Abstract Background and aim Due to the complexity of TCM ingredients and medication compatibility, TCM cannot be used like chemical medicines. The theory of “Four Natures and five Flavors” provides a theoretical basis for the use of TCM. “Four Natures and five Flavors” are originated from pharmacological rules based on clinical practices. Whereas, How to describe and characterize “Natures”(Warm, Hot, Cold and Cool) and “Flavors” (Pungent, Sour, Sweet, Bitter and Salty) scientifically remain the issue that needs to be solved. The aim of this study is to establish the TCM characterization models based on the integrated pharmacology network strategy and provide a deeper understanding of TCM theory. Experimental procedure Five “Pungent-Neutral”, nine “Sweet-Neutral and nine “Bitter-Neutral” TCMs were selected to characterize the “Flavors” (Pungent, Sweet and Bitter). Nine “Pungent-Warm” and nine “Bitter-Cold” TCMs were selected to characterize the “Natures” (Warm and Cold). The screened chemical ingredients were analyzed by classification and the screened characteristics targets were analyzed by GO and KEGG enrichment analysis. Results and conclusion In the “Pungent” group, flavonoids are the most. “Pungent” may have immune-regulatory effects and potential anticancer activity. In the “Sweet” group, isoflavones are the most. “Sweet” are related to effectively invigorate health. Fatty acids in the “Warm” group are the most. Flavonoids in the “Cold’ group are far more than other components. “Warm” and “Cold” are both related to fatty acid and energy metabolism. Keywords: Network pharmacology, Traditional Chinese medicine, Natures, Flavors, Neutral medicine Graphical abstract [33]Image 1 [34]Open in a new tab Highlights * • The network pharmacology integrated strategy were used to study “Flavors” and “Natures”. * • “Pungent, Sweet and Bitter” and “Warm and Cold” were characterized in this study. * • The characteristic compounds-targets were analyzed by classifying, GO and KEGG enrichment analysis. List of abbreviations TCM Traditional chinese medicine TCMSP Traditional chinese medicine systems pharmacology database and analysis platform ETCM The encyclopedia of traditional chinese medicine PN Pungent-neutral SN Sweet-neutral BN Bitter-neutral BC Bitter-cold PW Pungent-warm PPI Protein-protein interaction MF Molecular function BP Biological process CC Cell component KEGG Kyoto Encyclopedia Of Genes And Genomes 1. Introduction Traditional Chinese Medicine (TCM) has a rich history and is widely recognized for its efficacy in China. Among the multidimensional attributes of TCM, “Natures and Flavors”, “Tropisms”, “Ascending-Descending-Floating-Sinking”, “Toxic and Non-Toxic”, “Natures and Flavors” belong to the core theory of them. Understanding and characterizing them is a crucial scientific challenge in TCM modernization study. The “Natures” include “Warm, Hot, Cold and Cool” called “four Natures”. Some researchers have experimentally confirmed that some key players distinguishing Cold/Hot are affected by corresponding Cold/Hot ZHENG-oriented herbal treatments, respectively.[35]^1 It can be summarized that TCMs which can alleviate or eliminate fever syndrome generally belong to chill “Natures”, which can be divided into “Cold and Cool” in degree. On the other hand, The TCMs that can alleviate or eliminate chill syndrome generally belongs to thermal “Natures”, which can be divided into “Hot and Warm” in degree.[36]^2 The essence of the “Natures” is that they can regulate the body's energy production process.[37]^3 Some TCMs do not express the “Natures” trend that belongs to the “Neutral” attribute. Some researchers have found that the correlations between volatile oil and “Warm”, followed by “Cold” and alkaloids and organic acids, “Neutral” and saccharides and organic acids.[38]^4 “Flavors” include “Pungent, Sour, Sweet, Bitter and Salty” called “five Flavors”. The concepts of “Flavors” originated from people's taste perception of the TCM and gradually evolved into a high-level summary of TCM's efficacy. They are used to describe the pharmacological response of TCM to the body. The main ingredients of “Pungent” TCM include volatile oils, glycosides, and alkaloids.