基于液相色谱-质谱联用技术和随机森林算法，对苹果（蔷薇科苹果属）按地理来源、品种和生产方法进行分类

Classification of Apples ( borkh.) According to Geographical Origin, Variety and Production Method Using Liquid Chromatography Mass Spectrometry and Random Forest.

作者信息

Hansen Jule, Fransson Iris, Schrieck Robbin, Kunert Christof, Seifert Stephan

机构信息

Institute of Food Chemistry, Hamburg School of Food Science, University of Hamburg, Grindelallee 117, 20146 Hamburg, Germany.

Landeslabor Schleswig-Holstein, Max-Eyth-Str. 5, 24537 Neumünster, Germany.

出版信息

Foods. 2025 Jul 29;14(15):2655. doi: 10.3390/foods14152655.

DOI:10.3390/foods14152655

PMID:40807597

原文链接:https://pmc.ncbi.nlm.nih.gov/articles/PMC12345882/

Abstract

Apples are one of the most popular fruits in Germany, valued for their regional availability and health benefits. When deciding which apple to buy, several characteristics are important to consumers, including the taxonomic variety, organic cultivation and regional production. To verify that these characteristics are correctly declared, powerful analytical methods are required. In this study, ultra-high performance liquid chromatography quadrupole time-of-flight mass spectrometry (UHPLC-Q-ToF-MS) is applied in combination with random forest to 193 apple samples for the analysis of various authentication issues. Accuracies of 93.3, 85.5, 85.6 and 90% were achieved for distinguishing between German and non-German, North and South German, organic and conventional apples and for six different taxonomic varieties. Since the classification models largely use different parts of the data, which is shown by variable selection, this method is very well suited to answer different authentication issues with one analytical approach.

摘要

苹果是德国最受欢迎的水果之一，因其在当地易于获取且有益健康而受到重视。在决定购买哪种苹果时，几个特征对消费者来说很重要，包括分类品种、有机种植和产地。为了验证这些特征是否被正确标注，需要强大的分析方法。在本研究中，超高效液相色谱四极杆飞行时间质谱联用仪（UHPLC-Q-ToF-MS）结合随机森林算法应用于193个苹果样本，以分析各种认证问题。在区分德国产和非德国产、德国北部和南部产、有机苹果和传统苹果以及六种不同分类品种方面，准确率分别达到了93.3%、85.5%、85.6%和90%。由于分类模型很大程度上使用了数据的不同部分，这一点通过变量选择得到了体现，所以这种方法非常适合用一种分析方法来回答不同的认证问题。