化学计量学方法在食品分析中的应用

化学计量学是由数学、统计学、化学及计算机科学交叉形成的具有独特魅力的重要新兴学科。作为其核心理论体系的模式识别和多维校正方法,在海量数据信息挖掘与处理、分析信号分辨与解析中,体现了突出的优势,并能攻克传统分析方法难以处理的复杂难题,已广泛应用于食品分析领域。系统介绍了化学计量学核心理论方法及其在食品分析中的应用及研究进展,详细分析了模式识别和多维校正方法的基本原理及优缺点,指出了目前化学计量学方法在食品分析领域研究中需要解决的问题并对其未来发展方向进行了展望。     

连续小波变换-纯变量法解析气相色谱-质谱重叠信号

提出了连续小波变换-纯变量法(CWT-PVA)解析气相色谱-质谱(GC-MS)重叠信号的新方法.在CWT-PVA中,利用CWT对重叠色谱峰处理,得到重叠组分数目,再利用PVA解析重叠质谱数据,得到纯组分质谱信息,并进而计算出各组分色谱信息,在质谱和色谱两方面完成GC-MS重叠信号定性定量解析.结果表明,CWT-PVA是一种简便有效的GC-MS重叠信号解析方法.     

基于主成分分析法对一次性纸杯的拉曼光谱检验研究

为建立一种检验犯罪现场遗留一次性纸杯物证的科学有效的方法,利用拉曼光谱结合化学计量学对收集到的27个不同品牌、不同用途的一次性纸杯样本的拉曼特征峰峰位和峰强度的对比,对一次性纸杯样本进行区分。为了使分类结果科学准确,利用化学计量学的方法对光谱数据进行分析处理。为了减少聚类分析所用的变量个数,利用主成分分析法对拉曼光谱数据进行降维处理,指定提取3个因子,减少变量个数,保留大部分信息,其累计贡献率达99.09%。利用层次聚类和K-Means快速聚类将27个样本分为8类,并利用Fisher判别分析法验证了分类结果的科学合理性,从而对未知样本的类别判断提供依据。     

食品农药残留检测中的化学计量学研究进展

现阶段食品农药残留的检测主要依靠色谱法,但针对日趋复杂的食品农药残留分析,色谱信号中不可避免地出现背景基线和多组分重叠干扰。光谱法近年来凭借其快速无损的优势逐渐受到研究者的青睐,但常用的拉曼光谱和近红外光谱信号不仅容易受到高频噪声和背景基线的干扰,且信号中常出现谱峰重叠的干扰,因此大多需要建立多元校正模型才能实现定量分析;荧光光谱检测信号中也常常出现其它组分信号的干扰,同样较难实现对农药残留的准确分析。化学计量学方法利用"数学分离"的思想,能为食品中农药残留分析提供强大助力。因此,利用化学计量学方法可以用于信号中噪声的消除、背景基线的校正、多元校正模型的建立以及重叠信号的解析。文章总结评述了近十年相关化学计量学方法在食品农药残留分析中的应用研究,旨在为食品农药残留的分析方法和途径提供参考,同时思考了化学计量学方法在食品分析中的研究现状并对其进行了前景展望。     

便携式近红外水果糖度分析模型简化研究

为了实现便携式近红外光谱仪水果糖度现场快速分析,将桃、梨和苹果的光谱进行二阶导数和卷积平滑处理后,利用组合移动窗口偏最小二乘法选择信息变量建立PLS模型,利用遗传偏最小二乘法选择信息变量建立MLR模型.分析表明,桃、梨和苹果PLS模型的RMSEP分别为0.417、0.372和0.654,其RSDP分别为4.685%、3.348%和4.111%;MLR模型的RMSEP分别为0.381、0.382和0.550,其RSDP分别为4.281%、3.438%和3.457%,模型预测精度均满足现场检测应用要求.结果表明, 用SCMWPLS和GA-PLS可以提取最有效信息变量,模型更加简洁、数据运算量也更少,模型适用于便携式近红外光谱仪器.     