[39]^5^,[40]^6 “Sweet” TCM often contains nutrients such as saccharides, amino acids, vitamins, proteins sterols, and saponins, that maintain body metabolism.[41]^7^,[42]^8 Some studies have reported that “Bitter-Warm” TCM is mainly composed of volatile oil. “Bitter-Cold” is mainly composed of alkaloids and glycosides.[43]^9 The pH value of pure“Sour” TCM is mostly below five.[44]^10 The ingredients of “Sour” TCM include organic acids, flavonoids, tannins, and volatile oils.[45]^11^,[46]^12 Most “Salty” TCMs are derived from animal sources and contain inorganic salts and minerals. A kind of TCM can belong to multiple “Flavors”, but only belongs to one “Nature”. The theoretical explanations of “Natures and Flavors” are primarily documented in ancient books, and their efficacy are expounded within the theoretical framework of TCM. How to understand the scientific underpinnings of TCM while preserving and leveraging its distinctive strengths is a challenging and valuable issue. Some scholars have proposed the methods based on the separability and composability of TCM “Natures and Flavors” can promote to solve the TCM research issue.[47]^13 Network pharmacology integrates systems biology, bioinformatics, network science, and other disciplines and analyzes the molecular relationship between drugs and treatment objects to reveal the systematic pharmacological mechanism of drugs from an overall perspective of the system level and biological network, thereby guiding the drug development and clinical diagnosis and treatment.[48]^14^,[49]^15 The multi-level network models using network pharmacology technology to study TCM from a comprehensive perspective, have become novel strategies to elucidate the TCM effectiveness and scientific basis. At present, there are few studies on the theory of the “Natures and Flavors” by using network pharmacology technology. Jiang Miao proposed a new strategy to construct a classification model of “Cold and Hot” of TCM from pharmacological biological effects using network pharmacology.[50]^16 In our previous study, we characterized the “Natures” (Pungent, Sweet, Bitter) and “Flavors” (Warm, Cold) based on metabolomics.[51]^17 In this study, We collected “Pungent-Neutral”, “Sweet-Neutral, “Bitter-Neutral”, “Pungent-Warm” and “Bitter-Cold” TCMs to established the “Natures and Flavors” -ingredients-targets network models to characterize “Flavors” (Pungent, Sweet and Bitter) and “Natures” (Warm and Cold) using integrated network pharmacology. The characteristics ingredients were classified and analyzed, and the characteristics targets were analyzed using GO and KEGG enrichment analysis to gain further understanding of the “Natures and Flavors” mechanism. 2. Database and tools 2.1. Database TCMSP(old.tcmsp-e.com/tcmsp.php), contains 499 Chinese herbal medicines and approximately 29000 ingredients[52]^18; HERB(herb.ac.cn/), contains 263 Chinese herbal medicines and over 28000 related ingredients[53]^19; Shanghai Chemical Professional Database([54]www.organchem.csdb.cn/scdb/), contains over 22000 herbs and more than 19700 ingredients[55]^20; ETCM([56]www.tcmip.cn/ETCM/), contains more than 400 Chinese herbal medicines and over 7200 ingredients[57]^21; PubChem (pubchem.ncbi.nlm.nih.gov/), is the world's largest database for free access to chemical information. 2.2. Tools SwissADME ([58]http://www.swissadme.ch/), is an online tool that can calculate the physicochemical properties of compounds, and predict pharmacokinetic parameters and drug-like properties[59]^22; SwissTargetPredition ([60]http://www.swisstargetprediction.ch/), is an online tool for the most likely macro-molecular targets for predicting small compounds. The prediction tool is based on the 2D and 3D structural similarity of over 3000 proteins and 3.7 million known active compounds from three different species, including humans[61]^23; Metascape([62]https://metascape.org/), is an integrated and user-friendly online tool that provides comprehensive gene list annotation and analysis resources, which can be annotated and enriched based on targets obtained from network pharmacology.