curve was about 2.7 log₁₀ cfu/h, while for the other two species, chicken and pork, the estimate was about 2.2 log₁₀ cfu/h. Physiological state constants were estimated by minimizing the mean square error of predictions of the log₁₀ of the relative increase versus the corresponding observed quantities for the dynamic experiments: for beef the estimate was 0.007, while those for pork and chicken the estimates were about 0.014 and 0.011, respectively. For a hypothetical 1.5 h cooling from 54 °C to 27° and 5 h to 4 °C, corresponding to USDA-FSIS cooling compliance guidelines, the predicted growth (log₁₀ of the relative increase) for each species was: 1.29 for beef; 1.07 for chicken and 0.95 log₁₀ for pork. However, it was noticed that for pork in particular, the model using the derived q₀ had a tendency to over-predict relative growth when the observed amount of relative growth was small, and under-predict the relative growth when the observed amount of relative growth was large. To provide more fail-safe estimate, rather than using the derived value of q₀, a value of 0.04 is recommended for pork.     

Predictive model for growth of Clostridium botulinum from spores at temperatures applicable to cooling of cooked ground pork

Cooling deviations and temperature abuse are two main reasons leading to the risk of Clostridium botulinum outgrowth in cooked pork. The aim of this research was to create a model that could be used to estimate C. botulinum growth from spores in cooked pork at temperatures similar to those used to chill cooked pork in processing facilities and food establishments. A cocktail of proteolytic C. botulinum types A and B consisting of five strains per type were used to inoculate pork to a final spore concentration of approximately 2 log CFU/g and cooked to 71 °C to heat shock the spores and kill vegetative microbes. The growth of C. botulinum was established at constant storage temperatures from 10 to 46 °C. C. botulinum growth was also studied under dynamic temperature conditions with cooling set to start at 54.4 °C and end at 4.4 °C or 7.2 °C in monophasic or biphasic cooling profiles, respectively. Growth parameters were estimated using the Baranyi model as a primary model and growth rates were fitted using the modified Ratkowsky secondary model with respect to temperature. The R² values ranged from 0.7653 to 0.9995 indicating that the Baranyi primary model was well suited to the growth data. The modified Ratkowsky secondary model's R² was 0.9653 and its root mean square error (RMSE) was 0.0687. All 11 prediction error values computed were within the limit of acceptable prediction zone (−1.0 to 0.5) suggesting a good fit of the model. The predictive model can provide information for the safety of cooked pork exposed to longer chilling times or for customized process schedule development as cooling of larger diameter products presents a processing challenge in the meat process operations.     

Effect of fermentation temperature on the microbial and physicochemical properties of silver carp sausages inoculated with Pediococcus pentosaceus

The effect of fermentation with Pediococcus pentosaceus at different temperatures ranging from 15 to 37 °C on the quality characteristics of silver carp sausages was investigated. Higher temperature stimulated the rapid growth of lactic acid bacteria, resulting in a rapid decline in pH, and consequently suppressed the growth of Pseudomonas, Micrococcaceae and Enterobacteriaceae. However, increasing fermentation temperature gave a progressive increase in total volatile basic nitrogen and biogenic amines in fermented silver carp sausages. Histamine was the main biogenic amine, exceeding 100 mg/kg after 48 h of fermentation at temperatures above 30 °C. Higher content of non-protein nitrogen and α-amino nitrogen correlated with the electrophoretic studies, which showed that proteolysis of high molecular weight myofibrillar and sarcoplasmic proteins was more prominent at higher fermentation temperatures. Products fermented at 23–30 °C showed greatest consumer preference and most favourable textural properties.     

Development and survival of the cheese mites, Acarus farris and Tyrophagus neiswanderi (Acari: Acaridae), at constant temperatures and 90% relative humidity

Two species of acarid mites, Acarus farris and Tyrophagus neiswanderi, have been identified infesting Cabrales cheese in an Asturian maturing cave, the former being the prevalent species. The developmental rate and survival of immature stages of these mites were examined at constant temperatures, ranging from 7 to 29.7 °C for A. farris, and 10 to 31 °C for T. neiswanderi, and a relative humidity (r.h.) of 90±5%. The larval stage of A. farris was particularly susceptible to low and high temperatures with 81.7% and 95.2% mortality at 7 and 29.7 °C, respectively. Tyrophagus neiswanderi larvae also showed the greatest mortality at extreme temperatures among immature stages, though at a lower level than for A. farris (8.6% and 25.6% at 10 and 31 °C, respectively). The optimal temperature for development appeared to be 27-28 °C for both species and the developmental rates were higher for A. farris than T. neiswanderi within the range of the cooler temperatures prevalent in the cheese-maturing caves. The nonlinear Logan type-III model provided the best fit for the relationship between developmental rates and temperature (R_a²>0.99) for all immature stages of A. farris, whereas the development of T. neiswanderi was better described by the Lactin model (R_a²>0.97). The lower and upper developmental threshold temperatures predicted for each stage of A. farris were 3-4 °C lower than those predicted for T. neiswanderi. The differential temperature-development rate for each species might explain the greater abundance of A. farris compared to T. neiswanderi. Furthermore, manipulation of temperature based on modeling predictions may well be used to control mite populations during the cheese maturing process.     