二维相关可见-近红外光谱结合支持向量机评价猪肉新鲜度

为探究二维相关同步光谱优选生鲜肉新鲜度特征变量的可行性,采集生鲜猪肉在1～15 d共58 个样品的可见-近红外反射光谱信息,并参照国标方法测定其挥发性盐基氮值（total volatile basic nitrogen,TVB-N）。然后,以TVB-N为“外界扰动”,选择10 条代表性光谱并进行包络线去除,结合光谱差异选取了7?个子区间。通过对每个子区间作二维相关分析,解析其二维相关同步谱和自相关谱,获取与TVB-N变化密切相关的敏感变量。最后,利用所选特征变量,分别基于原始、标准正态变量变换预处理和归一化预处理的光谱,建立猪肉新鲜度的支持向量机（support vector machine,SVM）判别模型。结果表明,利用二维相关光谱分析共提取到17?个特征波长,仅占总变量个数的1.61%,建立的SVM模型总体正确率分别为94.83%、98.28%和98.28%。这表明所建立的模型具有较好的判别效果,也说明二维相关分析用于筛选与生鲜肉新鲜度相关特征变量的方法是可行的。这有利于解析生鲜肉在腐败变质过程中的光谱特征信息变化,也为近红外光谱分析中变量筛选提供了一种新的思路。     

CARS-SVM预测哈密瓜可溶性固形物含量

采用近红外光谱技术结合数据降维的方法,建立了哈密瓜可溶性固形物含量的预测模型：对比多种光谱预处理方法,确定二阶求导用于处理原始光谱;经预处理的光谱数据分别结合特征选择竞争性自适应重加权采样法（CARS）、蒙特卡罗无信息变量消除法（MC-UVE）提取特征波长,以及利用主成分分析进行降维;再使用特征选择和特征提取的光谱数据作为模型的输入变量,建立哈密瓜可溶性固形物含量预测模型。结果显示,CARS+SVM建立的预测模型最优,模型的校正集相关系数为0.981 4,预测集相关系数为0.900 2,模型能够准确预测哈密瓜可溶性固形物含量。     

BP 神经网络在大豆油酸价近红外光谱检测中的应用

对于大豆四级油,采用BP 神经网络对其近红外光谱数据建模,对系统的结构及参数选取进行了分析,对样本训练集的设计和网络输入端的主因子方面进行了处理。对于其他的多变量建模方法,分析了其对近红外光谱有用信息的提取作用。结果显示：多元线性回归、主成分回归和偏最小二乘法对大豆四级油酸价预测的标准偏差分别为0.1472%、0.1801% 和0.1576%,BP 神经网络的预测标准偏差为0.1387%。     

基于可见/近红外光谱的香梨贮藏时间区分方法

采用可见/近红外光谱方法区分不同贮藏时间的香梨。使用光纤光谱仪记录香梨样品的可见/近红外光谱反射信号,用随机共振提取香梨样品光谱数据的特征信息。与原始光谱数据比较,基于随机共振输出信噪比特征值的主成分分析法区分不同香梨样品的结果更好。该方法具有速度快、无损、成本低等优势,有望在食品品质快速分析领域得到应用。     

近红外光谱结合化学计量法快速无损鉴别燕麦

提出了一种基于近红外光谱技术与化学计量学的燕麦无损鉴别方法。通过近红外光谱仪测定了5个品牌与劣质燕麦的光谱曲线,利用连续小波变换方法对光谱进行预处理,然后基于标准偏差与相对标准偏差的变量筛选方法筛选出具有代表的15个波数点,最后结合主成分分析法对不同燕麦样品快速鉴别。结果表明:连续小波变换可以有效地消除光谱中的背景干扰,提取光谱有效信息,波长筛选方法可以大大提高主成分分析结果的鉴别能力。通过结合近红外光谱分析技术与化学计量学方法,可对中国国产品牌、进口品牌和劣质燕麦进行准确鉴别。     

Visual detection of apple bruises using AdaBoost algorithm and hyperspectral imaging

Hyperspectral imaging technique (400–1000 nm) was used for rapid and nondestructive recognition of bruises of apples. A total of 324 hyperspectral images were collected from 108 Fuji apples and the average spectral reflectance was extracted from the region of interest (ROI) of each image. The classification results of AdaBoost for the data pretreated by various existing methods were compared. Then, the correlation-based feature selection (CFS) algorithm was used to obtain characteristic wavelengths for reducing data redundancy. After pretreating with multiplicative scatter correction (MSC) and CFS, the average accuracy of the selected wavelengths was 97.63%. Then, an image processing algorithm based on the characteristic wavelengths selected before was proposed for the visual discrimination of bruises. This algorithm performed independent component analysis (ICA) transformation of the selected wavelengths, and chose the third component image of the ICA transform, then used adaptive threshold segmentation to obtain the bruise region of apples. The results showed that hyperspectral imaging technology could discriminate apple bruise, and this study can help to develop an online apple bruises detection system.     

