Herb Co-Occurrence Network Analysis of Formulas and Herbs in Systematic Differentiation of Warm Diseases

Abstract

Aim: Based on the method of complex network analysis, this study explores the medication patterns in the systematic differentiation of warm diseases (Wen Bing Tiao Bian). Method: From the book Wen Bing Tiao Bian, extract prescriptions and establish a database. After normalizing the data, develop software using Python to perform statistical analysis on medication frequency, drug relationships, and core herbs. Results: The total drugs used in the Three Jiao are 250. The drugs of the top 10 node degrees are as follows: ginseng, white peony root, Poria cocos, licorice, radix rehmanniae praeparata, almond, Evodia rutaecarpa, Danggui, fennel, Scutellaria baicalensis. The drugs can be divided into 5 clusters. The total drugs used in the upper warmer are 154. The drugs with high frequencies are licorice, almond, ginseng, fennel, clove and musk, radix rehmanniae praeparata, Danggui, Sichuan pepper, Evodia rutaecarpa, Ophiopogon japonicus, rosewood heart wood, faeces trogopterorum, ginger etc. The total drugs used in the middle warmer are 133. The drugs of the top 10 frequencies are as follows: licorice, Scutellaria baicalensis, Magnolia officinalis, Poria cocos, ginseng, dried orange peel, Coptis chinensis, Pinellia ternata, ginger, radix scrophulariae, talc, gypsum. The total drugs used in the Xia Jiao are 157. The drugs of the top 10 frequent are: poria cocos, ginseng, white peony root, radix rehmanniae praeparata, licorice, Danggui, ass hide glue, cassia twig, Evodia rutaecarpa, chicken egg yolk. This study is meaningful for further analysis of the drug compatibility characteristics and syndrome-differentiation-based medication patterns in Wen Bing Tiao Bian. Conclusions: Complex network methods can analyze the medication characteristics and relationship characteristics of Wen Bing Tiao Bian, which is meaningful for further analyzing the medication compatibility characteristics and syndrome differentiation medication rules in Wen Bing Tiao Bian.

Share and Cite:

Lu, W.J. (2026) Herb Co-Occurrence Network Analysis of Formulas and Herbs in Systematic Differentiation of Warm Diseases. Journal of Biosciences and Medicines, 14, 345-370. doi: 10.4236/jbm.2026.1411025.

1. Introduction

The Wen Bing Tiao Bian, a comprehensive systematic differentiation of warm diseases by Wu Jutong of the Qing dynasty, established the “Three Burners Differentiation” system and systematically consolidated the clinical experiences of physicians such as Ye Tianshi. Its pharmacological framework is centered on the core principle of “preserving Yin fluids and protecting Yang Qi” [1]. In recent years, under the backdrop of integrating classical theories with modern scientific and technological advancements, the research paradigm has shifted from traditional textual collation and empirical summation toward multidimensional explorations including mechanism elucidation, clinical evidence evaluation, and dosage form innovation.

Zhang et al. [2] proposed a three-dimensional network analysis framework for Wen Bing prescriptions of “pathogenesis-target-pathway”. They also discussed the application and potential role of network pharmacology in the Chinese herbal medicine. The developing tendency of network pharmacology is also summarized.

Yang et al. [3] organized 281 cases of Wu Jutong treating warm diseases. A total of 281 prescriptions were involved, from which 133 Chinese herbal medicines were extracted, with a total frequency of 2,625 occurrences. There were 24 high-frequency herbs, among which the top 5 most frequently used were Ophiopogon japonicus, Rehmannia glutinosa, Forsythia suspensa, Lonicera japonica, and Scutellaria baicalensis. The nature of the herbs was primarily cold, warm, and neutral; the flavors were predominantly bitter, sweet, and pungent; and the meridian tropisms mainly belonged to the Lung, Stomach, Heart, and Spleen meridians. Based on Ophiopogon japonicus, Rehmannia glutinosa, Forsythia suspensa, and Lonicera japonica, the core herbal formula was formed.

Sun et al. [4] studied the medication rules of damp-heat diseases in Wen Bing Tiao Bian selecting a total of 87 prescriptions and involving 167 Chinese herbal medicines with a cumulative frequency of 569 occurrences. Among the Three Jiao patterns, warm diseases characterized by damp-heat in the Middle Jiao were predominant; specifically, within summer-heat warm disease, latent summer-heat, and damp-warm disease, damp-warm disease was the most common. The frequently used herb natures and flavors were primarily bitter and warm. The meridian tropisms mainly targeted the Spleen, Lung, and Stomach meridians. Commonly used herbs included Pinellia ternata, Poria cocos, ginseng, Magnolia officinalis, Scutellaria baicalensis, Prunus armeniaca, and Coptis chinensis. Frequent association combinations identified were: ginseng and Coptis chinensis leading to Zingiber officinale and Pinellia ternata; Aurantium aurantium leading to Coptis chinensis; and Pinellia ternata with Scutellaria baicalensis leading to Coptis chinensis. Principal component analysis revealed that clinically common herbs for treating damp-heat warm diseases include heat-clearing and dampness-drying herbs (Scutellaria baicalensis, Coptis chinensis), Qi-moving and dampness-drying herbs (Magnolia officinalis, Citrus reticulata, Aurantii Fructus, Pinellia ternata, Prunus armeniaca), and bland-permeating dampness-leaching herbs (Poria cocos, talci, Tetrapanax papyrifer, Coix lacryma-jobi).

Wang et al. [5] studied the medication rules for damp-warm disease in Wen Bing Tiao Bian. They selected 6 prescriptions and 22 Chinese herbal medicines for the Upper Jiao, 45 prescriptions and 81 herbs for the Middle Jiao, and 21 prescriptions and 72 herbs for the Lower Jiao. The results showed that in the Upper Jiao, heat-clearing herbs, diuretic dampness-leaching herbs, and phlegm-resolving, cough-suppressing and asthma-relieving herbs accounted for a cumulative frequency of 73.33%. In the Middle Jiao, diuretic dampness-leaching herbs, heat-clearing herbs, dampness-transforming herbs, and Qi-regulating herbs reached a cumulative frequency of 55.32%. In the Lower Jiao, tonifying deficiency herbs, interior-warming herbs, diuretic dampness-leaching herbs, and heat-clearing herbs comprised a cumulative frequency of 62.17%.

Liu et al. [6] used the Traditional Chinese Medicine Inheritance Computing Platform (TCMICS), the frequency statistics, property-taste-meridian analysis, herbal efficacy evaluation, clustering analysis, and association rule mining were performed on the herbal data extracted from the Wen Bing Tiao Bian. A total of 43 target prescriptions were identified. The predominant herbal properties were warm, followed by cold and neutral. The most common tastes were sweet, followed by bitter and pungent. The meridians most frequently targeted were the Spleen, Lung, Heart, and Kidney. The most frequent herbal efficacy category was tonifying deficiency, followed by astringing and clearing heat. The four core herbs most frequently combined with Ginseng, in descending order of frequency, were Ophiopogon japonicus, processed glycyrrhiza, Danggui, and donkey-hide gelatin. Liao et al. [7] screened the active components and related targets of Yinqiao by traditional Chinese medicine system pharmacology database and analysis platform (TCMSP). Then the disease-related targets of URTI, Gene ontology (GO) enrichment analysis and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis were carried. They got the main active ingredients such as (-)-phillygenin, luteolin, quercetin, kaempferol and the key targets involved IL-6, TP53, MAPK1, APP, VEGFA, TNF and found 121 KEGG related pathways.

Tang et al. [8] employs network pharmacology and molecular docking to investigate the shared molecular mechanisms underlying Yin Qiao San’s application in treating different diseases with the same method—namely, acute upper respiratory infection in children (AURIC) and viral myocarditis in children (VMC). The primary findings reveal that Yin Qiao San exerts its therapeutic effects on AURIC and VMC primarily through key targets including STAT3, RELA, TNF, MAPK3, and AKT1, mediated by major bioactive components such as quercetin, kaempferol, luteolin, naringenin, and stigmasterol. These components modulate multiple signaling pathways, including the cancer-related pathways, PI3K/AKT signaling, TNF signaling, AGE/RAGE signaling, IL-17 signaling, and TLR-like receptor signaling pathways.

Tan et al. [9] retrieved the active ingredients and corresponding targets of the medicinal components in the Jianpi Yiqi Formula from the Traditional Chinese Medicine Systems Pharmacology Database and Analysis Platform (TCMSP). A protein-protein interaction (PPI) network, GO enrichment analysis, KEGG enrichment analysis and molecular docking were performed. They found that the Jianpi Yiqi Formula may exert a potential therapeutic effect on pediatric SS via core targets including STAT3, AKT1, and MAPK3.