Thermal inactivation of Yersinia enterocolitica in pork slaughter plant scald tank water

The objective of this study was to establish the time–temperature combinations required to ensure the thermal inactivation of Yersinia enterocolitica during scalding of pork carcasses. A 2 strain cocktail of Y. enterocolitica (bioserotypes 2/O:5,27 and 1A/O:6,30) was heat treated at 50, 55 and 60 °C in samples of scald tank water obtained from a commercial pork slaughter plant. Samples were removed at regular intervals and surviving cells enumerated using (i) Cefsulodin–Irgasan–Novobiocin Agar (CIN) supplemented with ampicillin and arabinose and (ii) Tryptone Soya Agar (TSA), overlaid with CIN agar with ampicillin and arabinose. The data generated was used to estimate D- and z-values and the formula D_x = log^− 1(log D₆₀ − ((t₂ − t₁)/z)) was applied to calculate thermal death time–temperature combinations from 55 to 65 °C. D₅₀, D₅₅ and D₆₀-values of 45.9, 10.6 and 2.7 min were calculated from the cell counts obtained on CIN agar, respectively. The corresponding D-values calculated from the TSA/CIN counts were 45.1, 11 and 2.5 min, respectively. The z-value was 7.8. It was concluded that a time–temperature combination of 2.7 min at 60 °C is required to achieve a 1 log reduction in Y. enterocolitica in pork scald tank water. The predicted equivalent at 65 °C was 0.6 min. This study provides data and a model to enable pork processors to identify and apply parameters to limit the risk of carcass cross-contamination with Y. enterocolitica in pork carcass scald tanks.     

Growth of Carnobacterium divergens M35 and production of Divergicin M35 in snow crab by-product, a natural-grade medium

This work aimed to study the factors influencing the growth of Carnobacterium divergens strain M35 and its ability to produce a new class IIa bacteriocin (divergicin M35) in various synthetic media and in medium supplemented with snow crab hepatopancreas (SCH), a natural-grade by-product of crustacean processing. C. divergens M35 growth and bacteriocin production in SCH-supplemented medium was evaluated at different temperatures in batch fermentations and under controlled pH. C. divergens M35 was shown to grow well in SCH medium at tested temperatures between 4 and 30 °C, except at 37 °C. Maximum divergicin M35 production was obtained after 10 h of growth in SCH medium at 25 and 30 °C with a total activity of 3.7 × 10⁴ AU mL⁻¹. Less growth was observed at 37 °C, for which a total bacteriocin activity of only 256 AU mL⁻¹ was obtained. The production of divergicin M35 was greatly influenced by the medium composition, especially by the type of added carbon source. The best production of divergicin M35 in SCH medium was observed at a controlled pH of 7.0. This study describes the optimal parameters for the growth of C. divergens M35 and production of divergicin M35 and demonstrates the effectiveness of the marine by-product as a medium supplement for this culture.     

Molecular identification of mesophilic and psychrotrophic bacteria from raw cow's milk

The aim of this study was to use molecular techniques to assess the microbiota of eight raw cow's milk samples at biotype and species level. Sixty-six isolates from raw milk samples were screened by Randomly amplified polymorphic DNA–PCR (RAPD–PCR) biotyping and representative strains of RAPD–PCR profiles were identified by 16S rRNA gene sequencing. Pseudomonas spp. were the most commonly occurring contaminants along with Enterobacteriaceae such as Hafnia alvei, Serratia marcescens and Citrobacter freundii. Moreover, Gram-positive isolates belonging to the genera Staphylococcus and Lactococcus were also found. Experiments of growth at different temperatures showed that more than 50% of the Gram-negative isolates could grow at chill temperatures and that 65% of the Pseudomonas spp. strains grew at 7 °C within 5 days. Only 13 Gram-negative isolates displayed proteolytic activity on milk agar, suggesting that not all the biotypes of milk contaminating species are able to perform this spoilage-associated activity. Among the Gram negative, the proteolytic strains were mainly Peudomonas spp. that displayed the activity at both 7 °C and 20 °C. A reliable molecular identification of raw milk microbiota is important for the study of the microbiological quality of raw milks and for the assessment of the ecology at species level in order to develop improved systems, preventing contamination and having the best conditions for the storage of milk.     

