Analysis of the Rule of TCM Compatibility in TCM Prescriptions Containing Ginseng Radix ET Rhizoma in Ancient Books for Xiaoke Bing

Background TCM considers that diabetes belongs to the scope of Xiaoke Bing. Compound prescriptions are characteristics of TCM. For a certain medicine, its compatibility with different medicines can exert different efficacies in different prescriptions. Using the TCM compound prescriptions containing Ginseng Radix ET Rhizoma in ancient books for Xiaoke Bing as an example, this study introduces new methods to investigate the rule of TCM compatibility. Methods Frequency analysis was accomplished by programs written in Perl. The R, Cytoscape, and DpClus software were used to carry out the association rules analysis, the construction of the TCM interconnection network, and the graph clustering analysis, respectively. Results Frequency analysis ranked the frequencies of medicine, medicinal flavors, properties, and meridian attributions, and it was found that some of them are significantly higher than others. Six association rules were obtained. The TCM interconnection network showed that there are close medicine associations among prescriptions, and we got 17 categories of closely related prescriptions from the network. Conclusions Ginseng Radix ET Rhizoma was widely used in treating Xiaoke Bing. Our results are consistent with the understanding of Xiaoke Bing in TCM; hence, it is demonstrated that the methods are effective for exploring the rule of TCM usage in prescriptions. This analysis could provide references for the treatment of diabetes.


Introduction
Because of the advantages and characteristics of traditional Chinese medicine (TCM), TCM compound prescriptions have been widely used for thousands of years.
ere are many factors controlling the efficacies of medicines, among which the most important factor is the performance of the compatible medicine. e compatibilities of a medicine with different medicines can induce differential effects. erefore, it is necessary to further study TCM compound prescriptions, especially for prescriptions that have been used for a long time, and to summarize the rule of TCM usage across different prescriptions. Such analysis may be helpful for enhancing the accuracy of treatments with clinical medications.
Panax ginseng Meyer (hereinafter referred to as P. ginseng as a plant and Ginseng Radix ET Rhizoma as a medicine in TCM) is one of the most valuable medicinal plants and is widely used in East Asia and North America, especially in Korea, China, and Japan [1]. P. ginseng has been used in TCM for more than 2000 years. P. ginseng and its components have been shown to exhibit a variety of pharmacological activities [2][3][4] through different mechanisms and pathways in vitro, in vivo, and in clinical models and have therapeutic effects [5][6][7][8][9][10] in a variety of human diseases. In recent years, with an increased focus on ginseng chemical components, the hypoglycemic activity of ginseng and its active ingredients has been gradually recognized by researchers. Niu et al. [11,12] studied the effects of ginsenoside 20(S)-Rg3 and ginseng polysaccharide in type-2 diabetic rats from the perspective of metabolomics by rapid liquid chromatography-mass spectrometry. Jiao et al. [13] studied the changes of antihyperglycemic activity of ginseng pectin induced by chemical and heat treatment, and Liu et al. [14] studied the therapeutic effects of malonyl ginsenosides on type-2 diabetic rats induced by high-fat diet and streptozotocin. e inspiration of these studies originates mainly from the treatment of Xiaoke Bing by P. ginseng in ancient books.
Xiaoke Bing in TCM is characterized by excessive drinking, eating, polyuria, emaciation or turbidity of urine, and sweetness of urine. us, TCM considers that diabetes belongs to the scope of Xiaoke Bing. TCM has a long history of knowledge of Xiaoke Bing. As early as the second century B.C., Huangdi Neijing, the earliest TCM masterpiece in China, pioneered the understanding of Xiaoke Bing. With the development of later generations, a complete diagnosis and treatment system has been formed [15].
ere are a large number of TCM compound prescriptions containing ginseng (while as a Chinese medicine, it is called Ginseng Radix ET Rhizoma) to treat Xiaoke Bing in ancient books. Collecting and sorting out these prescriptions and elucidating their rule of TCM usage across different prescriptions may be of clinical significance for further guiding the treatment of modern diabetes.

