Comparative Analysis of Machine Learning Methods to Predict Growth of F. sporotrichioides and Production of T-2 and HT-2 Toxins in Treatments with Ethylene-Vinyl Alcohol Films Containing Pure Components of Essential Oils

Machine Learning Applied to the Detection of Mycotoxin in Food: A Systematic Review

Mycotoxins, toxic secondary metabolites produced by certain fungi, pose significant threats to global food safety and public health. These compounds can contaminate a variety of crops, leading to economic losses and health risks to both humans and animals. Traditional lab analysis methods for mycotoxin detection can be time-consuming and may not always be suitable for large-scale screenings. However, in recent years, machine learning (ML) methods have gained popularity for use in the detection of mycotoxins and in the food safety industry in general due to their accurate and timely predictions. We provide a systematic review on some of the recent ML applications for detecting/predicting the presence of mycotoxin on a variety of food ingredients, highlighting their advantages, challenges, and potential for future advancements. We address the need for reproducibility and transparency in ML research through open access to data and code. An observation from our findings is the frequent lack of detailed reporting on hyperparameters in many studies and a lack of open source code, which raises concerns about the reproducibility and optimisation of the ML models used. The findings reveal that while the majority of studies predominantly utilised neural networks for mycotoxin detection, there was a notable diversity in the types of neural network architectures employed, with convolutional neural networks being the most popular.

Exploring the impact of lactic acid bacteria on the biocontrol of toxigenic Fusarium spp. and their main mycotoxins

The occurrence of fungi and mycotoxins in foods is a serious global problem. Most of the regulated mycotoxins in food are produced by Fusarium spp. This work aimed to assess the antifungal activity of selected lactic acid bacteria (LAB) strains against the main toxigenic Fusarium spp. isolated from cereals. Various machine learning (ML) algorithms such as neural networks (NN), random forest (RF), extreme gradient boosted trees (XGBoost), and multiple linear regression (MLR), were applied to develop models able to predict the percentage of fungal growth inhibition caused by the LAB strains tested. In addition, the ability of the assayed LAB strains to reduce/inhibit the production of the main mycotoxins associated with these fungi was studied by UPLC-MS/MS. All assays were performed at 20, 25, and 30 °C in dual culture (LAB plus fungus) on MRS agar-cereal-based media. All factors and their interactions very significantly influenced the percentage of growth inhibition compared to controls. The efficacy of LAB strains was higher at 20 °C followed by 30 °C and 25 °C. Overall, the order of susceptibility of the fungi to LAB was F. oxysporum > F. poae = F. culmorum ≥ F. sporotrichioides > F. langsethiae > F. graminearum > F. subglutinans > F. verticillioides. In general, the most effective LAB was Leuconostoc mesenteroides ssp. mesenteroides (T3Y6b), and the least effective were Latilactobacillus sakei ssp. carnosus (T3MM1 and T3Y2). XGBoost and RF were the algorithms that produced the most accurate predicting models of fungal growth inhibition. Mycotoxin levels were usually lower when fungal growth decreased. In the cultures of F. langsethiae treated with LAB, T-2 and HT-2 toxins were not detected except in the treatments with Pediococcus pentosaceus (M9MM5b, S11sMM1, and S1M4). These three strains of P. pentosaceus, L. mesenteroides ssp. mesenteroides (T3Y6b) and L. mesenteroides ssp. dextranicum (T2MM3) inhibited fumonisin production in cultures of F. proliferatum and F. verticillioides. In F. culmorum cultures, zearalenone production was inhibited by all LAB strains, except L. sakei ssp. carnosus (T3MM1) and Companilactobacillus farciminis (T3Y6c), whereas deoxynivalenol and 3-acetyldeoxynivalenol were only detected in cultures of L. sakei ssp. carnosus (T3MM1). The results show that an appropriate selection and use of LAB strains can be one of the most impacting tools in the control of toxigenic Fusarium spp. and their mycotoxins in food and therefore one of the most promising strategies in terms of efficiency, positive impact on the environment, food safety, food security, and international economy.

Lactic Acid Bacteria as Potential Agents for Biocontrol of Aflatoxigenic and Ochratoxigenic Fungi

Aflatoxins (AF) and ochratoxin A (OTA) are fungal metabolites that have carcinogenic, teratogenic, embryotoxic, genotoxic, neurotoxic, and immunosuppressive effects in humans and animals. The increased consumption of plant-based foods and environmental conditions associated with climate change have intensified the risk of mycotoxin intoxication. This study aimed to investigate the abilities of eleven selected LAB strains to reduce/inhibit the growth of Aspergillus flavus, Aspergillus parasiticus, Aspergillus carbonarius, Aspergillus niger, Aspergillus welwitschiae, Aspergillus steynii, Aspergillus westerdijkiae, and Penicillium verrucosum and AF and OTA production under different temperature regiments. Data were treated by ANOVA, and machine learning (ML) models able to predict the growth inhibition percentage were built, and their performance was compared. All factors LAB strain, fungal species, and temperature significantly affected fungal growth and mycotoxin production. The fungal growth inhibition range was 0-100%. Overall, the most sensitive fungi to LAB treatments were P. verrucosum and A. steynii, while the least sensitive were A. niger and A. welwitschiae. The LAB strains with the highest antifungal activity were Pediococcus pentosaceus (strains S11sMM and M9MM5b). The reduction range for AF was 19.0% (aflatoxin B1)-60.8% (aflatoxin B2) and for OTA, 7.3-100%, depending on the bacterial and fungal strains and temperatures. The LAB strains with the highest anti-AF activity were the three strains of P. pentosaceus and Leuconostoc mesenteroides ssp. dextranicum (T2MM3), and those with the highest anti-OTA activity were Leuconostoc paracasei ssp. paracasei (3T3R1) and L. mesenteroides ssp. dextranicum (T2MM3). The best ML methods in predicting fungal growth inhibition were multilayer perceptron neural networks, followed by random forest. Due to anti-fungal and anti-mycotoxin capacity, the LABs strains used in this study could be good candidates as biocontrol agents against aflatoxigenic and ochratoxigenic fungi and AFL and OTA accumulation.

Predicting the Growth of F. proliferatum and F. culmorum and the Growth of Mycotoxin Using Machine Learning Approach

In distinct parts of the food web, Fusarium culmorum and Fusarium preserving the relationship can germinate and grow zearalenone (ZEA) and fumonisins (FUM), accordingly. Antimicrobial drugs used to combat these fungi and toxic metabolites raise the risk of hazardous residue in food products, as well as the development of fungus tolerance. For modeling fungal growth and pathogenicity under separate water action (a q ) (0.96 and 0.99) and surface temp (20 and 28°C) tyrannies, several machine learning (ML) methodologies (artificial neural, regression trees, and extreme rise enhanced trees) and multiple regression model (MLR) were used also especially in comparison. GR and mycotoxin levels inside the environment often reduced as EOC concentrations grew, although some treatment in association with specific a q and temperature values caused ZEA production. In terms of predicting the growth rate of F. culmorum and F. maintaining the relationship and the production of ZEA and FUM, random forest techniques outperformed neural network models and extreme gradient boosted trees. The MLR option was the most inefficient. It is the first research to look at the ML potential of bio EVOH products containing EOCs and ambient variables of F. culmorum and F. proliferatum development, as well as the generation of zearalenone and fumonisins. The findings show that these entire novel wrapping technologies, in tandem using machine learning techniques, could be useful in predicting and controlling the dangers connected with fungal species or biotoxins in foodstuff.