肝细胞癌、肝硬化患者血清中代谢物组研究

采用核磁共振方法检测肝硬化、肝细胞癌(简称肝癌)患者和健康人血清中代谢物,研究3组血清代谢物组的差异.利用偏最小二乘法-判别式分析(partial least square-discriminant analysis, PLS-DA)对NMR谱数据进行模式识别分析,探讨利用基于1H NMR代谢组学技术诊断肝癌的可行性.结果表明,与健康人相比,肝硬化、肝癌患者血清中脂质(低密度脂蛋白和极低密度脂蛋白)、胆碱、乙酰乙酸等含量减少,谷氨酰胺、丙酮酸、苯丙氨酸、酪氨酸等含量增加.PLS-DA分析结果显示肝癌患者可与健康人、肝硬化患者鉴别开来,肝癌诊断灵敏度达921%,假阳性率为5.7%,优于血清甲胎蛋白(alpha-fetoprotein, AFP)检测.研究结果表明,基于1H NMR代谢组学技术结合PLS-DA的方法具有灵敏、准确、重复性好等优点,有助于肝癌早期诊断.     

阿卡波糖对Ⅱ型糖尿病大鼠尿液代谢轮廓的影响

采用超高效液相色谱-质谱联用(UPLC-MS/MS)方法研究了阿卡波糖对Ⅱ型糖尿病大鼠代谢轮廓的影响, 分析了健康组、 糖尿病模型组和糖尿病给予阿卡波糖组的大鼠尿样, 采用主成分分析法(PCA)和偏最小二乘法-判别分析(PLS-DA)对数据进行分析. PCA得分图表明, 健康组、 糖尿病组和阿卡波糖组的代谢轮廓有显著差别, 根据PLS-DA载荷图筛选, 将对各组分离贡献大的化合物的串联质谱分析数据经Human Metabolome Database(HMDB)和Mass Bank.jp等数据库检索, 进行质谱信息匹配, 鉴定出苯乙酰甘氨酸、 肌酐及葡萄糖酸等8种内源性代谢物为潜在生物标记物.     

基于核磁共振定量氢谱技术的乳品脂质分析及方法建立

基于核磁共振定量氢谱(q1H NMR)技术,建立了乳品脂质分析的方法,并对市场上有机牛奶、山羊奶、大豆奶中的ω-3,DHA,EPA等脂质进行定量分析。就ω-3而言,有机牛奶山羊奶大豆奶;就DHA而言,山羊奶有机牛奶大豆奶;大豆奶中不含EPA,有机牛奶中EPA山羊奶。此外,对不同种类奶中脂质的积分数据进行偏最小二乘判别分析(PLS-DA)及方差分析(ANOVA),进一步明确了其脂质种类及含量分布的差异,并对其进行初步营养学评价。     

多元曲线分辨-交替最小二乘方法在联用数据分辨中的应用

通过对模拟数据和高效毛细管电泳实验数据的分析,讨论了多元曲线分辨-交替最小二乘方法(MCR-ALS)在毛细管电泳-二极管阵列检测(CE-DAD)联用数据分辨中的应用.讨论了几种因素对MCR-ALS单个数据矩阵分辨结果的影响,包括待分析物光谱间的相似程度、浓度曲线的重叠程度以及由渐进因子分析(EFA)所得到的浓度初始值等.MCR-ALS还可用于多个数据矩阵的同时分析,即二阶MCR-ALS.结果表明,与一阶MCR-ALS相比,二阶MCR-ALS方法能够更好地解决各种分辨问题,得到合理和满意的分辨结果.     

气相色谱-质谱联用技术分析大气细颗粒物对小鼠肺组织代谢轮廓的影响

建立了基于气相色谱-质谱联用技术(GC-MS)分析大气细颗粒物(PM2.5)滴注对小鼠肺组织代谢轮廓的影响的研究方法.通过分析肺组织细胞内代谢物的变化,研究不同浓度PM2.5对小鼠肺组织代谢的毒性机制.鼻腔分别滴注0、7.5、20.0和37.5 g/L的PM2.5悬液,提取肺组织胞内物质,预处理后进行GC-MS分析,结合主成分分析法(PCA)、偏最小二乘判别分析法(PLS-DA)进行数据解析,通过PLS-DA得分图可将不同PM2.5染毒浓度下的肺组织胞内物质明显区分.运用PLS-DA载荷图及模型的变量重要性因子(VIP)值,发现了7种代谢物可作为区别不同浓度PM2.5下代谢组的潜在生物标志物,分别为丙氨酸、缬氨酸、亮氨酸、鸟氨酸、延胡索酸、柠檬酸、嘌呤(p<0.01).代谢途径分析结果表明,PM2.5滴注使小鼠肺组织受到氧化损伤,氧化应激反应增强,抑制了三羧酸循环(TCA循环)及嘌呤代谢.本研究为深入解析PM2.5致毒机理提供了新的方法及理论依据.     

