Objective To develop a method for the simultaneous determination of 18 kinds of polychlorinated biphenyls (PCBs) in pork by isotope dilution-gas chromatography-high resolution magnetic mass spectrometry. Methods The optimal conditions of the accelerated solvent extractor and fully automated clean-up apparatus were determined by optimizing the extraction rate of the target substances using an accelerated solvent extractor and a fully automated clean-up instrument combined with isotope dilution method coupled with gas chromatography-high resolution magnetic mass spectrometry. A 4-factor, 3-level orthogonal test was designed using the extraction temperature, static extraction time and the number of cycles as the main factors to investigate the effect of the accelerated solvent extractor on the extraction rate of the fat content of the samples. The effects of the accelerated solvent extractor on the extraction rate of fat content of the samples were investigated. Results The linear relationships of the indicator PCBs (PCB28, PCBB52, PCB101, PCB118, PCB138, PCB153, PCB180) was good in the range of 0.5 to 200.0 pg/µL, and the linear range of the DL-PCBs (PCB77, PCB81, PCB105, PCB114, PCB123, PCB126, PCB156, PCB157, PCB167, PCB169, PCB189) had a good linear relationship in the range of 0.10 to 40.0 pg/µL, the correlation coefficients were all above 0.999. The limits of detection were in the range of 0.001-0.004 pg/g, and the limits of quantitation were in the range of 0.004-0.012 pg/g; the accuracy and precision of the fishmeal matrix standard reference material were tested by this study, and the PCBs levels were within the range with the relative standard deviations of 1.0%-7.5%. For the determination of actual samples of pork, and the contamination levels of PCBs in 18 kinds of were 1978.9-2530.8 pg/g fat. Conclusion This study is efficient, accurate, sensitive and suitable for the determination of 18 kinds of PCBs in pork.

Objective To establish an analytical method for the simultaneous determination of 8 kinds of organophosphate ester flame retardants of major meat and cereal food products by QuEChERS purification technology coupled with ultra performance liquid chromatography-tandem mass spectrometry. Methods Separation was achieved on a Agilent EC-C₁₈ column (3 mm×150 mm, 2.7 μm) with a mobile phase consisting of 0.1% formic acid aqueous solution and acetonitrile under gradient elution at a flow rate of 0.4 mL/min. Detection was performed in positive electrospray ionization (ESI+) mode with multiple reaction monitoring (MRM). Results The 8 kinds of organophosphate ester flame retardants exhibited good separation and showed excellent linearity in the range of 1.0-50.0 ng/mL, with correlation coefficients all greater than 0.99. The limits of detection and limits of quantitation were 1.0 μg/kg and 2.0 μg/kg, respectively. The recoveries of this method were 70.0%-119.6%, and the relative standard deviations were 0.2%-12.7%. Conclusion The established method is simple, rapid and highly efficient in extraction, and has been successfully applied to the analysis of organophosphate ester flame retardants in food samples (including cereals and meat products).

Objective To investigate the current situation of cadmium contamination in food and explore the dietary exposure risk of cadmium in the daily diet of residents in the southeast coastal areas. Methods Random sampling was conducted in major large-scale agricultural markets, commercial supermarkets, restaurants and other places throughout the city, with common commercially available foods as the research object. A total of 1285 food samples were collected from 7 categories, including grains and their products, vegetables, aquatic animals and their products, meat products, nuts and seeds, fresh edible fungi and seasonings. Inductively coupled plasma mass spectrometry method (ICP-MS) was applied for quantitative detection of cadmium content. The single factor pollution index (P_i) and target hazard factor (THQ) were used to evaluate the quality and consumption safety of each category of samples. Results After statistical analysis, there were significant differences in cadmium content among different categories of food. In the tested samples, the total detection rate of cadmium was 54.24%, and the detection result was undetected-2.86 mg/kg. Among them, crustaceans (sea crabs, shrimp) had the highest average detection value (1.2264 mg/kg). Conclusion The dietary exposure risk assessment results of tested sample indicate THQ<1. Therefore, the cadmium contamination in this region shows a low health risk to residents.