Analysis of volatile compounds in New Zealand unifloral honeys by SPME–GC–MS and chemometric-based classification of floral source

New Zealand unifloral honeys have a higher commercial value than polyfloral honeys; however identification of floral source can be difficult and time-consuming. In this study, we aimed to establish a rapid and semi-automated method for identifying the floral source of New Zealand honeys. Volatile compounds from ten types of New Zealand unifloral honeys (a total of 234 samples) were analyzed by solid-phase microextraction (SPME) and gas chromatography coupled to mass spectrometry (GC–MS). For 37 compounds, probability plots of log₁₀[GC–MS peak area] versus cumulative probability enabled visual identification of those that could be possible markers used to discriminate floral source. GC–MS peak areas were also analyzed by hierarchical cluster analysis and principal component analysis. Results showed data falling into groups based on floral source, indicating that supervised pattern recognition could be used to build a model with which to classify honeys based on floral source. A model was built using WEKA (Waikato Environment for Knowledge Analysis) machine-learning software. The logistic model tree algorithm in WEKA produced a model that classified 89.8 % of samples correctly. Overall, results show that the methods employed here have the potential to be used as a basis for routine testing and classification of New Zealand unifloral honeys.     

Impact of trehalose,sucrose and/or maltodextrin addition on aroma retention in freeze dried strawberry puree

Trehalose was studied as a drying aid to establish its impact on aroma retention in freeze dried strawberry puree. The evaluation was done by sensory analysis and headspace‐solid phase microextraction‐gas chromatography with a mass detector (HS‐SPME‐GC‐MS); results were compared with those obtained using sucrose and maltodextrin (MD) as drying aids. The carbohydrate used significantly modified the type and concentration of volatiles retained during freeze drying. HS‐SPME‐GC‐MS outcomes showed that the use of trehalose (alone or with MD) resulted in the product with the chromatographic profile most similar to fresh strawberry puree, being different from sucrose and MD. Meanwhile, sensory analysis showed a similarity with the aromatic profiles when using trehalose or sucrose, remaining both different from MD. This study proved that the use of trehalose as a drying aid can be beneficial on volatile aroma retention and that different combinations of organic volatiles can lead to similar sensory profiles.     

Determination of Milk Fat Adulteration with Vegetable Oils and Animal Fats by Gas Chromatographic Analysis

This study assessed the potential application of gas chromatography (GC) in detecting milk fat (MF) adulteration with vegetable oils and animal fats and of characterizing samples by fat source. One hundred percent pure MF was adulterated with different vegetable oils and animal fats at various concentrations (0%, 10%, 30%, 50%, 70%, and 90%). GC was used to obtain the fatty acid (FA) profiles, triacylglycerol (TG) contents, and cholesterol contents. The pure MF and the adulterated MF samples were discriminated based on the total concentrations of saturated FAs and on the 2 major FAs (oleic acid [C18:1n9c] and linoleic acid [C18:2n6c], TGs [C52 and C54], and cholesterol contents using statistical analysis to compared difference. These bio‐markers enabled the detection of as low as 10% adulteration of non‐MF into 100% pure MF. The study demonstrated the high potential of GC to rapidly detect MF adulteration with vegetable and animal fats, and discriminate among commercial butter and milk products according to the fat source. These data can be potentially useful in detecting foreign fats in these butter products. Furthermore, it is important to consider that several individual samples should be analyzed before coming to a conclusion about MF authenticity.     

Tocopherol,fatty acids and sterol distributions in wild Ecuadorian Theobroma subincanum (Sterculiaceae) seeds

Ecuadorian Theobroma subincanum (Sterculiaceae) seed parts were analyzed to determine quali-quantitative tocopherol distribution. Fatty acids and sterols in the embryos, teguments and endosperm were also evaluated with an aim to better-characterize the plant matrix for its potential use as a source of phytochemicals, for the cocoa butter processing industry and/or as a new source of low cost natural products in the cosmetic, drug and alimentary/nutraceutical industries. HPLC for tocopherols and gas-chromatography (GC), GC–mass spectrometry (MS) analyses for fatty acids profile (FAP) and phytosterols were performed. Tocopherols were particularly abundant in the embryo, with quali-quantitative data similar to wheat germ oil whereas, in the teguments and endosperm, the concentrations of tocopherols were lower. The fatty acid profile and phytosterol characterization of the seed parts showed qualitative homogeneous data. In the endosperm, 80% of the entire FAP consisted of oleic and stearic acid while, among sterols, cycloartenol was more abundant in endosperm than in embryos and teguments. Accordingly, T. subincanum seeds can be proposed as possible substitutes in the cocoa processing industry and as a potential source of vitamin E isomers.     