基于色谱和光谱数据融合的不同植物源食用油判别方法与模型

利用气相色谱和近红外光谱技术对不同植物源的4种食用油(葵花籽油、大豆油、玉米油和花生油)进行表征分析,基于表征数据分别建立了偏最小二乘判别分析(PLS-DA)模型,并在此基础上探究了数据级数据融合方法,构建了基于色谱和光谱数据融合的不同植物源食用油判别方法与模型。主成分分析(PCA)结果显示,气相色谱判别分析主要是依据脂肪酸组成信息,近红外光谱主要是基于样本中含氢化学键的表征进行分类。数据融合模型的灵敏度和特异度均为1000,分类误差为0000,降低了交互验证的平均分类误差,模型具有良好的稳健性。与基于单一数据的模型结果相比,数据融合分析策略提高了模型的分类精度和鲁棒性。     

基于氢核磁共振技术和化学计量学方法鉴别蜂蜜品种

利用氢核磁共振(~1H NMR)技术结合化学计量学方法对不同品种的蜂蜜进行鉴别。采集33个洋槐蜜、48个油菜蜜、63个荔枝蜜的核磁指纹图谱,对数据进行不同方式的预处理后,采用有监督的偏最小二乘判别分析(PLS-DA)和正交偏最小二乘判别分析(OPLS-DA)建立判别模型。结果表明,不同的数据预处理方式对模型解释能力和预测能力的影响较大,自标度化(UV)模式更适于蜂蜜核磁数据的分析。建立的OPLS-DA模型可有效地分离判别3种蜂蜜,所建模型对3种蜂蜜的判别解释能力达95.8%,对未知样本的预测能力为90.5%。因此,利用~1H NMR结合OPLS-DA方法可有效地实现不同品种蜂蜜的快速鉴别。     

正交信号校正在正常成人血清1H NMR谱的代谢组分析中的滤噪作用评价

观察、比较正交信号校正(OSC)滤噪前后, 用不同的模式识别方法对正常成人血清代谢组¹H NMR谱进行分析的效果, 以探讨NMR代谢组学技术应用于临床研究和疾病早期诊断的可行性. 78例正常成人在采血前按常规要求禁食8 h, 记录血清一维600 MHz氢谱后, 分别采用主成分分析(PCA)、偏最小二乘法-判别分析(PLS-DA)以及簇类的独立软模式法(SIMCA)对氢谱进行模式识别分析. 结果表明: 虽然采血前并无其它诸如饮食、生活方式、生理周期等方面的严格限制, 采用OSC 滤噪后, PLS-DA能够完全区分不同性别的血清氢谱, 其判别能力优于PCA和SIMCA. 而且采用OSC滤噪与文献报道的未经OSC处理的PLS-DA法获得的与性别分类有关的主要NMR积分区段基本相同. 从OSC去除不同数目的隐变量后所致的PLS-DA模型的性能改变可见: OSC去除两个隐变量时, 前两个隐变量的特征值明显比后面的大; 剩余残差为20.82%, 即去除了79.18%的X变量中与反应变量Y不相关的系统变异. 此时PLS-DA计算所得的隐变量个数为1; 而不使用OSC或用OSC去除一个隐变量时, PLS-DA所得的隐变量个数分别为3和2. 作为PLS-DA模型质量的评价指标, R2X表示PLS-DA模型计算所获得的隐变量反映自变量X的变异的百分比, R2Y则表示隐变量反映因变量Y的变异的百分比, Q2 (cum)为交叉验证后PLS-DA模型所获隐变量能够预测X和Y变异的累计百分比. R2X在OSC去除两个隐变量时达到最低值, 表明此时PLS-DA计算模型包含的系统变异最少; R2Y与Q2 (cum)都达到80%以上并趋于稳定, 说明OSC去除两个隐变量时PLS-DA模型的质量优良. 显然, OSC可去除饮食、环境等因素的影响, 降低临床样本的不均一性, 这对于NMR代谢组学技术应用于临床研究至关重要. OSC滤噪去除的隐变量个数应根据剩余残差、去除隐变量的特征值大小、PLS-DA模型计算所得的隐变量个数和反映模型质量的相关指标加以判断.     

