Assessing the prediction ability of different mathematical models for the growth of Lactobacillus plantarum under non-isothermal conditions.

Mathematical models taking temperature variations into account are useful in predicting microbial growth in foods, like meat products, for which Lactobacillus plantarum is a mesophilic and one of the main spoiling bacterium. The current study assessed the ability of the main primary models and their non-isothermal versions to predict L. plantarum growth under constant and variable temperature. Experimental data of microbial growth were obtained in MRS medium under isothermal conditions (4, 8, 12, 16, 20, and 30°C) which were used to obtain the secondary models. The experimental data under non-isothermal conditions (periodically oscillating temperature between the plateaus 4-12, 5-15, and 20-30°C) were used to validate the non-isothermal models. The bias factors indicated that all assessed models provided safe predictions of the microorganism growth at the non-isothermal conditions. Overall, despite the very good performance of the primary models (isothermal), none of the models was able to predict with accuracy the L. plantarum growth under temperature variations, mainly when the temperature range was close to refrigeration temperature. Incorporating the complex microbial adaptation mechanisms into the predictive models is a challenge to be overcome.