Wang et al. [10] analyzed the medicinal herb usage in Shanghan Lun and Wen Bing Tiao Bian to investigate the inheritance relationship between the two classical works. The study revealed that among the 110 formulas in Shanghan Lun and 200 formulas in Wen Bing Tiao Bian, a total of 255 distinct herbal substances were involved (77 from Shanghan Lun and 251 from Wen Bing Tiao Bian). Among herbs used with a frequency of ≥30, in addition to commonly used herbs such as licorice, ginger, and jujuba, the top herbs in Shanghan Lun were cassia twig and white Paeonia lactiflora; in Wen Bing Tiao Bian, they were ginseng, poria cocos, white Paeonia lactiflora, Ophiopogon japonicus, Pinellia ternata, and Rehmannia glutinosa. When combining the herb usage across both texts, the ranked list of herbs with frequency ≥30 was: white Paeonia lactiflora, cassia twig, ginseng, Pinellia ternata, poria cocos, dried ginger, Scutellaria baicalensis, Aconitum carmichaelii, Coptis chinensis, Ophiopogon japonicus, Magnolia officinalis, Prunus armeniaca, Rheum palmatum, and Rehmannia glutinosa.

Zhang et al. [11] conducted statistical analysis on the drugs in the Chapter of Upper Energizer of Systematic Differentiation of Warm Diseases. The number, use frequency, frequencies of four properties and five flavors, texture of different parts of the drugs were analyzed. There are 58 items and 46 formulas, involving 11 diseases names of wind warm, warm heat, pestilence, warm toxin, winter warm, summer warm, latent summer heat, damp warm, cold damp, warm malaria and autumn dryness. Tang et al. [12] explored the core drug combinations and compatibility patterns of pungent drugs in Wenbing Tiaobian for treating warm pathogen diseases and systematically collected formulas containing pungent drugs from Wenbing Tiaobian. Using the TCMICS 3.5, frequency statistics, property, taste and meridian tropism analysis and association rule mining on extracted herbal data were conducted to identify core drug combinations. They found that the pungent herbs treat Upper Jiao syndromes by exerting the effects of expelling pathogens through outward penetration and clearing heat, act on Middle Jiao syndromes via the mechanisms of pungent-opening and bitter-downbearing as well as dampness resolution, and manage Lower Jiao syndromes with the combined actions of purgative downward conduction and tonification.

Through analysis of Wen Bing Tiao Bian, Wen et al. [13] identified 72 formulas in the “Xia Jiao Chapter” involving 162 distinct medicinal substances. For Warm-Heat type warm diseases, the high-frequency herbs are: licorice, white peony root, phiopogon root, prepared rehmannia root, donkey-hide gelatin, oyster shell, hemp seed, turtle carapace, tortoise plastron, ginseng. These primarily comprise Yin and blood nourishing agents. For Damp-Heat type warm diseases, the high-frequency herbs are: ginseng, poria cocos, white peony root, prepared aconite root, coptis rhizome, dried ginger, cassia twig, pinellia rhizome, Danggui, smoked plum, etc. These primarily comprise Qi supplementing and interior warming agents. For Cold-Damp type disorders, the high-frequency herbs are: prepared aconite root, poria cocos, cassia twig, licorice, Magnolia mark, dried ginger, galangal rhizome, fennel seed, tangerine peel, green tangerine peel, etc. These primarily comprise interior warming and Qi regulating agents.

Although considerable research has been conducted on the patterns of syndrome differentiation and treatment in Shanghan Lun, studies employing complex network analysis to explore medication rules based on the San Jiao differentiation classification in Wen Bing Tiao Bian remain scarce. This study adopts complex network analysis to separately examine subnetworks of prescriptions for the upper, middle, and Xia Jiao, as well as the overall prescription network, aiming to uncover medication patterns, the roles of individual herbs, and their interrelationships in Wen Bing Tiao Bian. This work can provide references for better treatment of warm diseases in the future

2. Material and Method

To investigate the medication patterns in the Wen Bing Tiao Bian, this study performed complex network analysis and clustering analysis on the prescriptions in Wen Bing Tiao Bian. All statistical computations and network analyses were conducted using software written in Python, with visualization of results carried out accordingly. The data used in this study were exclusively derived from relevant classics of the Wen Bing Tiao Bian [1] [14] [15]. All data normalization procedures are performed based on the annotated edition of Wen Bing Tiao Bian documented in Reference [15]. A comprehensive collection and statistical analysis of the 215 screened prescriptions in Wen Bing Tiao Bian, in which the Upper Jiao includes 50 prescriptions, Middle Jiao 93 prescriptions and Lower Jiao 72 prescriptions. During the inclusion process of the 215 prescriptions, only one copy of duplicate prescriptions is retained. Prescriptions with only one single herb modified from the original formula are excluded. All herbs using different medicinal parts of the same plant, as well as raw and processed herbal products, are distinguished as separate medicinal ingredients. For example, fresh ginger and dried ginger are classified as distinct herbs, raw hyacinth bean and hyacinth bean are regarded as the same herb, while Tinglizi (Descurainiae Semen) and Tingli (Drabae Herba) are treated as two separate herbs. Fried Qu and fried Shenqu are considered the same herb, and “Zhuyu” refers specifically to Evodiae Fructus (Wu Zhuyu). Prepared licorice root and raw licorice root are collectively referred to as licorice. Schizonepeta herb and carbonized schizonepeta herb are referred to as schizonepeta. Pear peel and pear flesh are collectively referred to as pear.

3. Complex Network Analysis

3.1. Introduction of Method

As an integrated multidisciplinary concept, network pharmacology, which is based on system biology and polypharmacology, affords a novel network mode of multiple targets, multiple effects, complex diseases and replaces the magic bullets by magic shotguns. Chinese herbal medicine (CHM) has been recognized as one of the most important strategies in complementary and alternative medicine. Though CHM has been practiced for a very long time, its effectiveness and beneficial contribution to public health has not been fully recognized. Also, the knowledge on the mechanisms of CHM formulas is scarce.

Complex network analysis is a perspective and method for studying complex systems, which focuses on the structure of individual interactions within the system. By calculating the topological features such as average degree, network diameter, average path length, and average clustering coefficient of nodes in the network, as well as the characteristics of dispersion, concentration, and density between elements, the distribution pattern of elements between things is described [16]. From the perspectives of openness, hierarchy and complexity, traditional Chinese medicine has the same characteristics as complex networks, which is also the basis for using complex networks to analyze traditional Chinese medicine data.

The use of complex network technology to explore the rules of traditional Chinese medicine’s syndrome differentiation and treatment, as well as the formulation of medication, is an effective way to reveal the inherent complex relationships of traditional Chinese medicine [17]. The node degree is one of the criteria for determining the importance (centrality) of traditional Chinese medicine nodes in network analysis. The larger the node degree is, the more times it appears in combination with other nodes is, and the higher the centrality of the node is, making it more important in the network [18]. The betweenness centrality refers to the frequency at which any node(drug) appears on the shortest path of the network, reflecting the degree of hub of the node(drug) in the network connection compatibility [15]. The above two indexes can measure the importance of nodes of traditional Chinese medicine in complex networks. The more frequent the node compatibility, the denser the lines between the combinations, and the greater the node degree, the stronger the compatibility and combination ability of traditional Chinese medicine nodes. In this paper, each node represents a single standardized Chinese herbal medicine entity, meaning one distinct herb corresponds to exactly one node. The node degree is defined as the number of edges directly connected to the corresponding node, which represents the total number of direct associated objects of that entity. For every co-occurrence of two Chinese herbal medicines in the same prescription, the corresponding edge weight is incremented by 1. The final edge weight is defined as the total cumulative co-occurrence count of the two herbs across all 215 prescriptions. The network adopts undirected edges, and no additional weight filtering threshold is set. During the clustering analysis, the implementation from the network library in Python 3.8.5 was used.

3.2. The Upper Jiao

3.2.1. The Degrees of Nodes

The total nodes are 154, The total edges are 1823 (Figure 1).

The drugs of the top 10 of the node degree are licorice 79, almond 76, ginseng 70, fennel, clove and musk 63, radix rehmanniae praeparata, Danggui and sichuan pepper 57, Evodia rutaecarpa 56, Ophiopogon japonicus 51, rosewood heart wood and faeces trogopterorum 50, ginger 49, white peony root 44. The bottom 10 are: Magnolia officinalis and red peony 13, Szechwan chinaberry fruit, folium Isatidis, radix bupleuri and loquat leaf 12, Tetrapanax papyriferus, lotus leaf, bark of luffa, sevenlobed yam rhizome, turtle shell, pear skin, puffball, desert cistanche, Chinese wolfberry, pilose antler, pearl, carborundum and watermelon peel 11, benzoin, silver foil, peucedani radix, fructus aurantii, carapax eretmochelydis, amber, buf-falo horn and Perilla frutescens 10, radix isatidis, bombyx batryticatus and Salvia miltiorrhiza 9, red bean, donkey-hide gelatin, Sesamum orientale, Belamcanda chinensis, Canella winterana, cardamonfruit and chrysanthemum 7, Itoa orientalis, radices trichosanthis, raw lentil beans, Polygonatum odoratum and bulb of thun-berg fritillary 6, seed-coat of green gram, semen benincasae, lotus seed and reed stem 5, water chestnut juice, lotus root juice, pear juice and Chinese mosla herb 4, muskmelon base 2.

Figure 1. Network for the prescription of Upper Jiao.

The edge weights over 1 are shown in Table 1. The rest edge weights are 1. It means that the edge weights are either 2 or 1.