Isolation of Alicyclobacillus and the influence of different growth parameters

Alicyclobacillus species are thermo-acidophilic, endospore-forming bacteria that are able to survive pasteurisation and have been implicated in a number of spoilage incidents involving acidic foods and beverages. The aim of this study was to compare three isolation methods used for the detection of Alicyclobacillus acidoterrestris and to investigate the influence of incubation temperature on the growth of A. acidoterrestris and A. acidocaldarius. Peach juice samples inoculated with A. acidoterrestris K47 were analysed using either the International Federation of Fruit Juice Producers (IFU) Method No. 12 (Method A), which involved spread plating onto Bacillus acidoterrestris (BAT) agar at pH 4.0; Method B, which involved pour plating using potato dextrose agar (PDA) at pH 3.7; or Method C, which made use of membrane filtration followed by incubation on K agar at pH 3.7. The performance of the three methods differed significantly, with the IFU Method No. 12 recovering the highest percentage of cells at 75.97%, followed by Method B at 66.79% and Method C at 3.43%. These findings strengthen the proposal of the IFU for the use of the IFU Method No. 12 as a standard international method for the detection of Alicyclobacillus. To investigate the effect on growth of different incubation temperatures A. acidoterrestris (three strains) and A. acidocaldarius (two strains) were incubated at either 45 °C or 25 °C. Growth at 25 °C was slower and maximum cell concentrations were lower (1 × 10⁵-10⁶ cfu/mL compared to 1 × 10⁷-10⁸ cfu/mL) than at 45 °C for A. acidoterrestris. A. acidocaldarius was unable to grow at 25 °C and cell concentrations decreased by 1-2 logs. Since a growth temperature of 25 °C could not inhibit growth of A. acidoterrestris, cooling to room temperature (20°-25 °C) is not an effective control measure for A. acidoterrestris inhibition.     

Thermal inactivation kinetics of Shiga toxin-producing Escherichia coli in buffalo Mozzarella curd

The use of raw milk in the processing of buffalo Mozzarella cheese is permitted, but the heat treatment used for stretching the curd must ensure that the final product does not contain pathogens such as Shiga toxin-producing Escherichia coli (STEC) that may be present on buffalo dairy farms. This study carried out challenge tests at temperatures between 68°C and 80°C for 2 to 10 min to simulate curd temperatures during the stretching phase. Curd samples were inoculated with 2 STEC strains (serotypes O157 and O26), and their inactivation rates were assessed in the different challenge tests. The curd samples were digested with papain to ensure a homogeneous dispersion of bacteria. The STEC cells were counted after inoculation (range 7.1–8.7 log cfu/g) and after heat treatments using the most probable number (MPN) technique. A plot of log MPN/g versus time was created for each separate experiment. The log linear model with tail was used to provide a reasonable fit to observed data. Maximum inactivation rate (k_max, min⁻¹), residual population (log MPN/g), decimal reduction time (min), and time for a 4D (4-log₁₀) reduction (min) were estimated at each temperature tested. A 4D reduction of the O26 STEC strain was achieved when curd was heated at 68°C for 2.6 to 6.3 min or at 80°C for 2.1 to 2.3 min. Greater resistance was observed for the O157 strain at 68°C because k_max was 1.48 min⁻¹. The model estimates can support cheesemakers in defining appropriate process criteria needed to control possible STEC contamination in raw milk intended for the production of Mozzarella.     