Practical non-targeted gas chromatography/mass spectrometry-based metabolomics platform for metabolic phenotype analysis

Gas chromatography coupled to mass spectrometry (GC/MS) is a core analytical method for metabolomics and has been used as a platform in non-targeted analysis, especially for hydrophilic metabolites. Non-targeted GC/MS-based metabolomics generally requires a high-throughput technology to handle a large volume of samples and an accumulated database (reference library) of the retention times and mass spectra of standard compounds for accurate peak identification. In this study, we provide a practical GC/MS platform and an auto peak identification technique that is not restricted to certain types of mass spectrometers. The platform utilizes a quadrupole mass spectrometer capable of high-speed scanning, resulting in greater output compared with Pegasus GC-time of flight (TOF)/MS, which has been an essential instrument for high-throughput experiments. Moreover, we show that our reference library is broadly applicable to other instruments; peak identification can be readily performed using the library without constructing a reference resource. The usefulness and versatility of our system are demonstrated by the analyses of three experimental metabolomics data sets, including standard mixtures and real biological samples.     

Impact of botanical source and processing conditions on physicochemical properties and antioxidant activity of honey in the northern part of Thailand

Phenolics, melissopalynological analysis, and selected physicochemical characteristics could differentiate floral honey origin. Botanical sources (longan, multifloral and Manuka) had more pronounced effect on honey properties than processing conditions. The conductivity was related to ash content and colour of honey. Total sugar, fructose and glucose content in all samples were not significantly different. Melissopalynological analysis was used to verify the botanical source of honey samples. Predominant pollen was identified. Results characterising antioxidant properties (DPPH radicals, hydroxyl radicals, and carbon-centred radicals scavenging activity) were confirmed via both Electron Spinning Resonance and oxygen radical absorbance capacity (ORAC). They showed that only hydroxyl radicals scavenging activity and ORAC were significantly different. Type and amount of phenolics in honey samples identified using HPLC could differentiate botanical source. Gallic acid and kaempferol were potential marker phenolics for longan honey. Multivariate statistical analysis allowed discrimination of honey from different botanical sources based on their phenolic profiles.     

Geographical origin identification and quality control of Chinese chrysanthemum flower teas using gas chromatography–mass spectrometry and olfactometry and electronic nose combined with principal component analysis

Aroma profiles and volatile profiles were established based on volatile and aroma ‐ active components identified by gas chromatography–mass spectrometry and olfactometry (GC–MS–O) and electronic nose (E ‐ nose). The two profiles were used for quality control and origin identification of chrysanthemum flower teas. Results showed that 86 volatile components were identified in five chrysanthemum flower teas, including terpenes, alcohols, ketones, aldehydes, esters and others. Of them, 33 aroma ‐ active components were recognised, including 10 aroma categories. The aroma profiles and volatile profiles were established by 10 aroma categories and E ‐ nose. Chrysanthemum flower teas were divided into five groups on PCA score plots based on their aroma profiles and volatile profiles, and the key volatile (aroma ‐ active) components resulting in the tea sample differences were determined. The quality of chrysanthemum flower teas could be evaluated according to aroma ‐ active components and aroma profiles and origins could be discriminated by PCA combined with GC–MS–O and E ‐ nose.     

Differentiation of fresh and frozen/thawed fish, European sea bass (Dicentrarchus labrax), gilthead seabream (Sparus aurata), cod (Gadus morhua) and salmon (Salmo salar), using volatile compounds by SPME/GC/MS

BACKGROUND: A simple method based on solid phase microextraction/gas chromatography/mass spectrometry (SPME/GC/MS) was applied for studying the volatile profiles of whole fish samples of European sea bass (Dicentrarchus labrax) and gilthead seabream (Sparus aurata) and fillets of cod (Gadus morhua) and salmon (Salmo salar) during frozen storage in order to be able to differentiate a fresh product from one that has been frozen. Analysis of volatile compounds was performed on these two product types, fresh and after freezing/thawing following storage at ? 20 °C for 30 and 90 days. RESULTS: More than a hundred volatile compounds were found by SPME/GC/MS. Statistical processing by principal component analysis and ascending hierarchical classification was used to classify the samples into categories and verify the possibility of separating fresh samples from those that had been frozen and thawed. The compounds to be used as differentiators were identified. Four compounds were common to all species: dimethyl sulfide, 3‐methylbutanal, ethyl acetate and 2‐methylbutanal. Not only were they found in larger quantities after thawing but they also increased with the duration of storage at ? 20 °C. CONCLUSION: These four compounds can therefore be considered as potential markers of differentiation between a fresh product and one that has been frozen. Copyright © 2012 Society of Chemical Industry