Table 1. The edge weights of the Upper Jiao network.

SDH-MD

2

SHJ-GUIZ

2

ZY-LQ

2

SDH-YS

2

JG-LG

2

ZY-XNJ

2

GC-JG

2

JG-LQ

2

ZY-DSN

2

XR-JG

2

ZY-JYH

2

NBZ-MAB

2

MD-ZY

2

3.2.2. Cluster Analysis of the Upper Jiao

For the clustering analysis, the K-means implementation from the sklearn.cluster library in Python 3.8.5 is adopted. The edge weights (i.e., the co-occurrence counts of herbal medicines) are incorporated into the clustering calculation. A random seed of 42 is set to ensure result reproducibility. All disconnected components in the network are included in the unified clustering workflow. Nodes are allowed to belong to multiple clustering modules based on their co-occurrence relationships, and the optimal clustering partition is finally selected using the modularity index.

It can be divided into 6 clusters. Acorus tatarinowii, betel nut, lindera root, Amomum tsao-ko, monkshood, dried ginger, Allium macrostemon, Asarum, cassia twig, radix bupleuri, toosendan Fructus are the first cluster; Chinese wolfberry, desert cistanche, turtle shell, pilose antler, carborundum are the second cluster. Benzoin, ox horn, Eretmochelys imbricata, amber, silver foil are the third cluster. Radix scrophulariae, dried orange peel, Tetrapanax papyriferus, Salvia miltiorrhiza, Magnolia officinalis, gypsum, licorice, talc, Anemarrhena asphodeloides, japonica rice, loquat leaf, spica schizonepetae, black bean sauce, hyacinth bean flower, moutan cortex, folium isatidis, paeoniae radix rubra, white amomun pericarp, donkey-hide gelatin, seed of flax are the fourth cluster. Yu Jin, pearl, Gardenia jasminoides, ox horn, hawksbill turtle, silver foil, amber are the fifth cluster. The rest form the sixth cluster.

3.3. Middle Jiao

3.3.1. The Degrees of Nodes of the Middle Jiao Network

The total nodes are 133. The edges are 1092. The drugs of the top 10 node degrees are: licorice 62, Scutellaria baicalensis 53, Magnolia officinalis 48, poria cocos 47, ginseng 44, dried orange peel 43, Coptis chinensis 42, Pinellia ternata 42, ginger 41, radix scrophulariae 41, talc 40, gypsum 38 (Figure 2).

The drugs of the bottom 10 node degrees are: ash bark 10, japonica rice 10, caulis bambusae in taeniam 10, papaya 9, myristicae semen 9, peucedani radix 9, Szechuan lovage rhizome 9, sea cucumber 9, pericarpium trichosanthis 9, Angelica pubescens 9, Atractylodes lancea 9, notopterygium root 9, silkworm excrement 8, red bean 8, orange red, lotus seed 8, Forsythia suspensa 8, Evodia rutaecarpa 7, divaricate saposhnikovia root 7, Tribulus terrestris 7, semen raphani 7, soybean

Figure 2. Network for the prescription of Middle Jiao.

roll 7, Sichuan pepper 7, Dalbergia odorifera 6, Croton tiglium 5, polygoni multiflori radix 5, Trichosanthes kirilowii 5, citri Reticulatae pericarpium viride 5, turtle carapace 5, Polygonatum odoratum 5, red bean 5, Pulsatilla 5, red peony 5, donkey-hide gelatin 5, hecaoya 5, Lygodium japonicum 5, chicken gizzard membrane 5, seed of hemp 5, adenophora root 5, oyster 5, Artemisia annua 5, mulberry leaves 5, betel nut 4, rock sugar 4, fistular onion stalk 4, suis fellis pulvis 4, Achyranthes bidentata Blume 4, halloysitum rubrum 3, common ginger 3, orange peel 3, persimmon calyx 3, salt 2, milk 1, white sugar 1 (Table 2).

Table 2. Node degree of Middle Jiao.

GC

62

MDP

26

FUZ

16

BPZ

12

CXD

8

JINJ

5

HQC

53

CAOG

25

HUAF

16

JBP

12

JUHN

8

MARN

5

HPU

48

HSS

25

ZHIS

16

XHS

12

LPZ

8

SHAS

5

FULN

47

SHM

25

LG

16

ZZZ

12

LQ

8

MULI

5

RS

44

YIRN

25

CZHU

16

RNH

12

ZHUY

7

QNHO

5

CHP

43

JG

25

GUIZ

15

WUME

11

FAGF

7

SY

5

HL

42

ZY

25

CHX

15

DAZ

11

BJLI

7

BALA

4

BANX

42

ZHS

25

DIX

15

JINZ

11

LFZI

7

BINT

4

SHJ

41

RSX

25

PUX

15

HQI

11

DADJ

7

CONB

4

YS

41

MUX

24

XIS

15

NJJ

11

CUNJ

7

ZUDZ

4

HSI

40

BAIZ

24

CIS

15

QINP

10

JIAX

6

NIUQ

4

SG

38

DAH

24

LYJ

15

JM

10

BADO

5

CSZ

3

ZHUL

35

GANJ

23

DC

14

ZURU

10

HESW

5

PAOJ

3

ZZI

35

YJX

23

ZHIQ

14

MUGW

9

GUAL

5

JUP

3

XR

34

CAIH

22

MANX

14

ROUG

9

QIGP

5

SHID

3

BAIS

33

FAGJ

22

GUYA

13

QINH

9

BIEJ

5

SHIY

2

LQ

32

DAGG

21

SENQ

13

CHAX

9

YUZU

5

NIUR

1

SDH

32

ZEX

21

MAIY

12

HAIS

9

CHID

5

BAIT

1

MD

32

DAFP

21

BH

12

GLPI

9

BTWN

5

BTC

31

HUGB

20

DQY

12

DUHO

9

CSY

5

XNJ

31

HUOX

19

JJS

12

MAOS

9

EJIO

5

ZM

30

YINC

18

NBZ

12

QAGH

9

SLY

5

JYH

28

SANZ

17

BKRN

12

CANS

8

HJSA

5

3.3.2. The Edge Weights of the Middle Jiao Network

Poria cocos (FULN)-Tetrapanax papyriferus (BTC) 4, Scutellaria baicalensis (HQC)-ginger (SHJ) 2, bamboo juice (ZHUL)-Tetrapanax papyriferus (BTC) 2. Table 3 shows the edge weights equal or larger than 2. The rest are all 1.

Table 3. The edge weights more than 1.

FULN-BTC

4

HQC-SHJ

2

ZHUL-BTC

2

GC-JYH

2

GANJ-CUNJ

2

SHJ-CAOG

2

BTC-BKRN

2

HPU-BTC

2

GANJ-FUZ

2

SHJ-ZHIS

2

BTC-HSI

2

HPU-FULN

2

RS-FUZ

2

SHJ-ZZI

2

BTC-LQ

2

FULN-HSI

2

RS-SHJ

2

ZM-HUAF

2

BTC-SG

2

FULN-MAIY

2

BAIS-HL

2

CAOG-FULN

2

BTC-XR

2

FULN-ZHUL

2

BAIS-HQC

2

FAGJ-BTC

2

LG-JYH

2

YS-MD

2

DAGG-SHJ

2

FAGJ-FULN

2

SDH-MD

2

ZY-MD

2

GC-SHJ

2

FULN-DAFP

2

YS-JYH

2

MD-JYH

2

3.3.3. Cluster Analysis of the Middle Jiao Network

It can be divided into 5 clusters.

Radix bupleuri, Tetrapanax papyriferus, bamboo juice, almond, Pinellia ternata, poria cocos, Pericarpium arecae, ginger, rhizoma alismatis, ginseng, atractlodis rhizoma, Amomum tsao-ko, white cardamom kernel, coix Seed, cassia twig, hawthorn, agastache, medicated leaven, malt, silkworm excrement, red bean, Tribulus terrestris, semen raphani are the same cluster.

Clove, sodium nitrate, magnet, rhinoceros horn, eaglewood, antelope’s horn, talc, licorice, aucklandiae radix, white peony root are the same cluster.

Rehmanniae radix praeparataa, Ophiopogon japonicus, radix scrophulariae, reed root, honeysuckle flower, bamboo leaf, golden juice, moutan cortex, Trichosanthes kirilowii pollen, folium isatidis, peppermint, burdock fruit, spica schizonepetae, Forsythia suspensa, lotus seed are the same cluster.

Milk (NIUR)-white sugar (BAIT) are connected separately and are not drawn.

3.4. Lower Jiao

3.4.1. The Connection Degree of Nodes

The total nodes are 157. The edges are 1556. The drugs of the top 10 node degrees are as follows: poria cocos 82, ginseng 76, white peony root 75, radix rehmanniae praeparata 61, licorice 56, Danggui 50, ass hide glue 48, cassia twig 47, Evodia rutaecarpa 46, chicken egg yolk 45 (Figure 3).

