Development of a predictive model describing the growth of Listeria Monocytogenes in Kimbab

This study is to develop mathematical models to predict the growth of Listeria monocytogenes in kimbab as a function of storage temperature. Kimbab which was inoculated with L. monocytogenes were incubated at 4, 10, 15, and 30°C. The primary model showed a good fit (R²=0.9845 to 0.9967) to a Gompertz equation to obtain specific growth rates (SGR) and lag time (LT) at each temperature. The SGR of L. monocytogenes in the kimbab increased and LT decreased by increasing temperature. Secondary polynomial model was developed using PRISM general nonlinear analysis software for SGR and LT. The secondary models were 0.1479−(0.02457×Temp)+(0.001296 ×Temp²) for SGR and 312.8−(30.21×Temp)+(0.6654 ×Temp²) for LT. This secondary polynomial model was judged as appropriate based on the coefficient of determination (R² of the SGR and LT model=0.9995, 0.9556), the bias factor (B _f of the SGR and LT model=0.97, 0.94), and the accuracy factor (A _f of the SGR and LT model=1.10, 1.68). Reliable predictions of L. monocytogenes SGR and LT in kimbab were based on temperature.     

Development of predictive model for the growth of Staphylococcus aureus in Kimbab

This study was conducted to develop predictive models for the growth of Staphylococcus aureus in kimbab as a function of storage temperatures (7, 10, 12, 14, 16, 20, 25, and 30°C). The growth data were fitted into the modified Gompertz model and the Logistic model, and the goodness-of-fit of primary models was compared using determination of coefficient, mean square error, and Akaike’s information criterion. The modified Gompertz model was found to be more suitable to describe the growth data. Therefore, the growth rate (GR) and lag time (LT) obtained from the modified Gompertz model were employed to establish the secondary models. The newly developed models were validated using root mean square error (RMSE), bias factor (B_f), and accuracy factor (A_f). The results showed that RMSE<0.20 and B_f and A_f values were within the reliable range, which indicated that the presented predictive models can be used to assess the risk of S. aureus infection in kimbab.     

Effect of temperature on the growth kinetics and predictive growth model of Aeromonas hydrophila on squid (Sepioteuthis sepioidea)

This study developed a predictive growth model of Aeromonas hydrophila on fresh squids as a storage temperature (5°C–40°C). The primary models of specific growth rates (SGR) and lag time (LT) fit well (R ²≥0.973). Secondary polynomial models were obtained by non-linear regression and calculated as: SGR=0.05152+0.00337*T+ 0.00039*T²; LT=50.51030?2.56290*T+0.03446*T². The appropriateness of the secondary model was verified by mean square error (MSE; 0.006 for SGR, 0.256 for LT), bias factor (B _f ; 0.999 for SGR, 1.007 for LT), accuracy factor (A _f ; 1.025 for SGR, 1.026 for LT), and coefficient of determination (r ²; 0.991 for SGR, 0.993 for LT). The secondary model is therefore in good agreement with the validation and may be used as a practical prediction for A. hydrophila growth on squid. Ultimately, the developed models are of importance in reducing A. hydrophila levels in the seafood production, processing, and distribution.     

Development of Predictive Models for the Growth of Escherichia coli O157:H7 on Cabbage in Korea

Abstract: Cabbage is the main material of coleslaw, a popular side dish in Korea as well as many other countries. In the present study, the combined effect of temperature (15, 25, and 35 °C) and relative humidity (60%, 70%, and 80%) on the growth of Escherichia coli O157:H7 on cabbage was investigated. The polynomial models for growth rate (GR), lag time (LT), and maximum population density (MPD) estimated from the Baranyi model were conducted with high coefficients of determination (R²> 0.98). Subsequently, performance and reliability of the models were assessed through external validation, employing three indices as bias factor (B_f), accuracy factor (A_f), and the standard error of prediction expressed in percentage (%SEP). The B_f, A_f, and %SEP values of the predictive models for GR were 1.008, 1.127 and 18.70%, while 1.033, 1.187 and 20.79% for LT and 0.960, 1.044 and 5.22% for MPD, respectively. The results demonstrated that the developed secondary models showed a good agreement between the observed and predicted values. Therefore, the established models can be suitable to estimate and control E. coli O157:H7 growth risk on cabbage at some steps from farm to table in Korea as a valuable tool. Practical Application: The combined effect of temperature and relative humidity on the growth or survival of Escherichia coli O157:H7 on cabbage was investigated. The validated predictive models are qualified to provide good predictions for E. coli O157:H7 growth, which can help to conduct the quantitative microbiological risk assessment (QMRA) of E. coli O157:H7 on cabbage from farm to table in Korea.     