[63]^24 Cytoscape 3.9.1, is an open-source freeware tool that can be used to analyze, edit and visualize the network data of the protein-protein interactions.[64]^25 3. Method 3.1. Candidate tool TCMs The “single-Flavor” TCMs were obtained and counted from the Chinese Pharmacopoeia (2020 edition). The appropriate TCMs were selected as the candidate tool TCMs for studying “Natures and Flavors” based on the statistical results. Finally, five “Pungent-Neutral”, nine “Sweet-Neutral” and nine “Bitter-Neutral” plant TCMs were selected to study “Flavors” (Pungent, Sweet, and Bitter). Nine “Pungent-Warm” and nine “Bitter-Cold” plant TCMs were selected to study the “Natures’ (Warm and Cold). The candidate TCMs divided into five groups as follows： Pungent-Neutral:Lobelia chinensis Lour.(PN1), Euphorbia humifusa Willd.(PN2), Selaginella tamariscina (Beauv.) Spring(PN3), Hibiscus mutabilis L.(PN4), Eupatorium fortunei Turcz.(PN5) Sweet-Neutral:Codonopsis pilosula (Franch.) Nannf.(SN1), Glycyrrhiza uralensis Fisch.(SN2), Lycium barbarum L.(SN3), Gastrodia elata Bl.(SN4), Cannabis sativa L.(SN5), Dioscorea opposita Thunb(SN6), Glycine max (L.).Merr.(SN7), Crocus sativus L.(SN8), Sesamum indicum L.(SN9) Bitter-Neutral:Nelumbo nucifera Gaertn.(BN1), Viscum coloratum (Komar.) Nakai(BN2), Dioscorea spongiosa J. Q.Xi, M. Mizuno et W.L. Zhao(BN3), Morus alba L.(BN4), Eupatorium lindleyanum DC(BN5), Prunus persica (L.). Batsch(BN6), Sargentodoxa cuneata (Oliv.) Rehd. et Wils.(BN7), Citrus reticulata Blanco(BN8), Vaccaria segetalis (Neck.) Garcke(BN9) Pungent-Warm:Angelica dahurica (Fisch.ex Hoffm.) Benth.et Hook.f.(PW1), Ligusticum chuanxiong Hort.(PW2), Eugenia caryophyllata Thunb.(PW3), Eucommia ulmoides Oliv.(PW4), Lindera aggregata (Sims) Kos-term.(PW5), Sinapis alba L.(PW6), Carthamus tinctorius L.(PW7), Magnolia biondii Pamp.(PW8), Asarumsieboldii Miq.(PW9) Bitter-Cold:Isatis indigotica Fort.(BC1), Iris tectorum Maxim(BC2), Belamcanda chinensis (L.). DC(BC3), Gardenia jasminoides Ellis(BC4), Scutellaria baicalensis Georgi(BC5), Sophora flavescens Ait.(BC6), Aloe barbadensis Miller(BC7), Rheum palmatum L.(BC8), Stephania tetrandra S.Moore(BC9) 3.2. Collection of chemical ingredients The ingredients obtained from various databases and literature sources. On the one hand, the pertinent chemical ingredients of each candidate TCM were gathered from TCM databases, such as TCMSP and HERB et al. Additionally, the ingredients mentioned the relevant activity in the published literature were added. All of them were searched and recorded through the PubChem database, which provided basic information such as chemical names, formulas and CAS numbers. Theirs SDF structural files were downloaded simultaneously. For the ingredients were not included in the PubChem database, We drew their structures using the software ChemDraw 19.0. Finally, all ingredients of each TCM were collected together to remove duplicates to establish a chemical ingredients set for each TCM. 3.3. Target prediction and screening of the active ingredients The TCMs mentioned above are all for oral use (not external use). All collected chemical ingredients were predicted their pharmacokinetics parameters in the body firstly using the SwissADME tool for ADME screening. The screening parameters were set to Gastrointestinal absorption = High, and the druglike parameters were at least two “Yes” were as the screening criteria. Gastrointestinal absorption is predicted with the BOILED-Egg model classification. The classification showed 10-fold cross-validation accuracy of 88 % for Gastrointestinal absorption. The Drug-likeness includes six rule-based methods include the Lipinski rule-of-five, the Ghose filter, the Veber filter, the Egan filter, the Muegge filter and the Bioavailability Score.