近红外光谱技术应用于云芝蛋白含量的定量分析

DETERMINATION THE CONTENT OF PROTEIN IN CORIOLUS VERSICOLOR USING NEAR INFRARED SPECTROSCOPY

应用近红外光谱结合偏最小二乘法(PLS)建立测定真菌云芝中蛋白含量定量分析模型。所建立的模型经过i选择最有效的光谱预处理方法,光谱区域和最适主因子数使模型最优化。实验结果表明:采用傅里叶变换在1330nm～1725nm光谱区域,主因子数为6,建立的模型最优。模型校正集的交互验证均方根误差(RMSECV)为0.010,交互验证所得校正集样品中的蛋白含量的预测值与真实值间相关系数(Rv)为0.968;应用此模型对预测集样品中的蛋白质含量进行预测,得到预测均方根误差(RMSEP)为0.009,预测集的相关系数(Rp)为0.990。

Partial least squares （PLS） method was applied to establish models with near infrared spectroscopy （NIRS） for quantitative analysis of protein content in Coriolus Versicolor. Via selecting the most efficacious spectra pretreating method, spectra regions and suitable factors, the optimum model was obtained. In the optimum model, the regression coefficient （Rv） between the actual values and predicted values obtained by cross-validation method was 0.968, the root mean square error of cross-calibration （RMSECV） was 0.010, and the root mean square error of prediction （RMSEP） was 0.009. The regression coefficient （Rp） between actual and predicted values in predictive set is 0.990.