Predictive model of Vibrio parahaemolyticus growth and survival on salmon meat as a function of temperature

The growth and survival curves of a strain of pandemic Vibrio parahaemolyticus TGqx01 (serotype O3:K6) on salmon meat at different storage temperatures (range from 0 °C to 35 °C) were determined. In order to model the growth or inactivation kinetics of this pathogen during storage, the modified Gompertz and Weibull equations were chosen to regress growth and survival curves, respectively, and both equations produced good fit to the observed data (the average R² value equals to 0.990 for modified Gompertz and 0.920 for Weibull equation). The effect of storage temperature on the specific growth rate (μ) was modeled by square root type equation, and the relationship between μ and lag time (λ) was described by a rule of μ × λ = constant. The shape factor (n) and scale factor (b) values of the Weibull equations versus the temperature (°C) were plotted and the temperature effects on these parameters were described by two linear empirical equations. The predicted growth and survival curves from the model were compared to real enumeration results, using the correlation coefficient (R²), bias factor (B_f) and accuracy factor (A_f), to assess the performance of the established model. The results showed that the overall predictions for V. parahaemolyticus TGqx01 growth or inactivation on salmon at tested temperatures agreed well with observed plate counts, and the average R², B_f and A_f values were 0.958, 1.019 and 1.035, respectively.     

Modeling the response of Listeria monocytogenes at various storage temperatures in pork with/without electrolyzed water treatment

The objective of this study was to develop a model of the growth of Listeria monocytogenes in pork untreated or treated with low concentration electrolyzed water (LcEW) and strong acid electrolyzed water (SAEW), as a function of temperature. The experimental data obtained under different temperatures (4, 10, 15, 20, 25, and 30°C) were fitted into the modified Gompertz model to generate the growth parameters including specific growth rate (SGR) and lag time (LT) with high coefficients of determination (R² >0.97). The obtained SGR and LT were employed to develop square root models to evaluate the effects of storage temperature on the growth kinetics of L. monocytogenes in pork. The values of bias factor (0.924–1.009) and accuracy factor (1.105–1.186), which were regarded as acceptable, demonstrated that the obtained models could provide good and reliable predictions and be suitable for the purpose of microbiological risk assessment of L. monocytogenes in pork.     

Validation of Lag Time and Growth Rate Models for Salmonella Typhimurium: Acceptable Prediction Zone Method

ABSTRACT: The prediction bias (B_f) and accuracy (A_f) factors are the most widely used measures of performance of predictive models for food pathogens. However, B_f and A_f have limitations that can produce inaccurate assessments of model performance. Consequently, an objective of the current study was to develop a method for quantifying model performance that overcomes limitations of B_f and A_f. Performance of published lag time and growth rate models for Salmonella Typhimurium were evaluated for data used in model development and for data not used in model development but that were inside (interpolation) or outside (extrapolation) the response surface of the models. In addition, performance of published models for growth of Escherichia coli O157:H7 was evaluated for data used in model development. Observed and predicted values were compared using B_f, A_f, and pRE, a new performance factor that quantified the proportion of relative errors (RE) in an acceptable prediction zone from an RE of‐0.3 (fail‐safe) to 0.15 (fail‐dangerous). A decision diagram based on criteria for test data and model performance was used to validate the models. When B_f and A_f were used to quantify model performance, all models were validated. In contrast, when pRE was used to evaluate model performance, 2 models for S. Typhimurium and both models for E. coli O157:H7 failed validation. Overall, pRE was a more sensitive and reliable indicator of model performance than B_f and A_f because unacceptable pRE, which indicated a performance problem, were obtained for 8 of 20 evaluations, all of which had acceptable B_f and A_f. Alimitation of pRE was the inability to distinguish between global and regional prediction problems. However, when used in combination with an RE plot, pRE provided a complete evaluation of model performance that overcame limitations of B_f and A_f.     