The drugs of the bottom 10 node degrees are as follows: Chinese bulbul 3, almond 3, tabanus 3, medical leech 3, rice vinegar 2, euphorbia pekinensis 2, white mustard seed 2, euphorbia kansui 2, asparagus fern 2, Yu Yuliang 2, milk 1, white sugar 1 (Table 4).

Table 4. The connection degree of nodes of Lower Jiao.

FULN

82

NIUQ

35

CUNJ

21

ZHUL

10

LIUH

5

RS

76

HQC

35

HUGB

21

JG

9

ZY

5

BAIS

75

DUZO

33

ZUGS

20

RDKO

9

YINC

5

SDH

61

FUZ

32

HSS

19

SBD

9

YITN

5

GC

56

ZISY

32

SHJ

18

SHAR

9

YJSI

5

DAGG

50

LUJO

32

XNJ

18

MAHW

9

BINT

4

EJIO

48

BIJI

31

QIGP

17

ZM

8

YUZU

4

GUIZ

47

GUIY

31

PAOJ

15

JIUZ

8

BQZ

4

HUIX

47

LNGU

31

CHX

15

CAOG

8

LG

4

ZHUY

46

SPXO

31

DIX

15

HULB

8

OZI

4

JZH

45

WUZG

31

PUX

15

QINP

7

CANS

4

BANX

43

XYSN

31

RSX

15

BALA

7

ZOJZ

4

BIEJ

43

DAH

31

SHM

15

BADO

7

CSZ

4

LPZ

43

HPU

30

YS

15

CLZI

7

JM

4

XIS

42

MDP

30

ZHS

15

DIYU

7

DANC

4

GANJ

42

TAOR

29

CIS

15

WUYO

7

TONB

4

BUGZ

41

MUX

28

HSI

15

YGPI

7

LZI

4

GUIJ

41

CAIH

28

LYJ

15

GAOL

6

QNHO

4

TUSZ

41

BAIZ

25

XIXN

15

ZXHT

6

SHAS

4

QINS

40

HL

24

DAZ

14

SUZS

6

BTWN

3

MULI

39

TLZI

23

YIRN

14

XFHA

6

XR

3

LURO

39

CHP

23

MAOS

14

XIAF

6

MANC

3

RCRO

38

FONW

22

SANY

14

HUOX

5

SUIZ

3

WWZ

38

LXHA

22

ZHIS

12

SHUQ

5

MICU

2

BAIM

37

QAGL

22

LIAJ

12

MANX

5

DAJI

2

GOQI

37

QUMA

22

JYH

12

RNZB

5

BJZI

2

BJLI

37

SHIW

22

SANZ

12

ZHEL

5

GANS

2

YYZI

37

TUBC

22

MARN

11

MUGW

5

TIND

2

BAOY

37

WUME

22

BAJT

11

JUHN

5

YUYL

2

HAIS

37

SEG

22

FUPZ

11

WULZ

5

NIUR

1

WUGJ

37

SUFU

22

ROUG

10

JUP

5

BAIT

1

MD

35

SG

22

3.4.2. The Edge Weights of the Lower Jiao Network

The edge weights are as follows: white peony root - ass hide glue 9, white peony root-Ophiopogon japonicus 8, white peony root - birthplace 8, ass hide glue - phiopogon japonicus 7, ass hide glue raw rehmannia 7, ophiopogon japonicus - birthplace 7, white peony root - licorice 6, white peony root - dried ginger 6, ass hide glue - licorice 6, white peony root - seed of hemp 5 (Table 5).

Table 5. The edge weights of drug network in Lower Jiao.

BAIS-EJIO

9

GUIZ-SHJ

5

HUIX-LURO

4

BIEJ-MULI

3

DUZO-HUIX

3

BAIS-MD

8

MARN-SDH

5

MARN-MD

4

BIEJ-SDH

3

DUZO-LURO

3

BAIS-SDH

8

BAIS-FULN

4

MULI-SDH

4

BIEJ-MARN

3

FULN-LPZ

3

EJIO-MD

7

BAIS-GUIZ

4

TUSZ-LURO

4

BJLI-GUIJ

3

FULN-QINS

3

EJIO-SDH

7

BAIS-HL

4

GOQI-GUIJ

3

BUGZ-LURO

3

FULN-LURO

3

MD-SDH

7

BAIS-HQC

4

BAIS-DAZ

3

CUNJ-GANJ

3

FUZ-BAIZ

3

BAIS-GC

6

DAGG-RS

4

BAIS-MDP

3

CUNJ-HL

3

FUZ-RS

3

BAIS-GANJ

6

DAGG-LURO

4

BAIS-RS

3

DAGG-DUZO

3

LPZ-QINS

3

EJIO-GC

6

DAZ-GC

4

BAIS-SHJ

3

DAGG-FUZ

3

LPZ-RS

3

BAIS-MARN

5

FULN-FUZ

4

BANX-GANJ

3

DAGG-HUIX

3

MDP-DAH

3

EJIO-MARN

5

FUZ-LURO

4

BIEJ-EJIO

3

DAZ-SHJ

3

MDP-SDH

3

FULN-RS

5

GC-SHJ

4

BIEJ-MD

3

DAZ-GUIZ

3

RS-WWZ

3

GC-RS

5

GC-GUIZ

4

3.4.3. The Cluster Analysis of the Lower Jiao Network

It can be divided into 5 clusters. exocarpium citri rubrum, faeces trogopteri, sulfur, toosendan fructus, lindera root, croton tiglium, betel nut, pericarpium citri reticulatae viride, galangal, calcitum, aucklandiae radix, mirabilite, musk, clove, antelope’s horn, eaglewood, cinnabar, cimicifugae rhizoma are the same cluster.

Ailanthus root bark, sanguisorba officinalis, bamboo juice, atractylod, seed of Chinese honeylocust, fenugreek, silkworm excrement, white sugar, yamaimo, raspberry, hawthorn, amur cork-tree, honeysuckle flower, papaya, Coptis chinensis, Ophiopogon japonicus, atractylodes macrocephala, common ginger, asari radix et rhizoma are the same cluster.

Fennel, achyranthes bidentata, cornus officinalis, psoralen, semen cuscutae, cistanches herba, American ginseng, keel, sevenlobed yam rhizome, longan, mantidis ootheca, sheep kidney, sea cucumber, turtle shell, hen egg yolk, pig’s bone marrow, Evodia rutaecarpa, lotus seed are the same cluster.

Dried ginger, scutellaria baicalensis, cassia twig, fructus aurantii immaturus, rheum officinale, peach kernel, Meng Chong, medical leech, saltpeter are the same cluster (Figure 3).

Milk (NIUR)-white sugar (BAIT) and euphorbia kansui (GANS)-white mustard seed (BJZI)-euphorbia pekinensis (DAJI) are connected separately respectively and not drawn in the figure.

Figure 3. Network for the prescription of Lower Jiao.

3.5. The Total of the Three Jiao

3.5.1. The Connection Degree of Each Node

The total nodes are 250. The total edges are 3541.

The drugs of the top 10 node degrees are as follows: ginseng 131, white peony root 115, poria cocos 111, licorice 111, radix rehmanniae praeparata 104, almond 98, Evodia rutaecarpa 95, Danggui 93, fennel 90, Scutellaria baicalensis 82 (Figure 4).

Figure 4. Network for the prescription of all the three Jiao.

The drugs of the bottom 10 node degrees are as follows: rock sugar 4, water chestnut juice 4, lotus root juice 4, seed of Chinese honeylocust 4, onion stalk 4, pig’s bile 4, mussel 4, urine of boys under twelve 4, pear juice 4, Chinese mosla herb 4, kaki calyx 3, NaCl 2, asparagus fern 2, Yu Yuliang, sweet melon stem 2 (Table 6).

Table 6. The node degree of all the three Jiao.