Materials and Methods
e prescriptions studied in this paper originate from ancient books of TCM, before the Republic of China included them in the Chinese Medical Code (Fifth Edition). First, we retrieved the Chinese Medical Code with 35 keywords of different kinds and names of ginseng that appeared at different periods and constructed a database for retrieving the results. Second, the database was searched with 21 keywords representing different types and symptoms of Xiaoke Bing that were summarized from the ancient books of TCM. Finally, a clinician filtered the prescriptions to determine the final data. e inclusion criteria of prescriptions were as follows: a clear prescription composition, dosage, and clinical indications; the main prescriptions had to contain Ginseng Radix ET Rhizoma; and the original prescription for clinical addition and subtraction had to contain Ginseng Radix ET Rhizoma.

Data Preprocessing.
Because different books and different historical periods have different appellations for the same medicine, the medicine names of constituent prescriptions have been unified according to the Chinese Pharmacopoeia and the Great Dictionary of Traditional Chinese Medicine. If a medicine was not included in these two books, the medicine name was unified according to provincial and ministerial local standards. TCMs processed by different methods were treated as different medicines. e dosage of each medicine was not considered in the following analysis.

Frequency Analysis.
In the present study, the computer programming language, Perl [16], was used to program the frequency analysis of all prescriptions, single medicines, and medicine combinations, as well as medicinal flavors, properties, and meridian tropisms.

Association Rules Analysis.
e association rules of medicines in prescriptions were analyzed by using the arules package in the open-source software, R [17][18][19]. e resulting association rules were in the form of X ⟶ Y, in which X represents LHS items and Y represents RHS items. Each association rule involved two indicators. e support degree denoted the probability that LHS and RHS were present in one prescription. e confidence level denoted the probability of the appearance of RHS in a prescription on the basis of LHS appearing in the same prescription. ese two indicators reflected the drug compatibility tendency in statistics.

TCM Interconnection Network Construction.
e TCM interconnection network consisted of nodes and edges that represented the following: nodes represented medicines, the size of which represented the frequency of medicine occurrence in all prescriptions; the edge connecting two medicines represented that both medicines appeared in one prescription; and the width of the edge represented the frequency of this appearance. e final TCM interconnection network was displayed by Cytoscape software [20].

Graph Clustering Analysis.
To identify the densely connected node groups in the TCM interconnection network (i.e., to identify highly overlapping prescription groups and to analyze the subtle differences of prescription medication), we used DPClus [21] software to perform graph clustering analysis. DPClus tends to isolate densely connected regions of a graph as clusters.

Prescriptions Overview.
In total, 303 prescriptions were ultimately obtained, and the authors and works of more than 10 corresponding prescriptions are shown in Table 1.
e number of repetitions of prescriptions may partly explain their wide applications. e most repeated prescription consisted of Rubi Fructus, Coptidis Rhizoma, Galli Gigerii Endothelium Corneum, Ginseng Radix ET Rhizoma, Cnidii Fructus, Dendrobii Caulis, Rehmanniae Radix Praeparata, Trichosanthis Radix, Scrophulariae Radix, Poria, and Dioscoreae Spongiosae Rhizoma. is prescription has been named Bai Fuling Pill and has been mainly used to treat Xiaoke Bing of the kidney type, which is manifested as the legs gradually thinning and the waist and feet becoming weak. e second-most repeated prescription was the prescription composed of Ginseng Radix ET Rhizoma and Trichosanthis Radix, and its main use had been for treating thirst and excessive drinking. e distribution of the number of medicines in prescriptions is shown in Figure 1. e average number of medicines per prescription was 10. Most prescriptions contained 9-12 medicines.

Frequency Analysis.
ere were a total of 244 medicines involved in all of the prescriptions. Aside from Ginseng Radix ET Rhizoma appearing in all prescriptions, the topfive most frequently used medicines were in the following order: Ophiopogonis Radix, Glycyrrhizae Radix ET Rhizoma, Poria, Trichosanthis Radix, and Anemarrhenae Rhizoma. Medicines with frequencies exceeding 25% are shown in Table 2.
e frequency distribution of TCM flavors is shown in Figure 2(a), in which sweet, pungent, and bitter were the dominant flavors. e frequency distribution of TCM properties is shown in Figure 2(b), in which warm, cold, and calm were the dominant properties. e frequency distribution of herbal meridian tropisms is shown in Figure 2(c), in which the lungs, spleen, kidney, stomach, and liver meridians were the most dominant. ere are 4 (1.6%) herbs not included in the Chinese Pharmacopoeia and the Great Dictionary of Traditional Chinese Medicine, so their flavors, properties, and meridian tropisms are not counted, and we label them as group "unknown" in Figure 2.
e following four groups of three-medicine combinations each had a frequency of more