Models of the behavior of Escherichia coli O157:H7 in raw sterile ground beef stored at 5 to 46 °C

Escherichia coli O157:H7 can contaminate raw ground beef and cause serious human foodborne illness. Previous reports describe the behavior of E. coli O157:H7 in ground beef under different storage conditions; however, models are lacking for the pathogen's behavior in raw ground beef stored over a broad range of temperature. Using sterile irradiated raw ground beef, the behavioral kinetics of 10 individual E. coli O157:H7 strains and/or a 5- or 10-strain cocktail were measured at storage temperatures from 5° to 46 °C. Growth occurred from 6 to 45 °C. Although lag phase duration (LPD) decreased from 10.5 to 45 °C, no lag phase was observed at 6, 8, or 10 °C. The specific growth rate (SGR) increased from 6 to 42 °C then declined up to 45 °C. In contrast to these profiles, the maximum population density (MPD) declined with increasing temperature, from approximately 9.7 to 8.2 log cfu/g. Bias (B_f) and accuracy (A_f) factors for an E. coli O157:H7 broth-based aerobic growth model (10 to 42 °C) applied to the observations in ground beef were 1.05, 2.70, 1.00 and 1.29, 2.87, 1.03, for SGR, LPD and MPD, respectively. New secondary models increased the accuracy of predictions (5 to 45 °C), with B_f and A_f for SGR, LPD, and MPD of 1.00, 1.06, and 1.00 and 1.14, 1.33, and 1.02, respectively. These new models offer improved tools for designing and implementing food safety systems and assessing the impact of E. coli O157:H7 disease.     

Modeling the influence of water activity and ascospore age on the growth of Neosartorya fischeri in pineapple juice

The ascospores of resistant fungi, Neosartorya fischeri, can survive commercial pasteurization, diminishing the shelf life of these products. The time that the ascospores remain in the environment and the effect that they can cause on mold growth are still unknown. This study is aimed to evaluate the influence of water activity (a_w) from 0.90 to 0.99 and the ascospore age (I) from 30 to 90 days of vitro incubation on the growth of N. fischeri in pineapple juice by mathematical modeling. The growth parameters on pineapple juice: adaptation phase (λ), maximum specific growth rate (μ_max) and maximum diameter reached by the colony (A) were obtained by fitting Modified Gompertz and Logistic models to the experimental data. Both models were able to describe microbial growth in pineapple juice, but the Modified Gompertz model presented a slightly superior performance based on statistical indices (correlation coefficients (R²), mean square error (MSE), Bias Factor and Accuracy Factor). The minimum values of λ and A, calculated by the Modified Gompertz model, were 64.7 h and 6.3 mm, while the maximum values were 178.2 h and 20.8 mm, respectively. The result showed that ascospore age did not influence the growth but a_w was statistically significant to the growth parameters λ and A.     

Modeling survival of high hydrostatic pressure treated stationary- and exponential-phase Listeria innocua cells

High Hydrostatic Pressure (HHP) inactivation (325–400 MPa; 0–20 min; maximum temperature 30 °C) of cells of Listeria innocua CECT 910 was studied in two different growth phases (exponential and stationary), and the corresponding survival curves were obtained for each case. The curves were fitted to two nonlinear models, the modified Gompertz equation and the Baranyi model. The kinetic constants calculated for both models, µ_max and k_max, indicated that cells in exponential growth phase were more sensitive to pressure than those in stationary phase. Both mathematical models were suitable for describing L. innocua HHP survival curves, rendering kinetic constants that increased with increasing pressure. When considering the experimental models validation, both Gompertz and Baranyi predicted in a similar way, however Baranyi had slightly lower A_f (Accuracy factor) and B_f (Bias factor) values, which indicated better prediction values. In summary, both mathematical models were perfectly valid for describing L. innocua inactivation kinetics under HHP treatment.Industrial relevanceThe mathematical models for inactivation and growth of microorganisms are the foundation of predictive microbiology and are used in risk assessments procedures as part of the food safety management system. Besides, these models together with those applied to inactivation of enzymes and destruction of quality factors are essential to optimize processes and thus to lay the foundations for industrial processing. It is therefore necessary to identify generally applicable kinetic models that will produce primary and secondary kinetic parameters and are statistically reliable as a key tool to predict the behaviour of microorganisms, enzymes and quality factors after processing.     