高效液相色谱全轮廓指纹图谱结合化学计量学区分不同地区的紫苏叶

利用高效液相色谱全轮廓指纹图谱结合化学计量学方法对不同栽培地区的紫苏叶样品(共84个)进行区分。全轮廓色谱数据经自适应迭代加权最小二乘法(airPLS)和相关优化翘曲法(COW)校正后,基线和保留时间漂移现象均得到明显改善。经预处理后的色谱数据采用主成分分析(PCA)进行解析,结果表明不同来源的样品能按其特性各自聚为一类;而分段间隔压缩变量后的色谱数据经主成分分析处理可得到与全轮廓色谱数据为输入变量时相一致的结果。此外,偏最小二乘判别分析(PLS-DA)对于紫苏叶样品分类的识别能力和预报能力分别为92.8%和89.6%。     

正交信号校正在正常成人血清1H NMR谱的代谢组分析中的滤噪作用评价

观察、比较正交信号校正(OSC)滤噪前后, 用不同的模式识别方法对正常成人血清代谢组¹H NMR谱进行分析的效果, 以探讨NMR代谢组学技术应用于临床研究和疾病早期诊断的可行性. 78例正常成人在采血前按常规要求禁食8 h, 记录血清一维600 MHz氢谱后, 分别采用主成分分析(PCA)、偏最小二乘法-判别分析(PLS-DA)以及簇类的独立软模式法(SIMCA)对氢谱进行模式识别分析. 结果表明: 虽然采血前并无其它诸如饮食、生活方式、生理周期等方面的严格限制, 采用OSC 滤噪后, PLS-DA能够完全区分不同性别的血清氢谱, 其判别能力优于PCA和SIMCA. 而且采用OSC滤噪与文献报道的未经OSC处理的PLS-DA法获得的与性别分类有关的主要NMR积分区段基本相同. 从OSC去除不同数目的隐变量后所致的PLS-DA模型的性能改变可见: OSC去除两个隐变量时, 前两个隐变量的特征值明显比后面的大; 剩余残差为20.82%, 即去除了79.18%的X变量中与反应变量Y不相关的系统变异. 此时PLS-DA计算所得的隐变量个数为1; 而不使用OSC或用OSC去除一个隐变量时, PLS-DA所得的隐变量个数分别为3和2. 作为PLS-DA模型质量的评价指标, R2X表示PLS-DA模型计算所获得的隐变量反映自变量X的变异的百分比, R2Y则表示隐变量反映因变量Y的变异的百分比, Q2 (cum)为交叉验证后PLS-DA模型所获隐变量能够预测X和Y变异的累计百分比. R2X在OSC去除两个隐变量时达到最低值, 表明此时PLS-DA计算模型包含的系统变异最少; R2Y与Q2 (cum)都达到80%以上并趋于稳定, 说明OSC去除两个隐变量时PLS-DA模型的质量优良. 显然, OSC可去除饮食、环境等因素的影响, 降低临床样本的不均一性, 这对于NMR代谢组学技术应用于临床研究至关重要. OSC滤噪去除的隐变量个数应根据剩余残差、去除隐变量的特征值大小、PLS-DA模型计算所得的隐变量个数和反映模型质量的相关指标加以判断.     

超高效液相色谱/飞行时间质谱法分析尿液中的代谢物用于区分人类性别的研究

尿中的代谢产物可以反映生命个体的生理状态。为了考察在非严格控制条件下(即对志愿者的饮食、生活方式以及样品采集时间等诸多条件不加以控制)基于尿中代谢物的指纹图谱对男女性别进行区分的可行性,采用超高效液相色谱/飞行时间质谱(UPLC/TOF-MS)联用技术分析了31个随机尿样,并用主成分分析法（PCA）和偏最小二乘法判别分析（PLS-DA）两种数据处理方法对数据进行处理,与PCA法比较,PLS-DA法能提高分类效果,并筛选出4种可能的与性别相关的生物标记物。研究结果表明,UPLC/MS联用技术通量高,数据量丰富;模式识别数据处理方法适合于从大量数据中提取信息,两者的结合有利于代谢组学的研究。     

