建立红外光谱快速鉴别牛肝菌种类及预测牛肝菌中重金属镉（Cd）含量的方法。 采集11种牛肝菌共98个子实体的红外光谱信息, 解析牛肝菌的红外光谱, 用一阶导数、 标准正态变量和多元散射校正对原始光谱进行预处理, 通过PLS-DA鉴别牛肝菌种类。 采用ICP-AES法测定牛肝菌中有毒重金属Cd含量, 分析牛肝菌对Cd的富集规律并与GB 2762—2012规定的食用菌中Cd限量标准比较, 评价牛肝菌的食用安全性。 以食用菌对重金属Cd的富集机理为切入点, 将牛肝菌红外光谱数据和Cd含量数据进行拟合, 用PLS模型快速预测牛肝菌的Cd含量。 结果显示: （1）牛肝菌红外光谱经过适当的预处理进行PLS-DA, 前三个主成分累积贡献率达到79.3%, PLS-DA的三维得分图能明显区分不同种类牛肝菌; （2）不同产地、 种类牛肝菌对重金属Cd的富集存在差异, 其含量在0.05～23.41 mg·kg-1 dw之间, 除了采自昆明五华区的灰疣柄牛肝菌外, 多数样品的Cd含量超过GB2762－2012的限量标准, 食用有一定的健康风险; （3）牛肝菌红外光谱数据与Cd含量拟合后进行正交信号校正-小波压缩优化处理, 用PLS模型预测牛肝菌的Cd含量; 训练集和验证集的R2分别为0.851 9和0.882 4, RMSEE和RMSEP分别为2.59和2.67, 大部分牛肝菌的Cd含量预测值与真实值较接近, 表明PLS模型可用于牛肝菌Cd含量快速预测。 傅里叶变换红外光谱结合化学计量学能实现牛肝菌种类快速鉴别及Cd含量准确预测, 为牛肝菌种类鉴别和质量控制提供快速、 有效的方法。

Fourier transform infrared (FTIR) spectroscopy was used to establish a rapid method for the identification of species of bolete mushrooms and content prediction of cadmium (Cd). The information of infrared spectra based on 98 fruiting bodies of 11 species of bolete mushrooms were collected and analyzed. The original infrared spectra were optimized by first derivative, standard normal variate (SNV) and multiplicative signal correction (MSC), and then identification of different species of bolete mushrooms was performed by partial least squares discriminant analysis (PLS-DA). The Cd contents were determined by inductively coupled plasma emission spectrometer (ICP-AES) and the accumulation regularity for Cd was analyzed in order to evaluate the food safety of boletes according to Chinese national food safety standards and limit contaminants in food (GB 2762—2012). Based on the accumulation mechanism of Cd in edible mushrooms, the infrared spectral and Cd content data of tested samples were integrated and PLS model was used to rapidly predict the Cd content in boletes. The results showed that: (1) The spectral data were analyzed by PLS-DA after appropriate pretreatments, and the cumulative contribution rate of the first three principal components was 79.3% as well as all the samples could be correctly classified according to the species in the three-dimensional score plot. (2) There were obvious differences in the accumulation of Cd contents in different samples and the Cd contents in the bolete mushrooms were ranged from 0.05 to 23.41 mg·kg-1 dw. In addition, there were some health risks for eating the mushrooms because Cd contents in most samples were higher than the standard GB 2762—2012 except the mushrooms collected from Wuhua district in Kunming. (3) The integrated data of the infrared spectra and Cd content were optimized by orthogonal signal correction-wavelet compression (OSCW) and the prediction of Cd content in boletes was performed by PLS model. The R2 of the training set and validation set were 0.851 9 and 0.882 4, respectively, while RMSEE and RMSEP were 2.59 and 2.67, respectively. The predictive values of Cd content in most boletes were approximate to the measured values which indicated that the model could be used for rapid prediction of Cd content in boletes. FTIR combined with chemometrics could be proposed to rapidly discriminate the species of bolete mushrooms and predict Cd content accurately. This study can provide a rapid and effective method for quality control and identification of wild-grown bolete mushrooms.