Estimation of feed crude protein concentration and rumen degradability by Fourier-transform infrared spectroscopy

Currently, rapid methods are needed for feed analysis. This study examined the potential of Fourier-transform infrared (FTIR) spectroscopy to predict the nutritional value of a wide range of feeds for ruminants, as an alternative to the in situ technique. Moreover, we investigated whether universal equations could be developed that would allow the low-cost determination of crude protein (CP) concentrations and their kinetics of degradation into the rumen. Protein nutritional values of 663 samples comprising 80 different feed types were determined in terms of concentrations of CP, water-soluble CP (CP_WS), total-tract mobile bag CP digestibility (CP_TTD), and in situ CP degradability, including the rumen soluble fraction (CP_A), the degradable but not soluble fraction (CP_B), rate of CP_B degradation (CP_C), effective degradability (CP_ED), and potential degradability (CP_PD). Infrared spectra of dry samples were collected by attenuated total reflectance from 4000 to 600 cm⁻¹. Models were developed by partial least squares (PLS) regression in a randomly selected subset of samples, and the precision of the equations was confirmed by using an external validation set. Analysis by FTIR spectroscopy was sufficiently sensitive to allow the accurate prediction of sample CP concentration (R² = 0.92) and to classify feeds according to their CP_WS concentrations using universal models (R² = 0.78) that included all sample types. Moreover, substantial improvements in predictions were observed when samples were subdivided in groups. Models for forages led to accurate predictions of CP_WS and fractions CP_A and CP_B (R² > 0.83), whereas models for CP_TTD and CP_ED could be used for screening purposes (R² > 0.67). This study showed that models for protein-rich concentrates alone could also be used for screening according to the feed concentrations of CP_WS, CP_TTD, CP_ED, CP_A, and CP_B, but models for energy-rich concentrates gave relatively poor predictions. The general difficulty observed in predicting CP_C is because of a low correlation between FTIR spectra and the kinetics of CP degradation, which may be the result of large variation in the reference method (i.e., in situ degradation studies) and perhaps also because of the presence of compounds that can modify the CP degradation pattern in the rumen. In conclusion, FTIR spectroscopy should be considered as a low-cost alternative in the feed evaluation industry.     

Modelling of Escherichia coli O157:H7 growth at various storage temperatures on beef treated with electrolyzed oxidizing water

The influence of storage temperature (4, 10, 15, 20, 25, and 30 °C) on the growth of Escherichia coli O157:H7 in beef untreated (control) and treated by acidic electrolyzed oxidizing water (AcEOW) or slightly acidic electrolyzed oxidizing water (SAcEOW) was examined. A Baranyi model was employed to describe growth parameters such as specific growth rate (SGR) and lag time (LT) as a function of storage temperature. SGR increased and LT declined with rising temperatures in all samples. There were no significant differences between the SGR and LT values obtained from beef treated with AcEOW or SAcEOW. Secondary models were established for SGR and LT to evaluate the effects of storage temperature on the growth kinetics of E. coli O157:H7 in treated and untreated beef. Mathematical evaluation was carried out to validate the performance of the developed models.     

Improving Microbial Growth Prediction by Product Unit Neural Networks

This article presents a new approach to the Artificial Neural Networks (ANN) modeling of bacterial growth; using Neural Network models based on Product Units (PUNN) instead of on sigmoidal units (multilayer perceptron type [MLP]) of kinetic parameters (lag‐time, growth rate, and maximum population density) of Leuconostoc mesenteroides and those factors affecting their growth such as storage temperature, pH, NaCl, and NaNO₂ concentrations under anaerobic conditions. To enable the best degree of interpretability, a series of simple rules to simplify the expression of the model were set up. The new model PUNN was compared with Response Surface (RS) and MLP estimations developed previously. Standard Estimation Error of generalization (SEP_G’) values obtained by PUNN were lower for lag‐time and growth rate but higher for maximum population density than MLP when validated against a new data set. In all cases, bias factors (B_f) and accuracy factors (A_f) were close to unity, which indicates a good fit between the observations and predictions for the 3 models. In our study, PUNN and MLP models were more complex than the RS models, especially in the case of the growth rate parameter, but they described lower SEP_G’. With this work we have attempted to propose a new approach to neural networks estimations for its application on predictive microbiology, searching for models with easier interpretation and with a great ability to fit the data on the boundaries of variables range. We consider that still there is a lot left to do but PUNN could be a very valuable instrument for mathematical modeling.     

