Application of PCA Method in Traditional Chinese Medicine and Natural Medicine Analysis

The extraction of important and organized information from the complex information system of these massive compounds and the joint application based on chemistry, mathematics and computer science are major problems that have to be faced in the research of pharmaceuticals. Multivariate analysis technology is a great tool to solve these problems. Among many multivariate analysis methods, principal component analysis (PCA) is the most basic method of drug analysis, and it is one of the most commonly used analytical methods for pharmaceuticals.

For complex traditional Chinese medicine and natural medicine systems, it is very difficult to directly classify or sample a large number of samples with complex data variables. Based on PCA's effective simplified data function, projecting the samples in multidimensional space into the two-dimensional or the three-dimensional scoring space allows for a more intuitive view of the aggregation of the sample spatial distribution while retaining most of the information of the sample, which plays an important role in the classification of the sample and the characteristic study of the category.

In the determination and analysis of the main bile acid components of the 13 batches of bear bile powder collected from the market¹, Shiyan et al found that the XD4 sample can be determined by the package indication, content determination results, and near-infrared and infrared spectra. Suspicious samples, see Figure 1A, XD4 is also away from other bear bile powder samples. If XD4 is excluded as an abnormal sample, the PCA score distribution results of the remaining 12 batches of bear bile powder are shown in Figure 1B. It can be seen that the PCA analysis results after XD4 sample removal are more reliable, which can reveal the true relationship between sample data and is more suitable for in-depth study of sample data.

Fig.1 Score plots of PCA for samples of bear bile powder

In the search for iconic chemical characteristics of traditional Chinese medicines and natural medicines, the purpose of using multivariate analysis methods is to discriminate between different types of samples. However, the data used for multivariate analysis often has more variables, and not every variable is helpful to distinguish the sample categories that are performed. Therefore, it is necessary to find the decisive variables, that is, the iconic chemical characteristics, from a large number of variables. For example, there are multiple batches of artificial bezoar samples produced by a number of different manufacturers ². Shiyan et al used the PCA method to analyze the results of the determination of the main bile acids in these samples. The distribution of artificial bezoar samples in DB and SF 2 is obvious. The PCA load map can be used to identify the iconic chemical characteristics of DB and SF 2 artificial bezoar samples, namely cholic acid (CA) and hyodeoxycholic acid (HDCA), as shown in Figure 2.

Fig.2 Plots of PCA for samples of artificial bezoar

Pharmaceuticals are complex systems with rich chemical components. With the rapid development of analytical instruments, more and more relevant test data will be obtained through PCA and other chemometrics technologies combined with computer technology in the analysis and research work. The application of technology can help researchers grasp these experimental data more clearly and accurately, and enable researchers to understand the clues and laws hidden behind the complicated data, so there is reason to believe that chemometrics technologies such as PCA plays an increasingly important role in pharmaceutical research³.

References

• SHI Yan, et al., (2016) ‘Quantification of the Major Bile Acids in Bear Bile Powder by HPLC-ELSD.’ Chinese Pharmaceutical Journal, 51(22):1958-1961.

• SHI Yan, et al., (2018) ‘Multi-component quantitation of artificial cow-bezoar and study on its markers of quality difference.’ China Journal of Chinese Materia Medica, 43(04): 659-664.

• SHI Yan, WEI Feng MA Shuang-cheng, (2018) ‘Discussion on application of principal component analysis in analysis of traditional Chinese medicines and natural drugs.’ China Journal of Chinese Materia Medica, 43(14): 3031-3035.

Alfa Chemistry is a professional analytical testing company, and has long been committed to promoting the long-term development of the analytical testing industry. We sincerely welcome colleagues from all walks of life to come to negotiate and cooperate.