Growth Prediction and Metabolite Profiling of Probiotic Lactobacilli Cultivated with Eggshell Membrane Protein: A Comparison of Kinetic Models and Machine Learning Methods
Fermentation · 15.11.2026 · 12 (11) · 534
DOI: 10.3390/fermentation12110534 ↗Kurz gefasst
Fachartikel in der Zeitschrift Fermentation, online erschienen am 15.11.2026. Verfasst von Tulay Ozcan und 4 weiteren Autor:innen. Die Arbeit nennt 52 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Bursa Uludag University.
Abstract (Originalsprache)
This study addressed two main scientific questions: whether eggshell membrane (ESM) protein influences probiotic growth and metabolite production in a strain-dependent manner, and which predictive modeling approach provides the most accurate estimation of bacterial growth. This study aimed to determine the effects of ESM protein on the growth kinetics and metabolite profiles of Lactobacillus acidophilus, Lacticaseibacillus paracasei subsp. casei, and Lacticaseibacillus rhamnosus strains, and to compare the performance of various predictive models. Microbial growth, pH, and metabolite profiles were evaluated during 24 h of fermentation in control and ESM-containing media. Growth data were modeled using the modified Gompertz, Baranyi–Roberts, Gaussian process regression (GPR), and support vector regression (SVR) methods, with validation performed by withholding one intermediate time point at a time within each growth curve. The impact of ESM on growth kinetics was found to be strain-dependent; according to the modified Gompertz model, the maximum specific growth rate of L. casei increased from 0.53 to 0.73 h−1, whereas those of L. acidophilus and L. rhamnosus decreased from 0.94 to 0.47 h−1 and from 0.72 to 0.56 h−1, respectively. GPR yielded the lowest mean cross-validated root mean squared error (CV_RMSE = 0.1344 log10 CFU/mL), followed by SVR (0.1359), Baranyi–Roberts (0.1533), and modified Gompertz (0.1657), although the best-performing method varied among conditions. Significant changes in amino acid profiles were observed in media containing ESM, with total amino acid content increasing, particularly in the L. acidophilus group, from 344.90 to 2773.70 mg/kg after 24 h of fermentation. Lactic acid and SCFAs were also influenced by both the microorganism and the culture medium composition, with the highest final lactic acid concentration (3082.88 mg/kg) observed in the control L. rhamnosus medium. The findings indicate that ESM can be considered an alternative medium component in strain-specific fermentation processes, and that the choice of model should be based on the experimental conditions and cross-validated predictive performance.
Förderung
- Bursa Uludag University
Zitieren
APA
Ozcan, T., Yilmaz-Ersan, L., Keser, G., Narli, M. B., & Ozcan, M. O. (2026). Growth Prediction and Metabolite Profiling of Probiotic Lactobacilli Cultivated with Eggshell Membrane Protein: A Comparison of Kinetic Models and Machine Learning Methods. Fermentation, 12(11), 534. https://doi.org/10.3390/fermentation12110534
Vancouver
Ozcan T, Yilmaz-Ersan L, Keser G, Narli MB, Ozcan MO. Growth Prediction and Metabolite Profiling of Probiotic Lactobacilli Cultivated with Eggshell Membrane Protein: A Comparison of Kinetic Models and Machine Learning Methods. Fermentation. 2026;12(11):534. doi:10.3390/fermentation12110534
BibTeX
@article{Ozcan2026,
author = {Ozcan, Tulay and Yilmaz-Ersan, Lutfiye and Keser, Gokce and Narli, Merve Begum and Ozcan, Murat Olcay},
title = {Growth Prediction and Metabolite Profiling of Probiotic Lactobacilli Cultivated with Eggshell Membrane Protein: A Comparison of Kinetic Models and Machine Learning Methods},
journal = {Fermentation},
year = {2026},
volume = {12},
number = {11},
pages = {534},
doi = {10.3390/fermentation12110534}
}
EUVIMED