光谱学与光谱分析, 2016, 36 (6): 1726, 网络出版: 2016-12-20   

基于主成分分析和聚类分析的不同产地绒柄牛肝菌红外光谱鉴别研究

Discrimination of Boletus Tomentipes from Different Regions Based on Infrared Spectrum Combined with Principal Component Analysis and Cluster Analysis
作者单位
1 云南农业大学农学与生物技术学院, 云南 昆明 650201
2 云南省农业科学院药用植物研究所, 云南 昆明 650200
3 玉溪师范学院资源环境学院, 云南 玉溪 653100
摘要
采用傅里叶变换红外光谱结合主成分分析和聚类分析建立快速鉴别不同产地绒柄牛肝菌的方法。 采集15个产地绒柄牛肝菌样品的红外光谱信息, 用多元散射校正(multiplicative signal correction, MSC)、 二阶求导(Second derivative, SD)、 Norris平滑的组合方法对原始光谱进行优化处理, MSC+SD+ND(15, 5)预处理后的光谱数据进行主成分分析和聚类分析, 并通过主成分载荷图分析不同产地绒柄牛肝菌样品差异的原因。 结果显示, 该方法的重现性, 精密度及稳定性的RSD值分别为0.17%, 0.08%, 0.27%, 表明方法稳定、 可靠。 主成分分析的前3个主成分累积贡献率达到87.24%, 能表达红外光谱的主要信息, 主成分得分散点图中同一产地样品成簇聚集, 不同产地样品分布于相对独立的空间, 能有效区分不同产地样品。 主成分载荷图显示, 随主成分贡献率降低, 主成分所捕获的样品信息减少, 其中PC1在3 571, 2 958, 1 625, 1 456, 1 405, 1 340, 1 191, 1 143, 1 084, 935, 840, 727 cm-1波数捕获大量样品信息, 归属为糖类、 蛋白质、 氨基酸、 脂肪、 纤维素等化学物质的吸收峰, 表明这些化学物质含量的差异是区分不同产地绒柄牛肝菌样品的主要依据。 基于离差平方和法(Ward method)及欧氏距离(Euclidean distance)进行聚类分析, 能直观显示不同产地样品的分类情况及样品之间的相关性, 15个产地样品基本能够按照产地来源正确聚类, 正确率为93.33%。 傅里叶变换红外光谱结合主成分分析和聚类分析, 可以有效鉴别绒柄牛肝菌产地来源, 并且能够分析不同产地样品具有差异的原因, 为野生食用菌的鉴别分类和应用研究提供可靠依据。
Abstract
With the aim of establishing a rapid method to discriminate Boletus tomentipes samples from different regions, FTIR spectroscopy with the aid of principal component analysis and clustering analysis were used in the present study. The information of infrared spectra of B. tomentipes samples originated from 15 regions has been collected. The original infrared spectra was pretreated by multiplicative signal correction (MSC) in combination with second derivative and Norris smooth. The spectral data were analyzed by principal component analysis and cluster analysis after the optimal pretreatment of MSC+SD+ND (15, 5), and the reasons for the differences of B. tomentipes samples from different regions could be explained through the principal component loading plot. The results showed that, the RSDs of repeatability, accuracy and stability of the method were 0.17%, 0.08% and 0.27%, respectively, which indicated the method was stable and reliable. The cumulative contribution of first three principal components of PCA was 87.24% which could reflect the most information of the samples. Principal component scores scatter plot displaying the samples from same origin could clustered together and samples from different areas distributed in a relatively independent space. Which can distinguish samples collected from different origins, effectively. The loading plot of principal component showed that with the principal component contribution rate decreasing, the captured sample information of principal component was also reducing. In the wave number of 3 571, 2 958, 1 625, 1 456, 1 405, 1 340, 1 191, 1 143, 1 084, 935, 840, 727 cm-1, the first principal component captured a large amount of sample information which attributed to carbohydrates, proteins, amino acids, fat, fiber and other chemical substances. Which showed that the different contents of these chemical substances may be the basis of discrimination of B. tomentipes samples from different origins. Cluster analysis based on ward method and Euclidean distance has shown the classification and correlation among samples. Samples originated from 15 regions could be clustered correctly in accordance with the basic origins and the correct rate was 93.33%. Which can be used to identify and analyze B. tomentipes collected from different sites. Fourier transform infrared spectroscopy combined with principal component analysis and cluster analysis can be effectively used to discriminate origins of B. tomentipes mushrooms and the reasons for the differences of B. tomentipes samples from different regions could be explained. This method could provide a reliable basis for discrimination and application of wild edible mushrooms.

杨天伟, 张霁, 李涛, 王元忠, 刘鸿高. 基于主成分分析和聚类分析的不同产地绒柄牛肝菌红外光谱鉴别研究[J]. 光谱学与光谱分析, 2016, 36(6): 1726. YANG Tian-wei, ZHANG Ji, LI Tao, WANG Yuan-zhong, LIU Hong-gao. Discrimination of Boletus Tomentipes from Different Regions Based on Infrared Spectrum Combined with Principal Component Analysis and Cluster Analysis[J]. Spectroscopy and Spectral Analysis, 2016, 36(6): 1726.

本文已被 1 篇论文引用
被引统计数据来源于中国光学期刊网
引用该论文: TXT   |   EndNote

相关论文

加载中...

关于本站 Cookie 的使用提示

中国光学期刊网使用基于 cookie 的技术来更好地为您提供各项服务,点击此处了解我们的隐私策略。 如您需继续使用本网站,请您授权我们使用本地 cookie 来保存部分信息。
全站搜索
您最值得信赖的光电行业旗舰网络服务平台!