云南野生食用牛肝菌多光谱信息融合质量评价体系研究

Wild-grown boletes are the important featured mushroom resources of Yunnan Province. Quality control and safety of bolete mushrooms is an important premise to guarantee the health of consumers and promote the related industry to develop deeply. So, it is necessary to establish a rapid, comprehensive and systematic technique in order to assess the quality of these resources as well as to provide technical support of quality control. Multi-spectral information fusion which is based on the substance system and functional group could reflect the difference of the substances from the chemical essence and realize the complementation of multi-source information. In addition, this method could also reflect the qualitative characteristics of samples more comprehensively and synthetically. Based on collecting a total of 20 species, 1349 fruiting bodies of boletes from 61 populations of Yunnan Province, this project intends to combine the modern spectral technology with multivariate statistics analysis and aims to (1) acquire the UV and FT-IR specra of bolete mushrooms, perform the chemical pattern recognition and screen the characteristic spectra information, (2) determine the uniformity of the recognition results, analyze the relevance and difference of the characteristic peaks of the two spectra and screen the characteristic peaks which have significant difference with each other, (3) fuse the screened characteristic spectra information and the peak areas of characteristic peaks which are significant different with each other, establish the quality evaluation system based on the multi-spectral information fusion and determine the influences of species and geographical origins to the quality. The results could provide a new method for the quality control of wild-grown edible mushrooms. Meanwhile, it could also provide the scientific basis for the development and utilization on the regional characteristic wild-grown edible mushroom resources.

野生牛肝菌是云南重要的特色食用菌资源，其质量安全可控是保障消费者健康以及产业深度发展的基础。有必要建立快速、全面、系统的质量评价方法，为野生食用牛肝菌质量监管提供技术支撑。多光谱信息融合基于物质体系及官能团信息，从化学本质上反映物质的不同，实现多源信息互补，能够更全面、综合地反映样品质量特征。本项目在已采集云南地区20个物种、61个居群共1349个野生食用牛肝菌子实体的基础上，采用现代光谱技术，结合多元统计分析方法：（1）采集牛肝菌UV和FT-IR光谱，进行模式识别，筛选特征光谱信息；（2）判断两种光谱识别结果的一致性，分析光谱特征峰的关联性与差异性，筛选具有差异性的特征峰；（3）拟合筛选的特征光谱信息以及具有差异性的特征峰峰面积，建立多光谱信息融合质量评价体系，探明物种和产地对其质量的影响，为野生食用菌质量控制提供一种新思路，同时为区域特色野生食用菌资源合理开发利用提供科学依据。

野生食用牛肝菌质量与其自身所含的化学成分密切相关，组成成分复杂，易受地域、气候、物种来源等因素影响，导致不同种和地区的子实体品质差异较大，仅通过指标性成分评价其质量，忽略多成分的协同作用，难以反映整体信息。本项目以云南野生牛肝菌子实体为研究对象，采用光谱技术结合数理统计分析，研究发现：紫外光谱274nm、285nm和296nm附近为样品共有峰，同一产地样品的化学成分类型相近；数据融合策略利用多维数据间的协同优势，较单一决策模型更具优异的分类性能；UV-Vis结合PLS-DA能够准确鉴别不同产地牛肝菌，部位信息融合技术具有较强的鉴别优势；牛肝菌的原始红外光谱较相似，共有峰主要归属为蛋白质、多糖、纤维素和氨基酸等物质中O-H、C=O、C-O-H、C=O、C-C等官能团的吸收峰；3300cm-1附近强吸收峰归属为蛋白质、多糖、纤维素等O-H伸缩振动或蛋白质的N-H伸缩振动；2931cm-1附近特征吸收峰主要由多糖、蛋白质的甲基对称伸缩振动引起；1634cm-1附近吸收峰为C=O伸缩振动，归属为蛋白质酰胺I带；1480cm-1附近代表亚甲基的弯曲振动；1400、1319、1253cm-1等附近为多糖、蛋白质等的C-O-H弯曲振动和亚甲基的变形振动；1078、1057cm-1附近分别为糖类的C-O和C-C伸缩振动；950-710cm-1范围有多个弱吸收峰，主要为糖类异构体的特征峰；红外光谱预处理与特征变量筛选结合，构建的模型分类性能显著增强；NIR和RI光谱特征变量结合高级数据融合策略可有效鉴别不同产地和种类牛肝菌；此外，建立了ICP-AES、UV-Vis和FT-MIR特征融合结合化学计量学对牛肝菌产地溯源的新方法。通过项目的实施，在国内外刊物发表研究论文30篇，其中SCI、EI收录25篇。培养硕士研究生15人，获硕士学位8人；晋升高职1人；培养云南省中青年学术和技术带头人1人。