A new mechanistic growth model for simultaneous determination of lag phase duration and exponential growth rate and a new Bĕlehdrádek-type model for evaluating the effect of temperature on growth rate

A new mechanistic growth model was developed to describe microbial growth under isothermal conditions. The new mathematical model was derived from the basic observation of bacterial growth that may include lag, exponential, and stationary phases. With this model, the lag phase duration and exponential growth rate of a growth curve were simultaneously determined by nonlinear regression. The new model was validated using Listeria monocytogenes and Escherichia coli O157:H7 in broth or meat. Statistical results suggested that both bias factor (B_f) and accuracy factor (A_f) of the new model were very close to 1.0. A new B?lehdrádek-type rate model and the Ratkowsky square-root model were used to describe the temperature dependence of bacterial growth rate. It was observed that the maximum and minimum temperatures were more accurately estimated by a new B?lehdrádek-type rate model. Further, the inverse of square-roots of lag phases was found proportional to temperature, making it possible to estimate the lag phase duration from the growth temperature.     

A model of the effect of temperature on the growth of pathogenic and nonpathogenic Vibrio parahaemolyticus isolated from oysters in Korea

Vibrio parahaemolyticus is recognized as the leading cause of human gastroenteritis associated with the consumption of seafood. The objective of this study was to model the growth kinetics of pathogenic and nonpathogenic V. parahaemolyticus in broth and oyster slurry. Primary growth models of V. parahaemolyticus in broth and oyster slurry fit well to a modified Gomperz equation (broth R(2)=0.99; oyster slurry R(2)=0.96). The lag time (LT), specific growth rate (SGR), and maximum population density (MPD) of each primary model were compared. The growth of nonpathogenic V. parahaemolyticus was found to be more rapid than that of pathogenic V. parahaemolyticus, regardless of the model medium. In addition, significant (P<0.05) differences in the growth kinetics between pathogenic and nonpathogenic V. parahaemolyticus in broth were observed at 10 degrees C. When compared to growth in broth, the growth of V. parahaemolyticus was delayed in oyster slurry, and growth was not observed at 10 or 15 degrees C. The Davey and square root models were identified as appropriate secondary models for predicting the LT and SGR, respectively. For the broth model, the average B(f) and A(f) values for LT were found to be 0.97 and 1.3, respectively, whereas the average B(f) and A(f) values for SGR were 1.05 and 1.11, respectively. The model generated in this study predicted an LT that was shorter and an SGR that was similar to those that were actually observed, which indicates that these models provide a reliable and safe prediction of V. parahaemolyticus growth.     

Comparison between a sire model and an animal model for genetic evaluation of fertility traits in Danish Holstein population

Comparisons between a sire model, a sire-dam model, and an animal model were carried out to evaluate the ability of the models to predict breeding values of fertility traits, based on data including 471,742 records from the first lactation of Danish Holstein cows, covering insemination years from 1995 to 2004. The traits in the analysis were days from calving to first insemination, calving interval, days open, days from first to last insemination, number of inseminations per conception, and nonreturn rate within 56 d after first service. The correlations between sire estimated breeding value (EBV) from the animal model and the sire-dam model were close to 1 for all the traits, and those between the animal model and the sire model ranged from 0.95 to 0.97. Model ability to predict sire breeding value was assessed using 4 criteria: 1) the correlation between sire EBV from 2 data subsets (DATA_A and DATA_B); 2) the correlation between sire EBV from training data (DATA_A or DATA_B) and yield deviation from test data (DATA_B or DATA_A) in a cross-validation procedure; 3) the correlation between the EBV of proven bulls, obtained from the whole data set (DATA_T) and from a reduced set of data (DATA_C1) that contained only the first-crop daughters of sires; and 4) the reliability of sire EBV, calculated from the prediction error variance of EBV. All criteria used showed that the animal model was superior to the sire model for all the traits. The sire-dam model performed as well as the animal model and had a slightly smaller computational demand. Averaged over the 6 traits, the correlations between sire EBV from DATA_A and DATA_B were 0.61 (sire model) versus 0.64 (animal model), the correlations between EBV from DATA_T and DATA_C1 for proven bulls were 0.59 versus 0.67, the correlations between EBV and yield deviation in the cross-validation were 0.21 versus 0.24, and the reliabilities of sire EBV were 0.42 versus 0.46. Model ability to predict cow breeding value was measured by the reliability of cow EBV, which increased from 0.21 using the sire model to 0.27 using the animal model. All the results suggest that the animal model, rather than the sire model, should be used for genetic evaluation of fertility traits.     