基于液相色谱-串联质谱的结肠癌血清氧固醇标志物筛选

结肠癌(CC)是全球常见恶性肿瘤之一,发病率呈逐年上升趋势,目前没有有效的标志物用于疾病早期诊断和干预跟踪。胆固醇及其氧化衍生物氧固醇在众多恶性肿瘤发生发展中发挥关键作用。该研究采用液相色谱-串联质谱(LC-MS/MS)技术,对CC临床血清样本中胆固醇及相关10种氧固醇代谢物进行了定性定量分析,并采用偏最小二乘判别分析(PLS-DA)和正交偏最小二乘判别分析(OPLS-DA)进行多元统计分析,发现上述目标代谢物能够较好地区分CC组与健康对照组。为防止数据过拟合,该研究在PLS-DA模型各代谢物变量投影重要性(VIP)基础上,结合最优组分数及K-均值聚类结果,筛选得到3种代谢标志物。通过受试者操作特征曲线(ROC)的曲线下面积(AUC)分析,发现筛选得到的3种潜在标志物联合预测CC达到0.998,说明模型性能优良。GO(基因本体论)富集分析显示3种潜在标志物主要分布在内质网和包被囊泡上,参与胆固醇代谢、运输、低密度脂蛋白重塑等生物进程,发挥胆固醇运输活性和低密度脂蛋白颗粒受体结合的分子功能。KEGG(京都基因与基因组百科全书)通路分析显示3种潜在标志物富集于类固醇生物合成、PPAR(过氧化物酶体增殖物激活受体)信号通路及ABC(ATP结合盒)转运等通路上。该研究为寻找CC标志物及进一步阐明胆固醇及氧固醇在CC发病过程中的作用奠定了一定的基础。     

顶空固相微萃取-气相色谱/质谱联用法结合化学计量学分析白肋烟烘焙前后挥发性、半挥发性成分

采用顶空固相微萃取-气相色谱/质谱法（HS-SPME-GC/MS）分析了白肋烟烟叶中挥发性、半挥发性成分。20 mg烟粉在60℃条件下孵化8 min,采用聚二甲基硅氧烷/二乙烯基苯（PDMS/DVB）65 μm纤维头萃取40 min,然后在250℃解吸3 min,通过与标准品和质谱数据库进行比对,初步定性了白肋烟烟叶中122种挥发性、半挥发性成分,并采用内标法进行半定量分析。通过主成分分析（PCA）和偏最小二乘-判别分析（PLS-DA）等化学计量学方法,直观反映了白肋烟烘焙前后挥发性、半挥发性成分的变化。该方法具有样品用量小、前处理简单、灵敏度高等特点,结合化学计量学方法可用于白肋烟烘焙前后化学成分变化分析,为白肋烟烘焙条件的优化提供了科学的检测方法。     

Discrimination of Osteonecrosis and Normal Tissues by Near-Infrared Spectroscopy and Successive Projections Algorithm-Linear Discriminant Analysis

Osteonecrosis of femoral head (ONFH) is a disease characterized by an impaired blood flow in the bone. The pathogenesis is still unknown, which makes an exact diagnosis troublesome and heavily dependent on experience. Exploring the information of molecular level by modern spectroscopy may help to discover the underlying pathogenesis and find its diagnostic application in clinical medicine. The study focuses on the combination of near-infrared (NIR) spectroscopy and classification models for discriminating ONFH and normal tissues. A total of 128 surgical specimens was prepared and NIR spectra were recorded by an integrating sphere. The experiment data set was divided into three subsets, i.e., the training set, validation set, and test set. Successive projection algorithm-linear discriminant analysis (SPA-LDA) was used to compress variables and build the diagnostic model. Partial least square-discriminant analysis (PLS-DA) was used as the reference. Principal component analysis (PCA) was used for exploratory analysis. The results showed that compared to PLS-DA, SPA-LDA provided a more parsimonious model using only seven variables and achieved better performance, i.e., sensitivity of 90.5 and 85%, and specificity of 100 and 95.5% for the validation and test sets, respectively. It indicated that NIR spectroscopy combined with SPA-LDA algorithm was a feasible aid tool for discriminating ONFH from normal tissue.     

