Predictive modeling of survival/death of Listeria monocytogenes in liquid media: Bacterial responses to cinnamon essential oil, ZnO nanoparticles, and strain

To predict Listeria monocytogenes population during storage (8 degrees C) as a function of time (1-16 days), cinnamon essential oil (EO), ZnO nanoparticles (NPs; 10-30 nm), and two different genotypes in liquid microbiological medium, an adaptive neuro fuzzy inference system (ANFIS) was developed. For this purpose, 32 modeling scenarios were investigated. The ANFIS scenarios were fed with 4 inputs of EO concentration (0, 0.8, 1.6, and 2.4%), ZnO NPs (0, 5, 10, and 15 mg/ml), strain (2 strains), and storage time (1-16 days). Our findings demonstrate that the final ANFIS architecture with triangular-shaped membership function (MF) provides the best prediction accuracy (RMSE = 0.214; R-2 = 0.974) over models with other MFs. Moreover, the effects of antibacterial activity of cinnamon EO were investigated in a food model system, vegetable broth. The bacterial counts decreased with increasing cinnamon oil and ZnO NPs concentrations; however, some strain variation was observed. These observations demonstrate the reliability of the ANFIS model for prediction of L. monocytogenes population and confirm its potential use as a supplemental tool in predictive microbiology. (C) 2016 Elsevier Ltd. All rights reserved.