Objective To explore the process of caffeine removal from green tea by supercritical carbon dioxide extraction. Methods Using Wuyuan green tea as raw material, supercritical carbon dioxide extraction was used to remove caffeine from green tea. Taking the caffeine removal rate as the investigation index, the effects of tea moisture content, extraction temperature, pressure, time, concentration and dosage of entrained agent on caffeine removal rate were studied by single factor experiment, and the caffeine removal process was optimized by orthogonal experiment. Results The optimal extraction process of caffeine removal in green tea was tea moisture content 40%, extraction pressure 30 MPa, extraction temperature 55 ℃, extraction time 240 min, concentration of entrainment agent (ethanol) 60%, and dosage of entrainment agent 1.2:1 (m:m). Under these conditions, the caffeine removal rate was 91.2%. Extraction pressure was the most important factor affecting caffeine removal rate, followed by extraction temperature, and extraction time had relatively little effect. Conclusion This study determines the best technology of caffeine removal in green tea by supercritical carbon dioxide extraction, which provides a certain reference for caffeine removal in green tea.

Objective To establish a method for the determination of 17α-estradiol and 17β-estradiol in aquatic products by QuEChERS combined with ultra performance liquid chromatography-tandem mass spectrometry (UPLC-MS/MS). Methods The sample was extracted with water-acetonitrile (2:8, V:V), purified with N-propylethylenediamine, C₁₈ and neutral alumina, blown to dryness with nitrogen, and then reconstituted with water-acetonitrile (8:2, V:V). The sample solution was analyzed using water and acetonitrile (both containing 5 mmol/L ammonia) as mobile phase, separated by C₁₈ chromatographic column, scanned and detected in multiple reaction monitoring (MRM) mode, and quantified by external standard method. Results The results showed that 17α-estradiol and 17β-estradiol showed a good linear relationship in the concentration ranges of 5 to 200 μg/L, and the correlation coefficients were 0.9998 and 0.9985, respectively. The average recoveries at three spiked concentration levels varied from 66.5% to 101.1% with relative standard deviations of 0.62% to 10.17%. The limit of detection was 0.85-0.90 μg/kg, and the limit of quantitation was 2.1-2.4 μg/kg. Conclusion This method is simple, fast, sensitive and accurate, suitable for qualitative and quantitative analysis of 17α-estradiol and 17β-estradiol in aquatic products such as fish, shrimp, crabs and shellfish.

Objective To optimize coffee peel beverage formula by response surface methodology with coffee peel and black tea as the main ingredients. Methods First, the ripe coffee fresh fruits were peeled. After homogenization and filtration, coffee peel liquid was obtained. This liquid was then mixed with black tea liquid, fructose and citric acid for formulation. Single-factor experiments and the response surface methodology were employed to investigate the effects of the addition amounts of coffee peel juice, black tea liquid, fructose, and citric acid on the coffee peel beverage, thereby determining the optimal formulation for it. Results The optimal addition amounts were 28% coffee peel juice, 40% black tea liquid, 5% fructose and 0.05% citric acid. This formulation yielded a product with protein content (5.13±0.34)%, fat content (4.94±0.78)%, dietary fiber (18.43±1.06)%, ash (1.00±0.01)%. The coffee peel juice beverage processed according to this process had a strong aroma, a deep amber color, and a delicious sour and sweet taste. Conclusion The application of response surface methodology to optimize the formula of coffee peel beverage provides another feasible way for the high value utilization of coffee peel and the development of functional products.

Objective To investigate the contamination of fusarium and alternaria toxins in wheat and wheat flour in Shaanxi Province. Methods A total of 140 samples of wheat and wheat flour were randomly collected in circulation of 9 cities from 2022 to 2024. Deoxynivalenol (DON), 3-acetyldeoxynivalenol (3-AcDON), 15-acetyldeoxynivalenol (15-AcDON), nivalenol (NIV), zearalenone (ZEN), tenuazonic acid (TeA), alternariol (AOH), alternariol monomethyl ether (AME) and tentoxin (TEN) were determined by high performance liquid chromatography-tandem mass spectrometry. Results The fusarium and alternaria toxins were detected in wheat and wheat flour. The detection rates of DON, 3-AcDON, 15-AcDON, NIV, ZEN, TeA, TEN, AME in wheat were 83.3% (50/60), 5.00% (3/60), 10.0% (6/60), 11.6% (7/60), 8.33% (5/60), 100% (60/60), 100% (60/60) and 98.3% (59/60), respectively. The detection rates of DON, TeA, TEN, AME in wheat flour were 87.5% (70/80), 96.2% (77/80), 88.7% (71/80) and 63.7% (51/80), respectively. A strong positive correlation between DON and 15-AcDON content was demonstrated in the positive samples, so was between TeA and TEN content. The DON dietary exposure was 959.8 ng/(kg·bw·d), which was less than the corresponding tolerable daily intake value. The TeA, TEN, AME dietary exposures were 159.2, 18.1 and 2.30 ng/(kg·bw·d), which were below the corresponding threshold of toxicological concern values in wheat flour. Conclusion Wheat and wheat flour in Shaanxi Province are generally contaminated with fusarium and alternaria toxins from 2022 to 2024, with the highest detection rates of TeA and TEN, following by DON. The fusarium and alternaria toxins exposure risk of wheat flour are within an acceptable risk range.