[65]^22 In addition, Some collected ingredients have been confirmed that their were pharmacological components through literature, which were regarded as the active ingredients without screening process. SwissTargetPrediction tool was used to predict the human targets for all screened active ingredients. The targets prediction are based on the similarity principle, in 2D and 3D, within a larger collection of 376342 compounds known to be experimentally active on an extended set of 3068 macromolecular targets. In validation, for 72 % molecules (total 500), at least one of the experimentally known targets can be found among the predicted top-15.[66]^23 Finally, ingredients with predicted targets and the corresponding targets were collected. The top 15 predicted targets of each ingredients with the probability greater than 0.1 were selected, as well the probability greater than 0.5. The probability values provided by SwitssTargetPrediction tool are calculated from the Combined-Scores of the most similar compounds to the query molecule (in 2D and 3D) known to be active on a given protein. Importantly, this value precisely illustrates the probability for a bioactive molecule to have a given protein as target, but not the probability of being bioactive.[67]^23 The data sets for the active ingredients-targets fo each candidate TCMs were obtained. 3.4. Protein-protein interaction network and screening of “natures and flavors” tool TCMs The predicted targets of each TCM were imported into the software Cytoscape 3.9.1. The GeneMANIA plugin tool in the software was used by setting the species as H.sapiens (Human) and performing protein-protein interaction (PPI) analysis to obtain the PPI network set of each TCM. The PPI network obtained for each TCM can be viewed as a compilation of all potential targets that the TCM can impact within the body, along with their corresponding interaction relationships. Except for the “Pungent-Neutral” group, which only contains five candidate TCMs, the other groups were compared and analyzed based on their collection information, including chemical composition, active ingredients and PPI network data of TCMs within their respective groups. Each groups was selected only six candidates as the tool TCMs. 3.5. Natures and flavors related targets and characteristic targets The Merge function in Cytoscape 3.9.1 was used to merge and extract same targets from the PPI targets network of five “Pungent-Neutral” tool TCMs to construct a “Pungent-Neutral” common targets network. Under the same operation, the common targets of other groups were merged and extracted. The “Sweet-Neutral” common target network, “Bitter-Neutral” common target network, “Pungent-Warm” common target network and “Bitter-Cold” common target network have been acquired. The schematic acquisition of Natures and Flavors targets are shown in [68]Fig. 1. The “Pungent-Neutral”, “Sweet-Neutral” and “Bitter-Neutral” common target networks were merged to eliminate the same targets of the three networks, and obtained their own exclusive target network regarded as the “Pungent” target network, “Sweet” target network, and “Bitter” target network respectively. The “Pungent-Warm” common targets were regarded as the “Warm” targets by eliminating the targets that same with the “Pungent-Neutral’ common targets. The “Bitter-Cold” common targets were regarded as the “Cold” targets by eliminating the targets that sa with the “Bitter-Neutral’ common targets. The Network Analyzer tool of Cytoscape 3.9.1was used to calculate the network topology parameters of the “Natures and Flavors” target networks. The core targets in each ““Natures and Flavors”” target network were screened by using the parameters (Degree, Betweenness, Closeness and other parameters) as the screening rules which were greater than the median or twice the median. The screened core targets of the “Pungent” targets network were regarded as “Pungent” characteristic targets. On the same operation, The characteristic targets of “Sweet”, “Bitter”, “Warm” and “Cold” were obtained, respectively. Fig. 1. [69]Fig. 1 [70]Open in a new tab The screening of Natures and Flavors targets (A)“Pungent-Neutral” common targets; (B)“Sweet-Neutral” common targets; (C)“Bitter-Neutral” common targets; (D)“Pungent-Warm” common targets; (E)The “Bitter-Cold” common targets. 