基于数据统计与化学模式识别分析结合的豆粕溶剂残留组分评价研究

Research on the evaluation of solvent residue components in soybean meal based on data statistics and chemical pattern recognition analysis combination

中文关键词:  豆粕  箱线图  分布拟合  聚类分析  主成分分析  化学模式识别

英文关键词:soybean meal  Box diagram  Distribution fitting  Cluster analysis  Principal component analysis  Chemical pattern recognition

中文摘要:

      目的 对市售不同用途豆粕中溶剂残留含量进行测定,分析豆粕中溶剂残留组分情况,并采用化学模式识别法分析比较不同豆粕的差异,为指导豆粕质量安全提供数据支撑。方法 参照GB5009.262-2016《食品安全国家标准 食品中溶剂残留量的测定》对32批次豆粕进行检测,并对豆粕数据进行直方图、频次图、箱线图和分布检验统计,借助主成分分析和聚类分析对不同种类豆粕进行分类并筛选出差异标志组分。结果 豆粕样品中溶剂残留为未检出~40.5 mg·kg-1,均低于GB 14932-2016《食品安全国家标准 食品加工用粕类》限量；食用低温豆粕样品溶剂残留检出率36.4 %,最高值31.5 mg·kg-1,食用高温豆粕样品溶剂残留检出率54.5 %,最高值40.5 mg·kg-1；饲用豆粕样品溶剂残留检出率90.0 %,最高值37.0 mg·kg-1。通过对豆粕峰面积进行主成分分析、聚类分析等可将32批豆粕整体分为低温豆粕、高温豆粕和饲用豆粕,同时显示2-甲基戊烷和3-甲基戊烷对豆粕分类起到主导作用。结论 豆粕的成分定量方法高效稳定,化学模式识别分析方法结果可靠,可为豆粕中溶剂残留情况及质量研究提供参考。

英文摘要:

      ABSTRACT:Objective: To determine the solvent residue content in different commercial uses of soybean meal, analyze the composition of solvent residues in soybean meal, and compare the differences among different soybean meals using chemical pattern recognition methods to provide data support for guiding the quality and safety of soybean meal. Methods: Referring to GB5009.262-2016, 32 batches of soybean meal were tested, and the data of soybean meal were analyzed by histogram, frequency chart, box plot, and distribution test statistics. Principal component analysis and cluster analysis were used to classify different types of soybean meal and screen out differential marker components. Results: The solvent residue in soybean meal samples ranged from undetected to 40.5 mg·kg-1, all below the limit specified in GB 14932-2016; the detection rate of solvent residue in low-temperature soybean meal samples for human consumption was 36.4%, with a maximum value of 31.5 mg·kg-1, the detection rate in high-temperature soybean meal samples was 54.5%, with a maximum value of 40.5 mg·kg-1; the detection rate in feed soybean meal samples was 90.0%, with a maximum value of 37.0 mg·kg-1. Through principal component analysis and cluster analysis of soybean meal peak area, the 32 batches of soybean meal could be divided into low-temperature soybean meal, high-temperature soybean meal, and feed soybean meal, showing that 2-methylpentane and 3-methylpentane played a dominant role in the classification of soybean meal. Conclusion: The quantitative method of soybean meal components is efficient and stable, and the results of chemical pattern recognition analysis are reliable, which can provide a reference for the study of solvent residue and quality in soybean meal.