Probing beer aging chemistry by nuclear magnetic resonance and multivariate analysis

This paper describes the use of nuclear magnetic resonance (NMR) spectroscopy, in tandem with multivariate analysis (MVA), for monitoring the chemical changes occurring in a lager beer exposed to forced aging (at 45 °C for up to 18 days). To evaluate the resulting compositional variations, both principal component analysis (PCA) and partial least squares-discriminant analysis (PLS-DA) were applied to the NMR spectra of beer recorded as a function of aging and a clear aging trend was observed. Inspection of PLS-DA loadings and peak integration enabled the changing compounds to be identified, revealing the importance of well known markers such as 5-hydroxymethylfurfural (5-HMF) as well as a range of other relevant compounds: amino acids, higher alcohols, organic acids, dextrins and some still unassigned spin systems. In addition, the multivariate analysis method of 2D correlation analysis was applied to the NMR data enabling the relevant compound variations to be confirmed and inter-compound correlations to be assessed, some reflecting common metabolic/chemical pathways and, therefore, offering improved insight into the chemical aspects of beer aging.     

Phospholipidomic identification of potential plasma biomarkers associated with type 2 diabetes mellitus and diabetic nephropathy

Type 2 diabetes mellitus (T2DM) and its attendant complications, such as diabetic nephropathy (DN), impose a significant societal and economic burden. The investigation of discovering potential biomarkers for T2DM and DN will facilitate the prediction and prevention of diabetes. Phospholipids (PLs) and their metabolisms are closely allied to nosogenesis and aggravation of T2DM and DN. The aim of this study is to characterize the human plasma phospholipids in T2DM and DN to identify potential biomarkers of T2DM and DN. Normal phase liquid chromatography coupled with time of flight mass spectrometry (NPLC-TOF/MS) was applied to the plasma phospholipids metabolic profiling of T2DM and DN. The plasma samples from control (n = 30), T2DM subjects (n = 30), and DN subjects (n = 52) were collected and analyzed. The significant difference in metabolic profiling was observed between healthy control group and DM group as well as between control group and DN group by the help of partial least squares discriminant analysis (PLS-DA). PLS-DA and one-way analysis of variance (ANOVA) were successfully used to screen out potential biomarkers from complex mass spectrometry data. The identification of molecular components of potential biomarkers was performed on Ion trap-MS/MS. An external standard method was applied to quantitative analysis of potential biomarkers. As a result, 18 compounds in 7 PL classes with significant regulation in patients compared with healthy controls were regarded as potential biomarkers for T2DM or DN. Among them, 3 DM-specific biomarkers, 8 DN-specific biomarkers and 7 common biomarkers to DM and DN were identified. Ultimately, 2 novel biomarkers, i.e., PI C18:0/22:6 and SM dC18:0/20:2, can be used to discriminate healthy individuals, T2DM cases and DN cases from each other group.     

Assessing corn stover composition and sources of variability via NIRS

Corn stover, the above-ground, non-grain portion of the crop, is a large, currently available source of biomass that potentially could be collected as a biofuels feedstock. Biomass conversion process economics are directly affected by the overall biochemical conversion yield, which is assumed to be proportional to the carbohydrate content of the feedstock materials used in the process. Variability in the feedstock carbohydrate levels affects the maximum theoretical biofuels yield and may influence the optimum pretreatment or saccharification conditions. The aim of this study is to assess the extent to which commercial hybrid corn stover composition varies and begin to partition the variation among genetic, environmental, or annual influences. A rapid compositional analysis method using near-infrared spectroscopy/partial least squares multivariate modeling (NIR/PLS) was used to evaluate compositional variation among 508 commercial hybrid corn stover samples collected from 47 sites in eight Corn Belt states after the 2001, 2002, and 2003 harvests. The major components of the corn stover, reported as average (standard deviation) % dry weight, whole biomass basis, were glucan 31.9 (2.0), xylan 18.9 (1.3), solubles composite 17.9 (4.1), and lignin (corrected for protein) 13.3 (1.1). We observed wide variability in the major corn stover components. Much of the variation observed in the structural components (on a whole biomass basis) is due to the large variation found in the soluble components. Analysis of variance (ANOVA) showed that the harvest year had the strongest effect on corn stover compositional variation, followed by location and then variety. The NIR/PLS rapid analysis method used here is well suited to testing large numbers of samples, as tested in this study, and will support feedstock improvement and biofuels process research.     