RS

131

XNJ

43

GUIY

31

RNH

16

YUZU

8

BAIS

115

XIS

42

LNGU

31

QINH

15

JIUZ

8

FULN

111

HAIS

42

SPXO

31

CHX

15

HMR

8

GC

111

BUGZ

41

WUZG

31

PUX

15

HULB

8

SDH

104

RCRO

41

XYSN

31

CIS

15

FAGF

7

XR

98

ZHUL

41

BH

30

LYJ

15

LFZI

7

ZHUY

95

TUSZ

41

SEG

30

SHAS

15

BKP

7

DAGG

93

LURO

41

CZHU

30

NJJ

15

DADJ

7

HUIX

90

GOQI

40

ZHS

29

MARN

14

DIYU

7

HQC

82

QINS

40

WUME

29

MUGW

14

JH

7

BANX

78

XIAF

40

LG

29

SBD

14

YGPI

7

MD

76

SUZS

40

SY

29

CXD

14

GAOL

6

BIEJ

75

FAGJ

39

HSS

27

BDH

14

ZXHT

6

DAH

73

ZY

39

SANZ

26

BKRN

14

ZHIP

6

GUIZ

71

ZM

38

BALA

26

SANY

14

QNHO

6

HPU

70

ZZI

38

YJX

25

GUYA

13

XINB

6

SHJ

70

DC

38

NBZ

24

QINP

13

XFHA

6

CUNJ

69

MICU

37

WUYO

24

JUHN

13

HESW

5

GANJ

64

RUX

37

TLZI

23

SENQ

13

GUAL

5

DIX

63

BAIM

37

FONW

22

BADO

12

LDP

5

RSX

63

YYZI

37

LXHA

22

CANS

12

CHID

5

CHP

62

BAOY

37

QAGL

22

MAIY

12

BTWN

5

TAOR

60

MOY

37

QUMA

22

DQY

12

SHUQ

5

YIRN

60

WUGJ

37

SHIW

22

ZZZ

12

CSZ

5

HL

56

XHS

36

TUBC

22

PPY

12

RNZB

5

ROUG

56

DAZ

36

HUOX

22

BAJT

11

ZHEL

5

JIAX

56

AICO

35

SUFU

22

FUPZ

11

DGZ

5

JG

55

YIMU

35

JM

22

HEY

11

SLY

5

HSI

55

AWEI

35

JUP

22

SGP

11

HJSA

5

WULZ

55

BIJI

35

ZHIS

21

LIP

11

JINJ

5

LQ

54

HUSO

35

HUAF

21

MAB

11

LIUH

5

EJIO

53

NIUQ

35

JJS

21

JINZ

11

WEIJ

5

MUX

53

SUIZ

35

DAFP

21

SHLH

11

YITN

5

WWZ

53

QIGP

35

ZUGS

20

XGCY

11

YJSI

5

FUZ

52

ZEX

35

BCQE

20

AXX

10

BINT

4

SG

52

CAOW

35

ZHIQ

20

YBP

10

BQZ

4

LPZ

50

GANQ

35

MAOS

20

DMP

10

OZI

4

MDP

50

JIAH

35

JIUB

20

HBO

10

ZOJZ

4

HUGB

49

MANC

35

SCPU

20

ZURU

10

CONB

4

JYH

49

PUHG

35

YINC

20

SUYE

10

ZUDZ

4

YS

48

SANL

35

HQI

20

RDKO

9

DANC

4

CAOG

47

SUMU

35

CLZI

17

SHAR

9

TONB

4

CAIH

46

ZHH

35

PAOJ

17

BLG

9

LZI

4

JZH

45

SHM

34

CSY

17

JIAC

9

XRU

4

LIAJ

44

DUZO

33

MANX

17

GLPI

9

SHID

3

BJLI

44

BTC

33

BPZ

16

DSN

9

SHIY

2

GUIJ

44

ZISY

32

JBP

16

DUHO

9

TIND

2

CHAX

43

LUJO

32

GEG

16

MAHW

9

YUYL

2

MULI

43

XIXN

31

SHQ

16

QAGH

9

TGD

2

BAIZ

43

3.5.2. The Edge Weights of All the Three Jiao

The edges weights are as follows: birthplace - Ophiopogon japonicus 11, birthplace-white peony root 9, birthplace - ass hide glue 7, dried ginger - white peony root 7, ginseng - licorice 7, white peony root - licorice 7, white peony root - Scutellaria baicalensis 7, ass hide glue - licorice 7, ass hide glue-Ophiopogon japonicus 7, ginseng - Danggui 6 (Table 7).

Table 7. The edge weights over 2.

SDH-MD

11

TUSZ-LURO

4

CUNJ-GUIZ

3

MUX-LYJ

3

SHM-YS

3

SDH-BAIS

9

LQ-ZY

4

CUNJ-HL

3

MUX-CHX

3

BTC-HSI

3

SDH-EJIO

7

BIEJ-EJIO

3

DAGG-DAZ

3

MUX-PUX

3

HSI-CIS

3

GANJ-BAIS

7

BIEJ-MD

3

DAGG-DUZO

3

MUX-RSX

3

HSI-CHX

3

RS-GC

7

BIEJ-MDP

3

DAGG-FUZ

3

MUX-SHM

3

HSI-DIX

3

BAIS-GC

7

BIEJ-MULI

3

DAZ-BAIZ

3

MUX-XIS

3

HSI-LYJ

3

BAIS-HQC

7

BIEJ-MARN

3

DUZO-LURO

3

MUX-YS

3

HSI-PUX

3

EJIO-GC

7

SDH-NIUQ

3

EJIO-HL

3

MUX-SG

3

HSI-RSX

3

EJIO-MD

7

SDH-SG

3

FULN-CAOG

3

MUX-XNJ

3

HSI-XNJ

3

RS-DAGG

6

SDH-SHAS

3

FULN-HSI

3

RCRO-LURO

3

HSI-YS

3

RS-FULN

6

HUIX-DUZO

3

FULN-JG

3

SHJ-JUP

3

HSI-SG

3

BAIS-HL

6

ZHS-CHX

3

FULN-LPZ

3

SHJ-ZHIS

3

ZY-JYH

3

DAZ-GC

6

ZHS-CIS

3

FULN-QINS

3

XIS-CHX

3

JYH-LG

3

SDH-MDP

5

ZHS-DIX

3

FULN-RCRO

3

XIS-CIS

3

JYH-YS

3

SDH-MARN

5

ZHS-GC

3

FULN-ZHUL

3

XIS-DIX

3

SG-CHX

3

HUIX-LURO

5

ZHS-HSI

3

FULN-MAOS

3

XIS-HSI

3

SG-CIS

3

RS-BAIS

5

ZHS-HSS

3

GC-BAIZ

3

XIS-HSS

3

SG-DIX

3

RS-FUZ

5

ZHS-LYJ

3

GC-CHX

3

XIS-LYJ

3

SG-LYJ

3

BAIS-FULN

5

ZHS-MUX

3

GC-CIS

3

XIS-PUX

3

SG-PUX

3

BAIS-GUIZ

5

ZHS-PUX

3

GC-DIX

3

XIS-SG

3

SG-RSX

3

DAGG-LURO

5

ZHS-RSX

3

GC-HSI

3

XIS-SHM

3

SG-XNJ

3

DAZ-SHJ

5

ZHS-SG

3

GC-JYH

3

XIS-RSX

3

SG-YS

3

EJIO-MARN

5

ZHS-SHM

3

GC-MD

3

XIS-XNJ

3

YS-CHX

3

FUZ-BAIZ

5

ZHS-XIS

3

GC-PUX

3

XIS-YS

3

YS-CIS

3

GC-GUIZ

5

ZHS-XNJ

3

GC-RSX

3

XR-SG

3

YS-LYJ

3

MARN-MD

5

ZHS-YS

3

GC-SG

3

XR-BTC

3

YS-DIX

3

BIEJ-SDH

4

GANJ-BANX

3

GC-SHM

3

ZM-SG

3

YS-PUX

3

DAH-MDP

4

GANJ-CUNJ

3

GC-XIS

3

HSS-CHX

3

YS-RSX

3

SDH-MULI

4

GANJ-HL

3

GC-XNJ

3

HSS-CIS

3

DIX-CHX

3

HUIX-DAGG

4

RS-EJIO

3

GC-YS

3

HSS-DIX

3

DIX-CIS

3

GANJ-FUZ

4

RS-MD

3

GC-HSS

3

HSS-HSI

3

DIX-PUX

3

RS-BAIZ

4

RS-CHP

3

GC-LYJ

3

HSS-LYJ

3

DIX-LYJ

3

RS-SHJ

4

RS-DAZ

3

GC-MUX

3

HSS-PUX

3

DIX-XNJ

3

RS-WWZ

4

RS-LPZ

3

GUIJ-LURO

3

HSS-RSX

3

RSX-CHX

3

BAIS-DAGG

4

RS-PAOJ

3

GUIZ-HPU

3

HSS-SHM

3

RSX-CIS

3

BAIS-DAZ

4

GOQI-GUIJ

3

HL-JYH

3

HSS-XNJ

3

RSX-LYJ

3

BANX-GC

4

GOQI-RCRO

3

HPU-BTC

3

HSS-YS

3

RSX-PUX

3

BANX-GUIZ

4

GOQI-LURO

3

HQC-MUX

3

HSS-SG

3

RSX-XNJ

3

BANX-HQC

4

BAIS-BANX

3

HQC-SHJ

3

JG-LG

3

CHX-PUX

3

CHP-FULN

4

BAIS-WUME

3

LPZ-QINS

3

JG-LQ

3

CHX-CIS

3

DAGG-SHJ

4

BAIS-CAIH

3

MD-JYH

3

SHM-CHX

3

CHX-LYJ

3

DAZ-GUIZ

4

BANX-CAIH

3

MD-LG

3

SHM-CIS

3

CHX-XNJ

3

FULN-BTC

4

BJLI-GUIJ

3

MD-SHAS

3

SHM-HSI

3

PUX-CIS

3

FULN-FUZ

4

BUGZ-LURO

3

MD-YS

3

SHM-LYJ

3

PUX-LYJ

3

FULN-LURO

4

CHP-BAIZ

3

MD-SG

3

SHM-SG

3

PUX-XNJ

3

FUZ-LURO

4

CHP-DAZ

3

MUX-CIS

3

SHM-DIX

3

CIS-LYJ

3

GC-JG

4

CHP-GC

3

MUX-DIX

3

SHM-PUX

3

CIS-XNJ

3

HL-HUGB

4

CUNJ-DAGG

3

MUX-HSI

3

SHM-RSX

3

LYJ-XNJ

3

MD-ZY

4

CUNJ-FUZ

3

MUX-HSS

3

SHM-XNJ

3

CSZ-JM

3

SHJ-CAOG

4

3.5.3. The Cluster Analysis of All the Three Jiao

It can be divided into 5 clusters. schisandra chinensis, raw lentil beans, pilose antler, seven lobed yam rhizome, cuttlebone, abalone, keel, yamaimo, pig bone marrow, raspberry, sheep kidney are the same cluster.