High-pressure destruction kinetics of Clostridium sporogenes ATCC 11437 spores in milk at elevated quasi-isothermal conditions

The high-pressure sterilization establishment requires data on isobaric and isothermal destruction kinetics of baro-resistant pathogenic and spoilage bacterial spores. In this study, Clostridium sporogenes 11437 spores (10⁷ CFU/ml) inoculated in milk were subjected to different pressure, temperature and time (P, T, t) combination treatments (700–900 MPa; 80–100 °C; 0–32 min). An insulated chamber was used to enclose the test samples during the treatment for maintaining isobaric and quasi-isothermal processing conditions. Decimal reduction times (D values) and pressure and temperature sensitivity parameters, Z_T (pressure constant) and Z_P (temperature constant) were evaluated using a 3 × 3 full factorial experimental design. HP treatments generally demonstrated a minor pressure pulse effect (PE) (no holding time) and the pressure hold time effect was well described by the first order model (R² > 0.90). Higher pressures and higher temperatures resulted in a higher destruction rate and a higher microbial count reduction. At 900 MPa, the temperature corrected D values were 9.1, 3.8, 0.73 min at 80, 90, 100 °C, respectively. The thermal treatment at 0.1 MPa resulted in D values 833, 65.8, 26.3, 6.0 min at 80, 90, 95, 100 °C respectively. By comparison, HP processing resulted in a strong enhancement of spore destruction at all temperatures. Temperature corrected Z_T values were 16.5, 16.9, 18.2 °C at 700, 800, 900 MPa, respectively, which were higher than the thermal z value 9.6 °C. Hence, the spores had lower temperature sensitivity at elevated pressures. Similarly, corrected Z_P values were 714, 588, 1250 MPa at 80, 90, 100 °C, respectively, which illustrated lower pressure sensitivity at higher temperatures. By general comparison, it was concluded that within the range operating conditions employed, the spores were relatively more sensitive to temperature than to pressure.     

Inactivation of Saccharomyces cerevisiae in pineapple, grape and cranberry juices under pulsed and continuous thermo-sonication treatments

Pineapple, grape and cranberry juice were thermo-sonicated (24 kHz, 400 W, 120 μm) at 40 °C, 50 °C and 60 °C during 10 min at continuous and pulsed mode. Inactivation of Saccharomyces cerevisiae was tested from 0 to 10 min; color and pH were measured. Survivor’s curves were fitted with Weibull distribution, four parameter model and modified Gompertz equation. The acoustic energy (AE) was also calculated. S. cerevisiae was inactivated in the treatments at 60 °C, with the continuous mode being more effective. Grape juice showed total inactivation (7-log) after 10 min. Results showed that pH and color changed significantly (p < 0.05); ultrasound may promote chemical reactions and extract some components. The modified Gompertz equation showed the best fit. Energy analysis showed that pineapple juice (4287.02 mW/ml) required a higher amount of energy; grape juice showed the lowest value (3112.13 mW/ml). Ultrasound represents a viable option for juice pasteurization.     

Modeling of pullulan fermentation by using a color variant strain of Aureobasidium pullulans

Pullulan, which is comprised of glucose units, is a simple linear polysaccharide produced by Aureobasidium pullulans. Pullulan has long been used in various applications such as blood plasma substitutes, food additives, adhesive additives, flocculants, and even environmental pollution control agents. Mathematical models of biomass, pullulan, and sucrose profiles during fermentation not only provide information about the kinetic-metabolic nature of pullulan, but also facilitate the control and optimization of pullulan production. In this study, several models were modified and tested in order to describe biomass, pullulan, and sucrose profiles during batch fermentation using a color variant strain of A. pullulans. The results demonstrated that the modified Gompertz model can serve as a universal equation to fit biomass production, pullulan production, and sucrose consumption. Furthermore, validation of this modified Gompertz model indicated that biomass (slope = 1.00, R² = 0.991), pullulan (slope = 1.10, R² = 0.991), and sucrose (slope = 0.96, R² = 0.991) were all predicted accurately.     

Effect of natural microbiota on growth of Salmonella spp. in fresh pork – A predictive microbiology approach

This study was undertaken to model and predict growth of Salmonella and the dominating natural microbiota, and their interaction in ground pork. Growth of Salmonella in sterile ground pork at constant temperatures between 4 °C and 38 °C was quantified and used for developing predictive models for lag time, max. specific growth rate and max. population density. Data from literature were used to develop growth models for the natural pork microbiota. Challenge tests at temperatures from 9.4 to 24.1 °C and with Salmonella inoculated in ground pork were used for evaluation of interaction models. The existing Jameson-effect and Lotka–Volterra species interaction models and a new expanded Jameson-effect model were evaluated. F-test indicated lack-of-fit for the classical Jameson-effect model at all of the tested temperatures and at 14.1–20.2 °C this was caused by continued growth of Salmonella after the natural microbiota had reached their max. population density. The new expanded Jameson-effect model and the Lotka–Volterra model performed better and appropriately described the continued but reduced growth of Salmonella after the natural microbiota had reached their max. population density. The expanded Jameson-effect model is a new and simple species interaction model, which performed as well as the more complex Lotka–Volterra model.     