3.6. “Natures and flavors” ingredient sets analysis The chemical ingredients corresponding to the characteristic targets of their respective “Natures and Flavors” tool TCMs were collected and formed into “Natures and Flavors” ingredients sets, respectively. The “Pungent” ingredients set was formed from five “Pungent-Neutral” tool TCMs. During the same operation, The “Sweet”, “Bitter”, “Warm” and “Cold” ingredients sets were obtained, respectively. Afterward, the above mentioned ingredients sets were analyzed by classification. 3.7. GO and KEGG enrichment analysis The online tool Metascape was used to conduct the GO gene functional analysis of the related targets and all characteristic targets, including molecular function (MF), biological process (BP), cell component (CC) analysis and KEGG pathway enrichment analysis. By setting the threshold at p < 0.05 and selecting the top enrichment results for visualization, and the mechanisms related to “Natures and Flavors” were discussed. 4. Results and discussion 4.1. Selection of candidate tool TCMs “Natures” and “Flavors” are different attributes of TCM, While TCM have the characteristics of “single-Nature and multi-Flavors” generally and it is difficult to control variables in research. From the viewpoint that the “Neutral” TCM does not make the body heat or cold and only exerts efficacy through itself “Flavors”.[71]^26 It is appropriate to study the “Flavors” in relation to the “Neutral” TCM. Thus, We counted only the “single-flavor” TCMs to simplify complexity of the “multi-flavors”. In our previous study[72]^17 according to [73]Table S1, we chose the “Pungent-Neutral”,“Sweet-Neutral and “Bitter-Neutral” TCMs for characterizing the “Flavors” (Pungent, Sweet and Bitter), and “Pungent-Warm” and “Bitter-Cold” TCMs for characterizing the “Natures” (Warm and Cold). due to ADME screening and target prediction tools only support small molecule compound analysis, so we can only select plant TCMs, and exclude animal, mineral and fungal TCMs. we preferentially selected the plant TCM with as many chemical ingredients as possible as the candidate TCMs in this work. 4.2. Collection and screening of ingredients, PPI network, and tool TCMs The results of collecting and screening TCM information are shown in [74]Table 1. In “Sweet-Neutral” Group, there are fewer PPI targets for SN7 and SN8, as well as fewer active ingredients in SN9. In “Bitter-Neutral” group, the PPI targets of BN6 are fewer, and the chemical components of BN7, BN8 and BN9 are fewer. Among “Pungent-Warm” group, there are fewer PPI targets for PW7, PW8 and PW9. Among “Bitter-Cold” group, there are fewer active ingredients in BC7 and BC8. There are fewer PPI targets for BC9. Finally, five “Pungent-Neutral”, Six “Sweet-Neutral”, five “Bitter-Neutral”, six “Pungent-Warm” and six “Bitter-Cold” candidate TCMs were selected as tool TCMs. The selected tool TCMs are marked with an asterisk (*) in the following tables. Table 1. Collection and Screening of candidate TCMs ingredients and targets. TCM Chemical components SwissADME Screening Active ingredients Predicting targets PPI targets PN1* 165 116 82 305 7826 PN2* 136 38 31 233 7162 PN3* 180 87 64 306 8334 PN4* 66 20 20 182 6510 PN5* 251 163 97 310 8038 SN1* 289 158 94 331 8390 SN2* 357 246 202 447 9249 SN3* 264 114 77 370 7558 SN4* 96 65 36 221 7146 SN5* 198 123 85 358 8270 SN6* 146 93 64 331 8395 SN7 174 88 61 247 6877 SN8 131 93 52 246 6424 SN9 133 57 20 – – BN1* 131 79 53 259 7032 BN2* 119 55 42 198 6495 BN3* 107 49 38 197 4958 BN4* 164 104 68 319 7824 BN5* 113 79 49 225 7087 BN6 87 63 31 97 2853 BN7 46 – – – – BN8 52 – – – – BN9 42 – – – – PW1* 306 187 124 294 8088 PW2* 293 178 96 314 7640 PW3* 234 131 68 285 7631 PW4* 191 104 69 327 7870 PW5* 183 131 79 338 7963 PW6* 178 99 72 214 7566 PW7 237 95 60 277 7333 PW8 222 143 80 257 6613 PW9 235 147 89 286 5517 BC1* 229 147 102 419 8592 BC2* 166 107 83 234 7638 BC3* __________________________________________________________________ 116 __________________________________________________________________ 66 __________________________________________________________________ 61 __________________________________________________________________ 260 __________________________________________________________________ 7856 __________________________________________________________________ BC4* 181 100 69 288 7908 BC5* 231 129 106 368 7742 BC6* 203 150 120 227 8452 BC7 129 50 35 302 5424 BC8 147 60 43 212 6531 BC9 105 74 57 294 5102 [75]Open in a new tab 4.3. Natures and flavors targets network and characteristic targets Firstly, the “Pungent-Neutral” common target network was established by the same targets of five “Pungent-Neutral” tool TCMs PPI network targets using Merge function of Cyctospae 3.9.1. The Merge function can merge and extract the same targets of multiple TCM target networks and preserve the relationship between the targets. Than, the “Pungent-Neutral” network contains 5732 target proteins and 131533 interaction relationships. Using the same method, we obtained “Sweet-Neutral” common target network with 4749 target proteins and 126280 interaction relationships. The “Bitter-Neutral” common target network has 3404 target proteins and 90427 interaction relationships. The “Pungent-Warm” common target network has 5521 target proteins and 130840 interaction relationships. The “Bitter-Cold” common target network has 6116 target proteins and 141319 interaction relationships. After merging the “Pungent-Neutral”, “Sweet-Neutral” and “Bitter-Neutral” common target networks, the “Pungent” target network contains 1316 target proteins and 891 interaction relationships. The “Sweet” target network contains 487 target proteins and 596 interaction relationships. The “Bitter” target network contains 157 target proteins and 67 action relationships. The “Warm” target network contains 481 target proteins and 594 interaction relationships by merging “Pungent-Warm” and “Pungent-Neutral” common target networks. The “Cold” target network contains 2896 target proteins and 8039 interaction relationships by merging “Bitter-Cold” and “Bitter-Neutral” common target networks. Through network topology analysis, the core targets were screened as characteristic targets. As shown in [76]Table S2 and Tabl S3, there are 175 characteristic targets for “Pungent’, 195 characteristic targets for “Sweet”, 95 characteristic targets for “Bitter’, 194 characteristic targets for “Warm”, and 199 characteristic targets for “’Cold”. The results are shown as follows. 4.4. Analysis of natures and flavors ingredients sets The chemical ingredients corresponding to the characteristic targets were obtained. As shown in [77]Table S4, the “Pungent” group includes 30 chemical ingredients and 6 related targets. The analysis results of the “Pungent” group are shown in [78]Fig. 2. Among them, The amount of flavonoids are the most, followed by alkaloids, monoterpenes and sesquiterpenes. It has been reported statistically that the main components of “Pungent’ TCMs are volatile oil (such as, monoterpenes and sesquiterpenes) and alkaloids.[79]^5^,[80]^6 In this study, it was found that the major components were flavonoids. Flavonoids, such as quercetin, apigenin, naringenin, fisetin and luteolin, mostly have anti-inflammatory, antibacterial, antiviral effects.[81]27, [82]28, [83]29, [84]30 These effects are correspond to the “Pungent” can relieve exterior syndrome. Ricinine and atropine have central nervous system excitatory effect,[85]^31^,[86]^32which emboding the inducing resuscitation effect of “Pungent”. Carvacrol has vasodilating effects and ellagic acid can treat cardiovascular diseases,[87]^33^,[88]^34 Acacetin has the effect to protect the heart.