Bamboo juice, four stamen stephania root, musk, magnet, calcitum, fructus aurantii immaturus, Magnolia officinalis, Scutellaria baicalensis, honeysuckle flower, talc, gypsum, Forsythia suspensa, radix scrophulariae, cinnabar, gardenia jasminoides, Yu Jin, cimicifugae rhizoma, eaglewood, cow bezoar, gold foil, borneol, pearl, caulis bambusae in taeniam, golden juice are the same cluster.

4. Compared with the Prescription Features of Shang Han Lun

In analysis, 112 prescriptions in Shan Han Lun and 215 prescriptions in Wen Bing Tiao Bian were adopted respectively. The prescriptions in Shang Han Lun were categorized into six types according to the Six Meridians, while those in Wen Bing Tiao Bian are analyzed in three categories based on the Three Jiaos.

The total drugs recorded in Wen Bing Tiao Bian are 250 while 90 in Shang Han Lun. In Shang Han Lun, the appearance frequency of roasted licorice, cassia twig, jujube, ginger, and peony was relatively high. The top 10 used drugs in Wen Bing Tiao bian are as follows: ginseng, white peony root, poria cocos, licorice, radix rehmanniae praeparata, almond, Evodia rutaecarpa, Danggui, fennel, Scutellaria baicalensis [19].

5. Discussion

The three sub-network analyses in Section 3 demonstrate that the Three Jiao classification system in Wen Bing Tiao Bian does not correspond to a single homogeneous repertoire of herbal combinations, but rather three distinct and characteristic co-prescription pairing patterns.

In terms of the relative density against network size, the Upper Jiao network exhibits the tightest node connectivity (154 nodes and 1823 edges in total). Its core herbs are mostly varieties that are light in nature, dispersing in action, aromatic, or specifically directed to the Lung meridian—Glycyrrhizae Radix (79 occurrences), Armeniacae Semen (76), Ginseng Radix (70), Foeniculi Fructus, Caryophylli Flos and Moschus (63), Rehmanniae Radix Praeparataa, Angelicae Sinensis Radix and Zanthoxyli Pericarpium (57), Evodiae Fructus (56), and Ophiopogonis Radix (51). This result fully aligns with Wu Jutong’s therapeutic principle for the Upper Jiao of “releasing the exterior, diffusing the Lung, clearing heat and resolving phlegm”. The Middle Jiao network contains 133 nodes and 1, 092 edges, with its core formulations centred on bitter-cold heat-clearing, Qi-regulating and dampness-transforming herbs (Scutellariae Radix 53, Magnoliae Officinalis Cortex 48, Poria 47, Coptidis Rhizoma 42, Pinelliae Rhizoma 42, Zingiberis Recens Rhizoma 41, Scrophulariae Radix 41, Talcum 40, Gypsum Fibrosum 38), which is consistent with the Middle Jiao’s therapeutic focus on “clearing heat, transforming dampness and harmonizing the centre”. The Lower Jiao network comprises 157 nodes and 1556 edges, where the dominant herbs shift to tonifying, Yin-nourishing and blood-enriching medicinal materials (Poria 82, Ginseng Radix 76, Paeoniae Radix Alba 75, Rehmanniae Radix Praeparataa 61, Glycyrrhizae Radix 56, Angelicae Sinensis Radix 50, Asini Corii Colla 48, Cinnamomi Ramulus 47, Evodiae Fructus 46, and Gallus Ovum Vitellus 45), corresponding to the Lower Jiao therapeutic strategy of “nourishing the Liver and Kidney, restoring the pulse and extinguishing wind”. In the pooled Three-Jiao network (250 nodes, 3541 edges, divided into 5 clusters), Ginseng Radix becomes the connectivity hub (with a degree of 131), which precisely echoes the consistent therapeutic thread of “protecting stomach Qi and preserving body fluids” that runs through all three Jiao layers. Taken together, the above results indicate that Jiao attribution not only predicts the distribution of different core herb categories, but also reflects the tightness of inter-herb associations and the variation characteristics of maximum co-occurrence weights of herb pairs.

At the same time, it should be noted that the analytical results of this study are affected by the following factors:

The first is the interference from formula size. A prescription containing n herbs generates n (n − 1)/2 unordered herb pairs. As a result, both the total number of edges and the weight of each individual edge increase with the number of ingredients in the prescription. An herb that frequently appears in large compound formulas will accumulate a high node degree and edge weight even if its pairing affinity with other herbs is not strong. This effect is clearly observed in our results: in the Upper Jiao sub-network, all edges with a weight greater than 1 have an exact value of 2 (Table 1). The vast majority of the 1823 edges carry only the minimum weight of 1, and only a small number of herb pairs show repeated co-occurrence. In contrast, the weight distribution of the pooled Three-Jiao network drops steeply from 11 to 4, and then extends into a long tail of edges with a weight of 3 (Table 7). Since the clustering step is also performed on these weighted edges, the influence of prescription length further permeates the clustering results (the Upper Jiao network yields 6 clusters, while the other three networks each produce 5 clusters). Therefore, conclusions such as “Poria is the herb with the highest connectivity in the Lower Jiao network” essentially indicate that Poria co-occurs with the largest number of distinct partners, and this outcome is partly a consequence of formula size. Future studies can separate genuine herbal compatibility rules from mere high-frequency co-occurrence phenomena by normalizing weights within each individual prescription (for example, using 1/(n − 1) transformation, or introducing Jaccard coefficient and association strength scaling), setting support or weight thresholds, and comparing the observed network with degree-preserving randomized null models.

The second limitation lies in the coverage scope of the dataset. Following the annotated edition of Wen Bing Tiao Bian [15], this study excluded duplicate records and near-identical prescriptions that differ from an existing formula by only a single herb, and finally retained 215 screened prescriptions: 50 for the Upper Jiao, 93 for the Middle Jiao, and 72 for the Lower Jiao. All three Jiao strata are covered by samples, but the herb inventory is curated according to predefined inclusion rules rather than being exhaustively collected. The final total number of herbs (250 in the pooled network, 154, 133 and 157 for the Upper, Middle and Lower Jiao respectively) is entirely determined by these inclusion criteria. Previous literature has fully demonstrated how such selection decisions affect results: Yang et al. [3] mined 281 medical cases and obtained 133 herbs, Sun et al. [4] analysed 87 prescriptions and identified 167 herbs, Wang et al. [5] included 6/45/21 prescriptions for the three Jiao corresponding to 22/81/72 herbs, Liu et al. [6] [7] incorporated 43 target prescriptions through the TCMICS platform, and Wen et al. [13] collected 72 formulas and 162 ingredients exclusively in the Lower Jiao chapter. The 72-prescription count of our Lower Jiao stratum exactly matches that reported by Wen et al. [13], and the total of 250 herbs in our pooled inventory is highly close to the 251 herbs of Wen Bing Tiao Bian recorded by Wang et al. [10].

Third, the accuracy of the herb inventory fully depends on the rigour of nomenclature standardization. In this study, different medicinal parts, raw products and processed products are listed as separate entries (for example, Zingiberis Recens Rhizoma and Zingiberis Rhizoma are set as independent nodes), while some variants under different aliases are merged (raw Lablab Semen and fried Lablab Semen are combined). “Tinglizi” (Descurainiae Semen) and “Tingli” are split into separate nodes, “Shenqu” and “fried Shenqu” are merged, and the term “Zhuyu” is explicitly restricted to Evodiae Fructus. Every such naming decision will alter the network structure. For instance, separating fresh ginger from dried ginger will distribute the Upper Jiao (49), Middle Jiao (41) and Lower Jiao occurrences of ginger across two nodes, and lower the weight of every pair that contains this herb; merging them, on the contrary, will raise both the node degree and the corresponding edge weights.

Fourth, although the patterns obtained in this study are mutually consistent with previous research, they do not directly prove any pharmacological effect. The core position of Ginseng Radix in the pooled network (degree 131) and its high-frequency co-occurring herbs Ophiopogonis Radix and Glycyrrhizae Radix are highly consistent with the ginseng-centred combinations extracted by Liu et al. [6] (Ophiopogonis Radix, Glycyrrhizae Radix Praeparataa, Angelicae Sinensis Radix and Asini Corii Colla), and also in line with the high ranking of Ginseng Radix, Poria and Paeoniae Radix Alba reported by Wang et al. [10]. The dominance of Coptidis Rhizoma, Scutellariae Radix, Magnoliae Officinalis Cortex and Pinelliae Rhizoma in the Middle Jiao agrees with the damp-heat medication rules summarized by Sun et al. [4], and the light, dispersing and aromatic core herbs in the Upper Jiao also match the herb texture, nature and flavour analysis conducted by Zhang et al. [11]. However, when the dataset or analytical method differs, the “core herbs” reported in previous studies will also vary, which means that the derived regularities are fully dependent on the adopted dataset and analytical approach.