TBARS predictive models of pork sausages stored at different temperatures

2-Thiobarbituric acid reactive substances (TBARS) is an important quality index for pork sausages. To study this in pork sausages during storage, kinetic models were developed to predict TBARS content changes of pork sausages at different temperatures. The predictive models of TBARS content with respect to storage time and temperature were developed based on primary and Arrhenius equations. The regression coefficients (R² > 0.95) indicated the acceptability of the primary reaction and Arrhenius model for predicting TBARS content changes of pork sausages. The activation energy (E_A) of TBARS is 14.12 kJ mol^− 1, and the corresponding rate constant (k₀) is 9.262 × 10¹⁰. Relative errors between predicted and measured values of TBARS content are all within ± 8%. Thus, the established model could effectively predict the TBARS content of pork sausages between 5 and 35 °C during storage.     

Survival of Pseudomonas fluorescens on beef carcass surfaces in a commercial abattoir

The influence of a commercial chilling process (18 h at 10 °C followed by up to 78 h at 2 °C) on Pseudomonas fluorescens inoculated on beef carcass surfaces at four sites, neck (NE), outside round (OR), brisket (BR) and foreshank/brisket (FB) before chilling (“hot inoculated”) or after chilling for 24 h (“cold inoculated”) was investigated. Pseudomonas counts increased significantly at all sites on “hot inoculated” carcasses during storage, but on “cold inoculated” carcasses, counts declined or remained unchanged. On hot and cold inoculated carcasses, differences in Pseudomonas growth or survival were demonstrated between sites. No clear relationships were observed between Pseudomonas growth or survival and chiller relative humidity (RH) or surface water activity (a_w) at the different sites. These results were unexpected, and are discussed in relation to environmental factors that affect the growth/survival of P. fluorescens on carcass surfaces during chilling i.e. temperature, RH, and the relationship of these parameters to surface water activity (a_w).     

Modelling the effect of water immersion thermal processing on polyacetylene levels and instrumental colour of carrot disks

C₁₇ polyacetylenes are a group of bioactive compounds present in carrots which have recently gained scientific attention due to their cytotoxicity against cancer cells. In common with many bioactive compounds, their levels may be influenced by thermal processes, such as boiling or water immersion. This study investigated the effect of a number of water immersion time/temperature combinations on concentrations of these compounds and attempted to model the changes. Carrot samples were thermally treated by heating in water at temperatures from 50–100 °C and holding times of 2–60 min. Following heating, levels of falcarinol (FaOH), falcarindiol (FaDOH), falcarindiol-3-acetate (FaDOAc) and Hunter colour parameters (L^∗, a^∗, b^∗) were determined. FaOH, FaDOH, FaDOAc levels were significantly reduced at lower temperatures (50–60 °C). In contrast, samples heated at temperatures from 70–100 °C exhibited higher levels of polyacetylenes (p < 0.05) than did raw unprocessed samples. Regression modelling was used to model the effects of temperature and holding time on the levels of the variables measured. Temperature treatment and holding time were found to significantly affect the polyacetylene content of carrot disks. Predicted models were found to be significant (p < 0.05) with high coefficients of determination (R²).     

Consumer acceptance of roasted peanuts affected by storage temperature and humidity conditions

Consumer acceptance and intensity ratings of roasted peanuts stored at temperatures of 23, 30, 35, and 40 °C, and water activities (a_w) of 0.33, 0.44, 0.54, 0.67 and 0.75 were determined over time. Consumer acceptance ratings, including overall, appearance, color, and texture were affected by storage water activity and time, but not by storage temperature. Consumer intensity ratings of crunchiness were affected by storage water activity and time, but not storage temperature. Aroma acceptance, flavor acceptance, and roasted peanutty and stale/oxidized/rancid intensity ratings of roasted peanuts were dependent on storage temperature, water activity and time.Shelf-life of roasted peanuts was predicted by all consumer attributes (R²>0.60) and was best predicted by aroma acceptance (R²=0.75). Using contour plots with ratings >5.0 for all acceptance attributes, the shelf-life of roasted peanuts stored at 23 °C and between 0.33 and 0.75 a_w was limited by overall acceptance and decreased by approximately 50% with each 0.1 a_w increase. At accelerated temperatures of 30, 35 and 40 °C, shelf-life of roasted peanuts was predominantly limited by flavor acceptance (>5.0), and to a lesser extent by overall acceptance (>5.0). Shelf-life of roasted peanuts stored at accelerated temperatures decreased by 50% or more with each 0.1 a_w increase.