Background: In Jordan as in other worldwide countries, mycotoxins are considered a serious national problem in food supplies. As a result, almost all nations are setting and adopting different regulations targeting the control of mycotoxins levels in the domestic food supply, including the problem of reliable sampling and analysis methods. Objective: It is necessary to improve and give evidence of analytical abilities of laboratories within Jordan and developing countries enabling them to monitor mycotoxins effectively in food to overcome non-tariff obstacles. Methods: We analyzed 40 samples from wheat, corn, dried fig and dried coffee beans for total aflatoxin content using High Pressure Liquid Chromatography (HPLC) and Enzyme Linked Immunesorbent Assay (ELISA) methods. Results: 40% of samples from wheat, 60% from corn, 30% from dried fig, and 50% from dried coffee beans were found positive when speaking of total aflatoxins, with average values between 1.14 and 4.12 μg/kg. Obtained results allow considering all tested food samples as fit for human consumption if compared with the labeled regulatory limit of allowed aflatoxins in the European Union. In detail, the limit of detection and the limit of quantification for methods used in this study were significantly lower than the maximum limits established by the European Union. Highlights: The procedure used in this study is suitable for detection of mycotoxins at very low concentration.

The incidence of contamination of aflatoxin M1 (AFM1) in milk samples collected from the Jordanian market was investigated by using the competitive enzyme linked immunosorbent assay (ELISA) technique. A total of 175 samples were collected during 2014-2015 year. All tested samples were contaminated with various levels of AFM1 ranging from 9.71 to 288.68 ng/kg. The concentration of AFM1 in 66% of fresh milk samples was higher than the maximum tolerance limit accepted by European Union (50 ng/kg) and 23% higher than the maximum tolerance limit accepted by USA (500 ng/kg). Percentages of contaminated raw cow, sheep, goat and camel milk exceeding the European tolerance limit were 60, 85, 75 and 0%, respectively. Twelve % of AFM1 contaminated pasteurized cow milk samples exceeded the European tolerance limit with a range of contamination between 14.60 and 216.78 ng/kg. For infant formula samples, the average concentration of AFM1 was 120.26 ng/kg (range between 16.55 to 288.68 ng/kg), the concentration of AFM1 in 85% of infant formula samples were higher than the maximum tolerance limit accepted by European Union and USA (25 ng/kg).

This study was undertaken to determine the presence of aflatoxin M1 (AFM1) in animal milk. In addition, exposure of infants to aflatoxin M1 (AFM1) and lactating mothers to aflatoxin B1 (AFB1) was examined using AFM1 in breast milk as a biomarker for exposure to AFB1. In total, 100 samples of fresh animal milk and fermented milk (buttermilk) and 80 samples of human breast milk were collected during the period 2011–2012. An enzyme-linked immunosorbent assay (ELISA) was used for the analysis of milk samples. AFM1 was detected in all animal fresh and fermented milk samples. The concentrations of AFM1 in 70 samples of fresh and fermented milk were higher than the maximum tolerance limit accepted by the European Union and the United States of 50 ng/kg. In human milk samples the average concentration of AFM1 was higher than the maximum tolerance limit accepted by the European Union and the United States of 25 ng/kg. Logistic regression analysis failed to show a correlation between AFM1 and type and amount of dairy consumption, vegetables, fruits, and meat. However, an association between AFM1 and cereal consumption was detected. This study is the first to report on the occurrence of AFM1 in milk consumed by the Jordanian population.

The presence of aflatoxin M1 in breast milk poses a serious risk to the health of infants because of its potential to cause cancer and have negative effects on development. Detecting AFM1 in milk samples using conventional methods is often time-consuming and may not provide real-time monitoring capabilities. The use of machine learning techniques to forecast aflatoxin M1 levels in breast milk samples is examined in this study. To develop predictive models of aflatoxin M1 in breast milk, we employed well-known supervised machine learning algorithms such as Random Forest and Gradient Boosting. The findings show that machine learning can be used for the identification of aflatoxin M1 in breast milk. By actively monitoring breast quality, this research highlights the significance of machine learning in protecting babies’ health and advances the prediction skills in food safety.