One-step kinetic analysis of competitive growth of Salmonella spp. and background flora in ground chicken

Abstract The present study was aimed at developing a mathematical model to predict the growth of Salmonella in ground chicken in the presence of the background microflora. Ground chicken was inoculated with a cocktail of S. Typhimurium and S. Enteritidis and incubated at various isothermal temperatures (8-33 °C). Salmonella grew at all temperatures, except at 8 °C, while the background microflora could grow under all conditions. The observed growth data of both Salmonella and background microflora at temperatures between 12 and 33 °C were analyzed simultaneously to develop the predictive growth models. The Huang-Jameson effect (HJE) model and Huang-Lotka-Volterra (HLV) model were used to describe the growth and interaction between Salmonella and background microflora. The Huang square-root model was used to evaluate the effect of temperature on the growth rates and lag times of Salmonella and background microflora. A one-step analysis method was used to directly build the tertiary models and to determine the kinetic parameters from the growth curves. The minimum growth temperature (Tmin) for Salmonella estimated by both the HJE and HLV models was 7.2 °C. The Tmin for background flora determined by HJE and HLV model was 1.3 and 1.8 °C, respectively. Under competition, the growth rate of Salmonella could be lower or higher than that of the native microbiota, depending on the storage temperature. Above 16.8 °C, Salmonella grew faster than the background microflora. With relatively low value of RMSE (0.3 log CFU/g), the HJE and HLV models could both successfully describe the growth of Salmonella and native microflora and the interaction between the two. Although the Akaike Information Criterion (AIC) value of the HJE model (−349.1) was slightly smaller than that of the HLV model (−339.4), both models were practically equal in accuracy for predicting the competitive growth of Salmonella and background flora in ground chicken. The HJE model and kinetic parameters were validated using separate isothermal and dynamic growth data. The validation results indicated the competition model was accurate. The RMSE of the predictions was only 0.3 log CFU/g. Overall, the residual errors of predictions followed a normal distribution, with more than 86% of them were within ±0.5 log CFU/g. The results from this study may be useful for microbial risk assessments of Salmonella and shelf-life prediction of ground chicken.