基于HPLC-DAD三维色谱指纹图谱的国产赤霞珠红葡萄酒产地识别研究OA
Geographical origin identification of Chinese Cabernet Sauvignon red wines based on HPLC-DAD three-dimensional chromatographic fingerprinting
[目的]丰富与发展葡萄酒真实性鉴别及产地溯源技术.[方法]采用高效液相色谱-光电二极管阵列检测(HPLC-DAD)法采集我国秦皇岛、银川和吐鲁番3 个著名产地共45 个赤霞珠红葡萄酒的三维色谱指纹图谱,并借助化学计量学多元曲线分辨-交替最小二乘(MCR-ALS)算法解析上述图谱数据,基于所得有效组分的相对浓度,利用主成分分析(PCA)、偏最小二乘-判别分析(PLS-DA)和支持向量机(SVM)3 种机器学习算法对红葡萄酒样品进行产地区分.[结果]通过 MCR-ALS 解析算法共获得56 个有效组分,PCA 得分显示各产地赤霞珠红葡萄酒有依照产地分类的趋势,且 PLS-DA 和 SVM 模型的分类效果均良好,训练集和预测集识别准确率都可达 100%.此外,基于 22 个差异变量建立的 VIP-PLS-DA 模型同样可对葡萄酒产地进行100%的准确判别.[结论]HPLC-DAD 三维色谱指纹图谱技术结合机器学习算法能建立稳定可靠的识别模型,实现3 个国产赤霞珠红葡萄酒产地客观、准确的鉴别.
[Objective]This study aimed to advance the authenticity identification and origin traceability techniques of wines.[Methods]The three-dimensional chromatographic fingerprints of 45 Cabernet Sauvignon red wines from three famous producing regions in China(Qinhuangdao,Yinchuan,and Turpan)were acquired using high-performance liquid chromatography with diode array detection(HPLC-DAD).The data were then resolved by multivariate curve resolution-alternating least squares(MCR-ALS).Based on the relative concentrations of the resolved components,three machine learning algorithms—principal component analysis(PCA),partial least squares-discriminant analysis(PLS-DA),and support vector machine(SVM)—were applied to discriminate the geographical origins of the red wines.[Results]A total of 56 resolved components were obtained by MCR-ALS analysis of the three-dimensional chromatographic fingerprints.PCA score plots showed a tendency of the wines to cluster according to their geographical origins.Both PLS-DA and SVM models achieved good classification performance,with 100%recognition accuracy for both the training and prediction sets.Furthermore,a VIP-PLS-DA model based on 22 differential variables enabled accurate discrimination of the geographical origins of the wines with 100%.[Conclusion]The HPLC-DAD three-dimensional chromatographic fingerprinting technique combined with machine learning algorithms can establish a stable and reliable recognition model,and is expected to provide objective and accurate identification of three geographical origins of Chinese Cabernet Sauvignon red wines.
邹华;潘圆;强海清;孟展;尹小丽;谷惠文
长江大学 化学与环境工程学院/生命科学学院,湖北 荆州 434023长江大学 化学与环境工程学院/生命科学学院,湖北 荆州 434023长江大学 化学与环境工程学院/生命科学学院,湖北 荆州 434023长江大学 化学与环境工程学院/生命科学学院,湖北 荆州 434023长江大学 化学与环境工程学院/生命科学学院,湖北 荆州 434023长江大学 化学与环境工程学院/生命科学学院,湖北 荆州 434023
轻工纺织
赤霞珠红葡萄酒HPLC-DAD三维色谱指纹图谱产地识别机器学习
Cabernet Sauvignon red wineHPLC-DADthree-dimensional chromatographic fingerprintingorigin identificationmachine learning
《轻工学报》 2026 (3)
10-18,9
国家自然科学基金项目(3237150132272409)宁夏回族自治区重点研发计划重点项目(2026BEG02035)
评论