Co-occurrence records document textual associative links in the literature, rather than direct pharmacological interactions. Herbs may be prescribed together to harmonize the overall medicinal property of a formula, correct its unpleasant taste, or moderate its harsh potency, while the current network does not incorporate any information on dosage, preparation methods, action targets or clinical endpoints. Therefore, the so-called “core herbs”, the strongest “compatibility pairs” (Paeoniae Radix Alba - Asini Corii Colla with a weight of 9 and Paeoniae Radix Alba - Ophiopogonis Radix with a weight of 8 in the Lower Jiao sub-network; Rehmanniae Radix Praeparataa - Ophiopogonis Radix with a weight of 11 in the pooled network) and the clustering assignments should all be interpreted as hypotheses about Wu Jutong’s prescribing logic, rather than direct evidence of herb synergy, efficacy or safety. Subsequent analysis needs to carry out three lines of work: first, integrating network pharmacology target prediction and molecular docking technology to analyse candidate herb pairs and their major active components; second, conducting pharmacological experimental validation in cell or animal models for the herb pairs with high topological strength and clinical plausibility (such as the Yin-nourishing Lower Jiao pairs and the Ginseng-Ophiopogon-Glycyrrhiza core combination); third, performing retrospective or randomized clinical evaluation on the formulas that embody the above patterns. The current analysis remains a descriptive, hypothesis-generating map of the formulation regularities in Wen Bing Tiao Bian, but cannot be directly equated to definitive conclusions about pharmacological compatibility or clinical application.

6. Conclusions

This study employs complex network theory to analyze the formula composition patterns, core herbs, inter-herb connectivity, and herb clustering in the systematic differentiation of warm diseases, categorized into four groups: Upper Jiao, Middle Jiao, Lower Jiao, and the overall prescription.

The total nodes are 154 and the total edges are 1823 of the networks formed by the drugs in the upper warmer. The drugs of the top 10 node degrees are as follows: licorice 79, almond 76, ginseng 70, fennel, clove and musk 63, Radix rehmanniae Praeparata, Danggui and Sichuan pepper 57, Evodia rutaecarpa 56, Ophiopogon japonicus 51, rosewood heart wood and faeces trogopterorum 50, ginger 49, white peony root 44. The drugs can be divided into 6 clusters. The edge weights greater than 1 are all 2 (Table 1), and the remaining edges have weight 1.

The total nodes are 133 and the total edges are 1092 of the networks formed by the drugs in the middle warmer. The drugs of the top 10 node degrees are as follows: licorice 62, Scutellaria baicalensis 53, Magnolia officinalis 48, Poria cocos 47, ginseng 44, dried orange peel 43, Coptis chinensis 42, Pinellia ternata 42, ginger 41, Radix Scrophulariae 41, talc 40, gypsum 38.

The largest edge weight is 4 (between Poria cocos and Tetrapanax papyriferus); the other edges with weight ≥ 2 are listed in Table 3, and all remaining edges have weight 1. The drugs in the middle warmer can be divided into 5 clusters.

The total nodes are 157 and the edges are 1556 of the networks formed by the druge in the total drugs in the Xia Jiao. The drugs of the top 10 node degrees are as follows: poria cocos 82, ginseng 76, white peony root 75, radix rehmanniae praeparata 61, licorice 56, Danggui 50, ass hide glue 48, cassia twig 47, Evodia rutaecarpa 46, chicken egg yolk 45.

The total nodes are 250 and the total edges are 3541 of the networks formed by the drugs in the total drugs in the Three Jiao. The drugs of the top 10 node degrees are as follows: ginseng 131, white peony root 115, Poria cocos 111, licorice 111, radix rehmanniae praeparata 104, almond 98, Evodia rutaecarpa 95, Danggui 93, fennel 90, Scutellaria baicalensis 82.

The drugs can be divided into 5 clusters. In the pooled network, the largest edge weights are 11 (birthplace-Ophiopogon japonicus), 9 (birthplace-white peony root) and 7 (birthplace-ass hide glue, dried ginger-white peony root, ginseng-licorice, white peony root-licorice, white peony root-Scutellaria baicalensis, ass hide glue-licorice and ass hide glue-Ophiopogon japonicus), followed by 6 (ginseng-Danggui) (Table 7).

