Study on quality grade evaluation of Andrographis Herba based on principal component clustering and PLS regression
10.7501/j.issn.0253-2670.2019.13.029
- Author:
Dan-Dan CUI
1
Author Information
1. School of Chinese Medicine, Guangdong Pharmaceutical University
- Publication Type:Journal Article
- Keywords:
Andrographis Herba;
Grade standard;
PLS regression analysis;
PLS-DA analysis;
Principal component clustering analysis
- From:
Chinese Traditional and Herbal Drugs
2019;50(13):3200-3206
- CountryChina
- Language:Chinese
-
Abstract:
Objective: To combine macroscopical characteristic indices and chemical indices of Andrographis Herba to evaluate its quality grade. Methods: Both macroscopical characteristic indices and chemical indices (the content of four active diterpenoids and the content of ethanol-soluble extractives) of different batches of Andrographis Herba were determined. The macroscopical characteristic indices were encoded using the method of numerical taxonomy, and the content of four active diterpenoids were determined by HPLC. To screen out the appropriate indices for classification, the correlational analyses were conducted between encoded macroscopical characteristic indices and chemical indices. The quality grade was made by principal component clustering analysis according these evaluation indices, and then was analyzed through partial least squares discriminant analysis (PLS-DA). Furthermore, a partial least squares (PLS) regression was constructed for the quality grade prediction of Andrographis Herba. Results: It showed that the samples could be divided into three grades according to the principal component clustering analysis, and was reasonable evaluating by PLS-DA. The PLS regression model for quality grade of Andrographis Herba was constructed as follows: grade Y=3.761-0.020×the leaf content-0.388×the content of andrographolide-1.117×the content of neoandrographolide-0.274×the content of deoxyandrographolide-0.287×the content of 14-deoxy-11,12-didehydro-andrographolide-0.302×the content of four active diterpenoids-0.104×the content of ethanol-soluble extractives-0.015×the color of stem-0.008 4×the color of leaf-0.003×the diameter of base part of stem+0.020×the number of branch+0.137×the diameter of the upper stem+0.011×plant height, if Y=0.7-1.3, the predicted quality was grade A, if Y=1.7-2.3, then B grade, and if Y=2.7-3.3, C grade or qualified product. Conclusion: The model of grade evaluation we constructed using principal component clustering analysis combing with PLS regression analysis performed well, which was applicable in evaluating the quality grade of Andrographis Herba and other traditional Chinese medicines. It also provided a new strategy for study on grade standards of traditional Chinese medicines.