X‐tended target projection (XTP)—comparison with orthogonal partial least squares (OPLS) and PLS post‐processing by similarity transformation (PLS + ST)

Target projection (TP) also called target rotation (TR) was introduced to facilitate interpretation of latent‐variable regression models. Orthogonal partial least squares (OPLS) regression and PLS post‐processing by similarity transform (PLS + ST) represent two alternative algorithms for the same purpose. In addition, OPLS and PLS + ST provide components to explain systematic variation in X orthogonal to the response. We show, that for the same number of components, OPLS and PLS + ST provide score and loading vectors for the predictive latent variable that are the same as for TP except for a scaling factor. Furthermore, we show how the TP approach can be extended to become a hybrid of latent‐variable (LV) regression and exploratory LV analysis and thus embrace systematic variation in X unrelated to the response. Principal component analysis (PCA) of the residual variation after removal of the target component is here used to extract the orthogonal components, but X‐tended TP (XTP) permits other criteria for decomposition of the residual variation. If PCA is used for decomposing the orthogonal variation in XTP, the variance of the major orthogonal components obtained for OPLS and XTP is observed to be almost the same, showing the close relationship between the methods. The XTP approach is tested and compared with OPLS for a three‐component mixture analyzed by infrared spectroscopy and a multicomponent mixture measured by near infrared spectroscopy in a reactor. Copyright © 2008 John Wiley & Sons, Ltd.     

基于质子转移反应-飞行时间质谱快速鉴别不同产地闽北水仙茶

采用质子转移反应-飞行时间质谱仪(PTR-TOF-MS), 构建了3个产地(武夷山、建阳、建瓯)113个闽北水仙茶样品香气的化学指纹图谱, 对所得的闽北水仙茶香气指纹图谱进行主成分分析(PCA), 获得了不同产地闽北水仙茶样品的质谱信息特征, 然后采用软独立建模分类法(SIMCA)、K最邻近结点算法(KNN)、偏最小二乘判别分析法(PLS-DA)对闽北水仙茶的质谱信息进行了模式识别.结果表明, PTR-TOF-MS结合分类识别模式能有效区分不同产地的闽北水仙茶.PCA 提取了3个主成分, 累计贡献率为84.66%;3个识别模型的校正集判别正确率分别为89.38%、100.00%和100.00%, 预测集的判别正确率分别为83.18%、 96.46%和95.57%.基于此成功建立了不同产地的闽北水仙茶识别模型.本方法无需样品预处理、分析速度快、灵敏度高、对茶叶无损伤, 为茶叶产地溯源提供了新方法.     

基于超高效液相色谱-四极杆飞行时间质谱的非靶向代谢组学用于不同来源单花蜜的差异分析

不同的蜜源植物具有结构多样的次生代谢产物。该研究以8种不同蜜源单花蜜(洋槐蜜、枣花蜜、荆条蜜、椴树蜜、荞麦蜜、麦卢卡蜜、枸杞蜜、益母草蜜)为研究对象,建立了基于超高效液相色谱-四极杆飞行时间质谱技术(UPLC-Q-TOF-MS^E)的非靶向代谢组学方法,考察了不同蜜源中次生代谢产物的差异。该研究采用固相萃取前处理方法和UPLC-Q-TOF-MS^E方法,获得不同蜜源单花蜜的植物代谢组信息,并构建了多变量统计分析模型,对不同来源的单花蜜进行模式识别和差异分析,发现洋槐蜜、枣花蜜、荆条蜜、椴树蜜、荞麦蜜、麦卢卡蜜相互间存在不同程度的显著差异。结合模型的变量重要性投影、方差分析与最大差异倍数值,根据精确前体离子和碎片离子质量信息检索Chemspider、HMDB数据库,该研究筛选并鉴定出32个代谢差异化合物,其中黄酮类化合物18个、酚酸类化合物7个、苯苷与萜苷类化合物6个、甾体类化合物1个;研究发现麦卢卡蜜和荞麦蜜以黄酮类化合物为主要差异代谢物,荆条蜜中酚酸类化合物为特征性表达,苯苷与萜苷类化合物主要为椴树蜜的特征代谢物。该研究从植物代谢组学角度初步揭示了不同单花蜜的代谢产物差异性以及特征化合物,为基于化学分析技术的蜂蜜溯源识别与质量评价提供了有效的研究策略。