Quantitative Models Of Components Of Eucommia Ulmoides Oliver Leaf And Identification Of Its Varieties And Seasons Based On Near-Infrared Spectroscopy

Eucommia ulmoides Oliver leaves are frequently used in the food,pharmaceutical,and feed industries and are high in nutrients and active compounds.The development of a quick,non-destructive Eucommia leaf analysis method based on green,environmentally friendly,effective,precise,and inexpensive near-infrared technology can raise the technical level of the Eucommia industry.At the moment,determining the active ingredients of Eucommia leaves requires a lot of time and is a complicated process.In this study,a total of 175 near-infrared spectra of Eucommia leaves were collected from 6 varieties in 3 seasons of the same origin,and the quantitative analysis models of crude protein,total flavonoids and total sugars of Eucommia were established and optimized by spectral preprocessing,eigenband optimization and different modeling methods,and then decision trees,random forests,BP neural networks and support vector machines（SVMs）were applied to classify and predict spectral information to identify the season and variety source of Eucommia leaves,and finally polyphenol metabolome technology was used to try to reveal the differences in active components of different varieties of Eucommia leaves.The main findings are as follows:（1）A total of 175 crude protein contents of Eucommia leaves were measured in six varieties of Qinzhong No.1,No.2,No.3,No.4,Huazhong No.5 and No.8 in spring,summer and autumn by Kjeldahl method,with a total of 175 crude protein content of Eucommia leaves in autumn（w/w）,16.73%-24.40%in spring and 9.25%-13.98%in summer;The total sugar content of Eucommia leaves was 8.21%-21.35%in autumn,5.99%-20.72%in spring and13.39%-24.49%in summer.The total flavonoid content of Eucommia leaves was 5.91%-12.61%in autumn,4.06%-8.47%in spring and 7.33%-14.24%in summer.There was no significant difference in the contents of the three substances of different cultivars,and the harvest season had a significant effect on the content of the three nutrients and active substances（p<0.001）.（2）Based on the near-infrared spectroscopy data and chemical measured values,through data optimization,preferred band and comparison of modeling methods,a quantitative model of crude protein,total sugar and total flavonoids of Eucommia leaf was established.After spectral smoothing,baseline correction and normalization preprocessing of spectral data,the modeling bands 5203.27-4751.98 cm^-1 and 7000.69-6499.26 cm^-1 were selected,and the crude protein quantitative model was established by PCR method,with the correlation coefficient R²=0.9897,after smoothing,baseline correction and normalization pretreatment,the genetic algorithm selected the modeling band,and the PLS method established a quantitative model of total flavonoids,with the correlation coefficient R²=0.9571.Normalized pretreatment,genetic algorithm selection modeling band,random forest algorithm to establish a quantitative model of total sugar,correlation coefficient R²=0.8913.The verification results showed that the established quantitative model could accurately determine the crude protein（relative error less than 0.1%）,total sugar（relative error less than 0.3%）,and total flavonoid content（relative error less than 0.2%）.（3）Based on near-infrared spectroscopy,the season and variety source of Eucommia leaf were identified by classification methods such as SMICA,decision tree,random forest,BP neural network and SVM.The seasonal recognition model of Eucommia leaf established by all methods can accurately identify the samples of Eucommia leaf in three seasons.The variety classification model was established by SIMCA,decision tree,random forest,BP neural network and SVM,and it was found that the SVM model could identify 57%of the Eucommia leaf samples,and the classification models of Qin Zhong and Huazhong series established by SVM method could identify 88.57%of the Eucommia leaf samples.（4）Using polyphenol metabolome technology,38 differential metabolites were screened among 6 varieties,including 26 flavonoids,7 phenolic acids,3 lignans and coumarins,and 2tannins.Qin Zhong No.1 had high content of 8 flavonols,1 lignan,1 flavonoid and 1 phenolic acid;Qin Zhong No.2 had higher content of 2 flavonols,and Qin Zhong No.3 and No.4 had higher content of 1 flavonoid and 5 flavonols.Huazhong No.5 and No.8 had higher phenolic acid content of 2 and 3 phenolic acids,respectively.PCA and cluster analysis showed that there were significant differences in the metabolite composition of Eucommia leaf metabolites in the Qinzhong and Huazhong series.In this paper,a near-infrared spectroscopy method was established that can quickly determine the crude protein,total sugar,and total flavonoid content of Eucommia leaf,as well as identify the seasonal source of Eucommia leaf,the source of Qin Zhong and Huazhong series varieties,which can be used for rapid detection of Eucommia leaf quality,which is of great significance to the industrial development of Eucommia leaf.