Modeling Penicillium Expansum Growth Response to Thyme Essential oil at Selected Water Activities and pH Values Using Surface Response Methodology

Our objective was to evaluate, using a full factorial design, the effects of selected water activities (0.990, 0.945, or 0.900), pHs (5, 4, or 3), and thyme essential concentration (TEO, 0, 25, 50, or 100ppm) oil on Penicillium expansum lag time (λ) and radial growth rate (μ_m) obtained by modeling mold response using Gompertz equation, and corresponding polynomial quadratic models. Potato-dextrose agar formulated with every studied factor combination was inoculated with 10³ spores/ml, and incubated at 25°C up to 30 days. Mold colony diameter was periodically measured during incubation and adjusted with Gompertz equation to determine λ and μ_m. Decreasing a_w and pH, and increasing TEO concentration decreased μ_m and increased λ. At low a_w and pH, the increase in TEO concentration had a dramatic effect on P. expansum response since 25ppm of TEO inhibited its growth for 30 days at 25°C. Gompertz parameters exhibited that P. expansum was sensitive to the evaluated combined factors, allowing us to construct a secondary predictive growth model. TEO in combination with a_w and pH reduction effectively inhibited P. expansum growth.     

Growth kinetics and predictive model of Aeromonas hydrophila in a broth-based system

This study developed a predictive model of Aermonas hydrophila in tryptic soy broth for any combination of temperatures (5 to 40°C), pH (6 to 8), and NaCl (0 to 5%) using a response surface model. A. hydrophila tended to grow within a pH range of 6.0 to 8.0 and could not tolerate NaCl up to 5.0%. The interaction of pH and NaCl did not affect the specific growth rates (SGR). The primary model to obtain the SGR showed a good fit (R²≥0.980). A secondary model was obtained by non-linear regression analysis and calculated as: SGR= 0.4577+0.0529X₁−0.1641X₂−0.1493X₃−0.0016X₁X₂−0.0001X₁X₃+0.0115X₂X₃0.0006X₁ ²+0.0114X₂ ²+0.0150X₃ ² (X₁=temperature, X₂=pH, X₃=NaCl). The appropriateness of the polynomial model was verified by the mean square error (0.0023), bias factor (0.922), accuracy factor (1.343), and coefficient of determination (0.937). The newly secondary model of SGR for A. hydrophila could be incorporated into the tertiary model to predict the growth of A. hydrophila.     

Estimating the maximum growth rate from microbial growth curves: definition is everything

The maximum growth rate (μ_max) is an important parameter in modelling microbial growth under batch conditions. However, there are two definitions of this growth parameter in current use and some of the comparisons of data made in the literature fail to acknowledge this important fact.We compared values of μ_max obtained by applying the Gompertz, logistic and Baranyi–Roberts models to experimental data on the growth of Listeria monocytogenes and Listeria innocua using both absorbance and viable counts measurements of cell concentration. All three models fitted the experimental data well, however, the values of μ_max obtained using the Gompertz and logistic models were similar to each other but substantially different from those predicted by the Baranyi–Roberts model. The latter growth model was used to derive a second estimate of μ_max based on the slope at the inflection point of the growth curve function; this value was in closer agreement with those obtained using the Gompertz or logistic models. Conditions were identified when values of μ_max based on different definitions would converge towards one another.     

Modelling the effect of temperature and water activity on the growth of Aspergillus parasiticus on irradiated Argentinian flint maize

A full factorially designed experiment including storage temperature (10, 20, 30 and 37 °C) and water activity (0.88, 0.92 and 0.96) was undertaken to study the growth of Aspergillus parasiticus in maize samples. Kinetic parameters such as specific growth rate (μ), lag phase duration and maximum logarithmic increase were determined by fitting the Modified Gompertz equation to the viable mould count data (N in CFU/g) as a function of time collected in twelve experiments. The average coefficient of determination (R²) was 0.987, being the mean standard deviation of the estimate of 0.216 in units of log₁₀N. In the practical range of 10-30 °C, the relationship of the three kinetic parameters with temperature was described by second order polynomial expressions, whose parameters, in turn, depended on water activity. The combined or full model i.e., the Modified Gompertz model with its parameters expressed as a function of temperature and water activity, was able to predict log₁₀N with an average percentage error of 4.3, so agreement with the experimental data was highly satisfactory.In a simulation exercise, the full model was able to predict the viable mould count, given an initial value and grain temperature and water activity histories, with promising results for maize storage.