Appendix: Drug Name Abbreviation Reference

Honeysuckle flower

JYH

Cimicifugae rhizoma

SHM

Raw lentil beans

SBD

Fructus aurantii immaturus

ZHIS

Persimmon calyx

SHID

Forsythia suspensa

LQ

clove

DIX

Radices trichosanthis

HUAF

Sesamum orientale

MARN

Seed of Chinese honeylocust

ZOJZ

Peppermint

BH

Sodium nitrate

PUX

Seed coat of green gram

LDP

Oyster

MULI

Agastache

HUOX

Burdock fruit

NBZ

Saltpeter

XIS

Seed of flax

HMR

Chicken egg yolk

JZH

Medicated leaven

MAIY

Spica Schizonepetae

JJS

Radix isatidis

BLG

Donkey hide gelatin

EJIO

Sweet wormwood herb

QNHO

Common ginger

PAOJ

Black bean sauce

DC

Puffball

MAB

Perilla frutescens

SUYE

Urine of boys under twelve

TONB

Soybean roll

DADJ

Platycodon grandiflorum

JG

Dried orange peel

CHP

Peucedani radix

QINH

Mussel

DANC

Hawthorn

SANZ

Licorice

GC

Bombyx batryticatus

JIAC

Fructus aurantii

ZHIQ

Trichosanthes kirilowi

GUAL

Woodlouse

SUFU

Bamboo leaf

ZY

Amur cork tree

HUGB

Evodia rutaecarpa

ZHUY

Halloysitum rubrum

CSZ

Pyrrosia lingua

SHIW

Reed root

LG

Rheum officinale

DAH

Radix bupleuri

CAIH

Rock sugar

BINT

Rice grain sprout

GUYA

Mulberry leaves

SY

Frankincense

RUX

White peony root

BAIS

Achyranthes bidentata Blume

NIUQ

Dianthus superbus

QUMA

Chrysanthemum

JH

Myrrh

MOY

Toosendan fructus

CLZI

Golden juice

JINZ

Chinese trumpet creeper

LXHA

Almond

XR

Astragalus membranaceus

HQI

Fennel

HUIX

Sea cucumber

HAIS

Ground beetle

TUBC

Gypsum

SG

Atractlodis rhizoma

CZHU

Cinnamomum cassia

ROUG

Eupatorium

ZHEL

Honeycomb

FONW

Anemarrhena asphodeloides

ZM

Rhizoma alismatis

ZEX

Radix aconiti kusnezoffii

CAOW

Sediment of human urine

RNZB

Dung beetle

QAGL

Japonica rice

JM

Medicated leaven

SHQ

Turmeric

JIAH

Sorghum

GAOL

Tribulus terrestris

BJLI

Ginseng

RS

Orange peel

JUP

Sichuan pepper

CUNJ

Maltose

MYT

Chicken gizzard-membrane

JINJ

Rehmanniae radix praeparataa

SDH

Atractylodes macrocephala

BAIZ

Tabanus

MANC

White sugar

BAIT

Lygodium japonicum

HJSA

Ophiopogon japonicus

MD

Danggui

DAGG

Rhizoma sparganii

SANL

Milk

NIUR

Radix scrophulariae

YS

Green tangerine peel

QIGP

Cattail pollen

PUHG

Yinchen

YINC

Semen raphani

LFZI

Rhinoceros horn

XNJ

Pueraria lobata

GEG

Saffron

ZHH

Bamboo juice

ZHUL

Polygoni multiflori radix

HESW

Pear juice

LZI

Schisandra chinensis

WWZ

Sappan wood

SUMU

Dark plum

WUME

Antifebrile dichroa branchlet and leaf

SHUQ

Water chestnut juice

BQZ

Ginger

SHJ

Perilla seed frost

SUZS

Caulis bambusae in taeniam

ZURU

Angelica dahurica

BAZI

Lotus root juice

OZI

Jujube

DAZ

Faeces trogopteri

WULZ

Sulfur

LIUH

Yamaimo

SANY

Gardenia jasminoides

ZZI

Chinese mosla herb

XRU

Rosewood heart wood

JIAX

Gypsum

YJSI

Semen euryales

QINS

Sweet melon stem

TGD

Magnolia officinalis

HPU

Dried lacquer

GANQ

Orange red

JUHN

Morinda officinalis

BAJT

Red bean

CXD

Hyacinth bean flower

BDH

Rhizoma cyperi

XIAF

Asparagus fern

TIND

Raspberry

FUPZ

Birthplace

SDH

Watermelon peel

XGCY

Corydalis

HUSO

Inula flower

XFHA

Notopterygium root

QAGH

Salvia miltiorrhiza

DSN

Bark of luffa

SGP

Medical leech

SUIZ

Atractylod

MAOS

Angelica pubescens

DUHO

Coptis chinensis

HL

Lotus leaf

HEY

Ferula asafoetida

AWEI

Polyporus umbellatus

ZhLN

Ailanthus root bark

YGPI

Moutan Cortex

DQY

Pinellia ternata

BANX

Rhizome of chuanxiong

CHAX

Ash bark

QINP

Sanguisorba officinalis

DIYU

Lotus seed

LPZ

Poria cocos

FULN

Galangal

LIAJ

Euphorbia kansui

GANS

Yu Yuliang

YUYL

Cow bezoar

RNH

Coix seed

YIRN

Carbonized mugwort leaf

AICO

Euphorbia pekinensis

DAJI

Myristicae semen

DOUK

Musk

ZZZ

Red peony

CSY

Extract of Leonurus sibiricus

YIMU

White mustard seed

BJZI

Eucommia ulmoides

DUZO

Cinnabar

ZHS

Adenophora root

SHAS

Turtle carapace

BIEJ

Semen cuscutae

TUSZ

Divaricate saposhnikovia root

FAGF

Realgar

XHS

White cardamom kernel

BKRN

Rice vinegar

MICU

Fenugreek

HULB

Chinese bulbul

BTWN

Scutellaria baicalensis

HQC

Tetrapanax papyriferus

BTC

Carborundum

SHLH

Psoralen

BUGZ

Nutmeg

RDKO

Yu Jin

YJX

Burdock fruit

NBZ

Pilose antler

LURO

Allium tuberosum

JIUZ

Amomum villosum

SHAR

Borneol

BPZ

Belamcanda chinensis

SEG

Chinese wolfberry

GOQI

Papaya

MUGW

Black bone chicken

WUGJ

Gold foil

JBP

Loquat leaf

PPY

Desert cistanche

RCRO

Pericarpium arecae

DAFP

Abalone

BAOY

Buffalo horn

NJJ

Reed stem

WEIJ

Sevenlobed yam rhizome

BIJI

Furnace soil

ZXHT

Sheep kidney

YYZI

Amber

HBO

Semen benincasae

DGZ

Turtle shell

GUIJ

Onion stalk

CONB

Pig’s bone marrow

ZUGS

Hawksbill turtle

DMP

Peach kernel

TAOR

Amomum tsao-ko

CAOG

Pig’s bile

ZUDZ

Whitish honey

BAIM

Benzoin

AXX

Cassia twig

GUIZ

Allium macrostemon bunge

JIUB

Ephedra

MAHW

Amethyst

ZISY

Silver foil

YBP

Dried ginger

GANJ

Lindera root

WUYO

NaCl

SHIY

Antler

LUJO

Calcitum

HSS

Monkshood

FUZ

Acorus tatarinowii

SCPU

Semen lepidii

TLZI

Mantidis ootheca

SPXO

Magnet

CIS

White amomun pericarp

BKP

Fourstamen stephania root

FAGJ

Hecaoya

SLY

Cuttlebone

WUZG

Talc

HSI

Pear skin

LIP

Betel nut

BALA

Croton tiglium

BADO

American ginseng

XYSN

Antelope’s horn

LYJ

Bulb of thunberg fritillary

XINB

Cubeb

BCQE

Pericarpium trichosanthis

GLPI

Keel

LNGU

Aucklandiae radix

MUX

Itoa orientalis

ZHIP

Asarum

XIXN

Silkworm excrement

CANS

Longan

GUIY

Eaglewood

CHX

Polygonatum odoratum

YUZU

Mirabilite

MANX

Conflicts of Interest

The author declares no conflicts of interest regarding the publication of this paper.

References

[1] Wu, J.T. (1979) Treatise on Systematic Differentiation of Warm Diseases. People’s Health Publishing House.
[2] Zhang, G.B., Li, Q.Y., Chen, Q.L. and Su, S.B. (2013) Network Pharmacology: A New Approach for Chinese Herbal Medicine Research. Evidence-Based Complementary and Alternative Medicine, 2013, Article ID: 621423.[CrossRef] [PubMed]
[3] Yang, C.C., Lan, W.Y. and Li, S.L. (2025) Exploration of Medication Rules of Wu Jutong in Treating Warm Diseases Based on Data Mining, Journal of Hubei Minzu University (Medical Edition), 42, 26-32. (In Chinese)[CrossRef]
[4] Sun, S.L. and Ma, X.B. (2025) Exploration of Medication Rules for Treating Damp-Heat Warm Diseases in Systematized Identification of Warm Diseases. Journal of Sichuan of Traditional Chinese Medicine, 43, 196-201. (In Chinese)[CrossRef]
[5] Wang, T., Li, C., Jiang, M.Q., Shen, S.Y., Wang, Y., Xia, S.J. and Li, C.D. (2023) Exploration of Medication Rules for Damp-Warm Disease in Systematized Identification of Warm Diseases Based on Association Rules and Latent Structure Models. Journal of LiShizhen Traditional Chinese Medicine, 34, 3065-3069. (In Chinese)[CrossRef]
[6] Liu, D.X., Lin, C.S., Wang, J.J., Sun, X.H., Pang, J., Wang, M.Y., Li, J.Y. and Zhang, F.L. (2024) Study on the Rule of Using Panax Ginseng Prescription to Treat Warm Diseases in Detailed Analysis of Warm Diseases Based on Data Mining and Network Pharmacology. Practical Pharmacy and Clinical Remedies, 27, 815-823. (In Chinese)[CrossRef]
[7] Liao, C., Xia, S.L., Li, X.F. and Dai, R. (2019) Research on Mechanism of Yinqiao Powder in the Treatment of Upper Respiratory Tract Infection Based on Network Pharmacology. Journal of Yunnan University of Traditional Chinese Medicine, 42, 55-61. (In Chinese)[CrossRef]
[8] Tang, W., Xu, Y. and Shang, L.L. (2023) Exploration of the Mechanism of Yinqiao San in Treating Pediatric Acute Upper Respiratory Infection and Pediatric Viral Myocarditis with “Different Diseases, Same Treatment” Strategy Based on Network Pharmacology and Molecular Docking. Journal of Anhui Traditional Chinese Medical College, 42, 83-91. (In Chinese)
[9] Tang, W., Fan, S.J., Liao, T.T., Wang, Y.L. and Shang, L.L. (2026) Exploring the Mechanism of Action of Jianpi Yiqi Formula in the Treatment of Pediatric Short Stature Based on Network Pharmacology and Molecular Docking. Journal of Youjiang Medical University for Nationalities, 48, 278-285. (In Chinese)[CrossRef]
[10] Wang, T., Xia, S.J., Wang, Y., Mi, L.F., Jiang, M.J. and Li, C.D. (2025) Exploration of the Inheritance Relationship of Medication between Systematized Identification of Warm Diseases and Treatise on Cold Damage Based on Data Mining Techniques. Journal of Basic Chinese Medicine, 31, 60-64. (In Chinese)[CrossRef]
[11] Zhang, H.K., Guo, X.X. and Liu, Z.Y. (2017) Drug Use Rule of Warm Disease of the Upper Energizer in Systematic Differentiation of Warm Diseases from the Aspects of Texture, Four Properties and Five Flavors of the Drugs. Henan Traditional Chinese Medicine, 37, 1541-1545. (In Chinese)[CrossRef]
[12] Tang, Y., Chen, X.B., Li, N. and Du, H.G. (2026) Medication Law of Pungent Drugs in Wenbing Tiaobian Based on Data Mining. Chinese Journal of Library and Information Science for Traditional Chinese Medicine, 50, 16-20. (In Chinese)[CrossRef]
[13] Wen, Y.Y. (2015) Study on Medication Compatibility Rules in the Lower-Jiao Chapter of Systematized Identification of Warm Diseases, Henan University of Chinese Medicine. Master’s Thesis, Henan University.
[14] Yang, J. (2017) Systematized Identification of Warm Diseases. People’s Medical Publishing House.
[15] Zhang, S.C. (2022) Commentaries on Wen Bing Tiao Bian. People’s Medical Publishing House.
[16] Si, S.K. and Sun, X.J. (2015) Algorithms and Applications of Complex Networks. National Defense Industry Press.
[17] Hu, Y., Zhang, Q., Kong, X., Xu, H. and Wang, Y. (2017) Progress of Studies on Traditional Chinese Medicine Based on Complex Network Analysis. World Journal of Traditional Chinese Medicine, 3, 28-37.[CrossRef]
[18] Wan, H., Moens, M., Luyten, W., Zhou, X., Mei, Q., Liu, L., et al. (2015) Extracting Relations from Traditional Chinese Medicine Literature via Heterogeneous Entity Networks. Journal of the American Medical Informatics Association, 23, 356-365.[CrossRef] [PubMed]
[19] Lu, W.J. (2026) On the Medication Rules in Shang Han Lun Based on Complex Network Analysis. Journal of Biosciences and Medicines, 14, 149-181.[CrossRef]

Copyright © 2026 by authors and Scientific Research Publishing Inc.

Creative Commons License

This work and the related PDF file are licensed under a Creative Commons Attribution 4.0 International License.