Work

Machine learning for detoxification of aflatoxin M1 by Lactococcus lactis probiotic in kashk production

Year: 2025

Type: article

Abstract: This study investigates the potential of Lactococcus lactis in reducing the concentration of aflatoxin M1 (AFM1) in Iranian dairy products (kashk). First, different samples were prepared and then inoculated with a certain dose of L. lactis , and then AFM1 levels were evaluated during storage periods. The maximum activity of L. lactis in the removal of AFM1 toxin was observed at the level of 10 9 CFU/mL after 30 days of exposure and the lowest level of AFM1 removal was observed at the level of 10 7 CFU/mL after 5 days of exposure. The results showed that L. lactis , especially at higher concentrations, had a more significant potential in reducing the level of AFM1 in the samples. A model of ANN (Artificial Neural Network) and PCA (principal component analysis) are employed to efficiently address the intricate challenge of detoxifying AFM1 with the aid of L. lactis . The optimized neural network architecture (3-10-1) featuring 10 neurons demonstrates impressive correlation coefficient values (0.98). The findings offer valuable insights that highlight the ANN model’s ability to make accurate predictions. Therefore, the detoxification of AFM 1 using probiotics combined with machine learning methods presents a practical, feasible, and simple method for predicting detoxification processes based on various parameters related to the probiotic application in managing aflatoxin in dairy products.

Source: Food Safety and Risk

DOI: 10.1186/s40550-025-00114-0

Authors: Maryam Jafari, Roshanak Rafiei Nazari, Mohammad Rezaei, Mojtaba Moazzen, Nabi Shariatifar