Objective To optimize the ultrasonic extraction process of dihydromyricetin from Ampelopsis grossedentata using artificial neural networks combined with response surface methodology. Methods The stems and leaves of Ampelopsis grossedentata were used as the research material. An ultrasonic extraction system for dihydromyricetin was established, and the process parameters were systematically optimized using a combination of single-factor experiments, response surface methodology and artificial neural network models optimized by genetic algorithms. The extraction yields of dihydromyricetin from different parts of Ampelopsis grossedentata were then analyzed under optimal conditions. Results The artificial neural network model exhibited superior accuracy and predictive capability in comparison to the response surface methodology. The optimal extraction conditions were determined to be an ultrasonic power of 360 W, a temperature of 42 ℃, a liquid-to-solid ratio of 20:1 (mL:g), and an extraction time of 35 min. Under these conditions, the actual extraction yield of dihydromyricetin was (39.83±0.01)%, with a relative error of only 0.36% compared to the artificial neural network-predicted value of 40.19%. Furthermore, the extraction yield of dihydromyricetin from various parts of Ampelopsis grossedentata under optimal ultrasonic conditions followed the sequence: Stems and leaves of Ampelopsis grossedentata>branches of Ampelopsis grossedentata>pruned branches of Ampelopsis grossedentata. Conclusion This study successfully optimizes the ultrasonic extraction process to enhance the extraction efficiency of dihydromyricetin from Ampelopsis grossedentata and reveals significant differences in dihydromyricetin extraction yields among different parts of the plant.

Objective To explore the mineral element differences in different visceral by-products of Simmental beef cattle in Ningxia. Methods Using post-slaughter by-products from Ningxia Simmental cattle (beef heart, beef liver, beef tripe, beef intestines) as the research subject, the content differences of 42 kinds of mineral elements in the 4 kinds of by-products were detected and analyzed. And verified through discriminant analysis. Results The results indicated that there were significant differences in the content of 38 kinds of and 41 kinds of mineral elements respectively in the 4 kinds of beef cattle by-products from Guyuan and Wuzhong regions (P>0.05). The content of Pb, Cd, Hg, As, Cr in all beef cattle by-products is below the limit standards. Beef tripe has a strong adsorption and accumulation effect on heavy metals such as Ni and Cr, rare earth elements Eu, Gd, Nd and Sm, as well as other elements included Nb, Sb and Sn; beef liver had a strong absorption and enrichment effect on heavy metals like Cd, trace elements Cu and Mo. Based on mineral elements, effective differentiation of various by-products and samples from different origins could be achieved. Selected Cs, Na, Li, Cd, Tl, Rb, Cr, U, Pt, Ta, Co, Sc, Mo, Ni, Zn, Ir, Sn and Fe as the key characteristic elements for distinguishing between cow stomach, cow liver, cow heart, and cow intestine. Eleven key elements had been identified to distinguish the origin of beef by-products from Guyuan and Wuzhong, included Cs, Cr, Rb, As, Sn, Nb, Ni, Li, Sb, Ti and Hg. Conclusion Beef by-products have their own elemental characteristics, and different by-products and samples from different origins can be effectively distinguished by the content of mineral elements. The research findings will help establish a quality database for beef by-products, providing data support for the development of new products by enterprises.

In the face of global population growth and agricultural production pressure, as well as the serious challenges of agricultural product quality and safety issues, the traditional methods of agricultural product integrity monitoring and risk prediction have become insufficient. The rapid development of machine learning technology provides new solution ideas for agricultural product integrity monitoring and risk prediction. This paper systematically summarized the applications of machine learning technology in agricultural product safety risk monitoring (including physical, chemical and biological risks), agricultural product authenticity and traceability assurance, and agricultural product risk assessment prediction based on historical data. Machine learning technology undoubtedly improves the efficiency of agricultural product integrity monitoring effectively, realizes early detection and prevention of risks, and provides new solution for constructing a safer and more reliable food supply chain provides new solutions. Although these applications show great promise, there are still challenges to artificial intelligence in the field of agricultural produce integrity monitoring and risk prediction. Based on summarizing the literature, this paper further explored the prospects and directions of this trend, and presented the importance of machine learning model interpretability and trust issues, as well as problems in data acquisition and use, to improve the application of machine learning in agricultural product integrity monitoring and risk prediction.