Association Rules Analysis.
Six association rules were obtained after setting the support degree to be greater than 0.1 and the confidence level to be greater than 0.6, as shown in Table 3. Taking rule Poria and Puerariae Lobatae Radix ⟶ Glycyrrhizae Radix Et Rhizoma as an example, its support degree was 0.1, which denotes that the probability of three medicines appearing simultaneously in a TCM compound prescription was 10% and that its confidence degree was 1. is indicates that when Poria and Puerariae Lobatae

Evidence-Based Complementary and Alternative Medicine
Radix are included in a TCM compound prescription, there is a 100% possibility of Glycyrrhizae Radix ET Rhizoma also appearing in the same prescription. Moreover, association rule analysis may help to identify the commonly used medicine compatibilities. And from the 6 rules, we can find that the Glycyrrhizae Radix Et Rhizoma, Puerariae Lobatae Radix, and Poria are the most commonly used medicine compatibilities.

Construction of TCM Interconnection Network. Since
Ginseng Radix ET Rhizoma appeared in every prescription, all of its associations were removed in the construction of the TCM interconnection network. e final network consisted of 243 nodes and 3,552 edges, as shown in Figure 3. As can be seen from Figure 3, the nodes in the middle part of the network were closely related and were loosely surrounded.
is pattern indicates that there are some prescriptions that mostly contain the same medicines. Identifying these prescription groups and elucidating their rules and subtle differences may help in providing more accurate application of TCM.

Graph Clustering Analysis.
Using the parameters d min � 0.9, cp min � 0.7, and minimum cluster size � 10, the graph cluster analysis extracted 24 clusters from the TCM interconnection network. Except for 7 ineffective clusters corresponding to only one TCM compound prescription, the other 17 clusters of the prescription were highly coincident in the inclusion of medicines. ese nine prescriptions all contained Ginseng Radix ET Rhizoma, Puerariae Lobatae Radix, Gypsum Fibrosum, Scutellariae Radix, Anemarrhenae Rhizoma, and Adenophora Trachelioides Maxim. In addition to these common medicines, the subtle differences of medicines among these prescriptions are shown in Figure 4.
In summary, the related prescriptions of a cluster had the following characteristics. First, they contained most of the same medicines, which represented the same main efficacies of the prescriptions. e small number of different medicines suggests that the symptoms treated by these prescriptions belong to a specific type of disease or due to the similar etiology and/or pathogenesis, or the doctors at each time period may have adjusted and added or subtracted the prescriptions for treating different clinical manifestations of the same disease syndrome, or simply because doctors have different medication habits for the same disease. is analysis is of great significance for studying the development and change of the same disease over different periods and the rule of medicinal usage for the same disease in different patients or different stages of development. Second, the prescriptions appearing in different ancient books with the same name may represent the same prescription in the process of inheritance in different periods. is is of great significance to the study of the historical evolution and traceability of a prescription.

Discussion
For the high-frequency medicines that we identified, their efficacies for treating primary symptoms were in accordance with the symptoms of Shangxiao and Xiaxiao of TCM; their meridian tropisms conformed to the understanding of the location and classification of Xiaoke Bing in TCM; their flavors and properties were in line with the understanding of the etiology and pathogenesis of Xiaoke Bing in TCM. e disease location of Xiaoke Bing is closely related to the lung, spleen, stomach, and kidney, and according to these locations, TCM has divided Xiaoke Bing into the following three categories: Shangxiao, Zhongxiao, and Xiaxiao. e etiology and pathogenesis consist of impairment of body fluid due to lung heat, excessiveness of stomach fire, and uncontrolling of the losses or overflow of Qi, blood, and body fluid in the body due to kidney

Conclusions
Our results showed that P. ginseng has been widely used in treating Xiaoke Bing, and there were different medicine combinations for different types of syndromes. e research analysis and results demonstrate that the methods and software introduced in this present study may effectively analyze the rule of usage of TCM medications within prescriptions.  Data Availability e data used to support the findings of this study are available from the corresponding author upon request.

Conflicts of Interest
e authors declare that they have no conflicts of interest.