Assessing the growth of Listeria monocytogenes in salmon with or without the competition of background microflora — A one-step kinetic analysis

Abstract This study was aimed at investigating the growth of Listeria monocytogenes (L. monocytogenes) in irradiated and raw salmon. Mathematical models were developed to describe the growth of L. monocytogenes with or without the presence of the native microbiota in salmon. At temperatures between 4 and 35 °C, the growth data of L. monocytogenes in sterile salmon were collected to develop the non-competitive growth model, while the observed growth data of both L. monocytogenes and background microflora in raw salmon were analyzed simultaneously to develop the competitive model. The Huang square-root model was used to evaluate the effect of temperature on growth rates and lag times of L. monocytogenes and the native microbiota. A one-step analysis approach was used for model development and the kinetic parameters estimation. The results showed that the background microflora in raw salmon have inhibiting effect on the growth of L. monocytogenes. The lag time of L. monocytogenes in raw salmon was 2.849 times longer than that in sterile salmon, while its specific growth rate was decreased by 78.7% with the presence of the native microbiota. The minimum growth temperature for L. monocytogenes ( T m i n , L m ) in raw salmon estimated by the competitive model was 1.3 °C, which was close to T m i n , L m in sterile salmon (1.2 °C) determined by the non-competitive model. With the relatively low value of RMSE (0.3 Log CFU/g), the non-competitive and competitive models were accurate to describe the growth of L. monocytogenes in sterile salmon and the interaction between L. monocytogenes and the native microflora in raw salmon, respectively. The predictive models and the associated kinetic parameters were validated under isothermal temperature conditions. The validation results indicated that the non-competitive and competitive models were accurate. Overall, the residual errors of predictions of these two models both followed a Laplace distribution. The RMSE values of the models were both 0.4 Log CFU/g and more than 79% of prediction errors were within normal experimental errors (±0.5 Log CFU/g). The results from this study may be useful for the future risk assessments of L. monocytogenes and microbiological shelf-life prediction of salmon.