[89]^35 Shikimic acid and its derivatives have antibacterial, antitumor, antithrombotic and anti-cerebral ischemia effects,[90]^36 which these components effects on cardiovascular diseases coincide with the promoting blood circulation effect of “Pungent’. Fig. 2. [91]Fig. 2 [92]Open in a new tab Classification of chemical compositions of the “Pungent”. As shown in [93]Table S5, the “Sweet” group includes 106 chemical ingredients and 19 related targets. The analysis results of the “Sweet” group are shown in [94]Fig. 3. Among them, The amount of isoflavones are the most, followed by glucosides, amino acids, phenols, steroids, terpenes, coumarins and others. The “Sweet” substance group has many nutrients, which may be related to the fact that ‘Sweet” TCMs have a tonifying effect and can supplement the substances needed by the body. Among the isoflavones, Formononetin and Ononin have antibacterial, anti-apoptosis and anti-inflammatory activities, and can improve myocardial damage in diabetic mice through the AKT/FoxO1 signaling pathway.[95]^37^,[96]^38 Tectorigenin has antioxidant and scavenging free radicals, regulating immunity, anti-inflammatory and protecting liver effects. Paeonol has neuroprotective and hepatoprotective effects,[97]^39 l-glutamic acid, l-tryptophan and l-citrulline and other amino acid nutrients have immune regulation, protection of the cardiovascular system, and other extensive health benefits. The effects of these components are related to enhancing the body's physical fitness, reflecting the gain compensation characteristic effect of “Sweet”. Fig. 3. [98]Fig. 3 [99]Open in a new tab Classification of chemical compositions of the “Sweet”. As shown in [100]Table S6, the “Bitter” group includes 12 chemical ingredients and 4 related targets. The analysis results of the “Bitter” group are shown in [101]Fig. 4. It cannot be found which chemical compositions are most strongly associated with the “Bitter” by the component classification. It may be related to the high molecular weight of many “Bitter” chemical components, which cannot predict the targets by the selected screening tool. On the other hand, pharmacological studies have shown that “Bitter” TCMs mainly excite or inhibit the pathological decline or hyperactivity.[102]^40 These may cause fewer same targets of “Bitter”. Furthermore, studies have indicated that the chemical ingredients of “Bitter-Neutral” TCMs are complex and the chemical compositions cannot fully reflect their main characteristics.[103]^41 As a result, there were fewer chemical components of the “Bitter” group in this study. Fig. 4. [104]Fig. 4 [105]Open in a new tab Classification of chemical compositions of the “Bitter”. As shown in [106]Table S7, the “Warm” group includes 82 chemical ingredients and 15 related targets. The analysis results for the “Warm” group are shown in [107]Fig. 5. The amount of fatty acids in the “Warm” group are the most, followed by alkaloids, terpenes and phenolic acids. Fatty acid metabolism can provide the energy required by the body, and is closely related to glycolipid metabolism and energy metabolism.[108]^42^,[109]^43 This reflects that the material basis of “Warm” is to directly provide the energy substances mainly fatty alcohols, such as, octanol, heptanol, nonanol, decanol, undecanol, dodecanol, thirteenol, tetradecyl alcohol, pentadecanol, cetyl alcohol and heptadecanol. In addition to fatty acids, “Warm” is also closely related to volatile oil (such as, monoterpenes and sesquiterpenes), as well as tryptophan-derived and lysine-derived alkaloids. Fig. 5. [110]Fig. 5 [111]Open in a new tab Classification of chemical compositions of the “Warm”. As shown in [112]Table S8, the “Cold” group includes 43 chemical ingredients and 9 related targets. The analysis results for the “Cold” group are shown in [113]Fig. 6. The amount of flavonoids in the “Cold’ group are significantly more than the other components. Flavonoids, such as, quercetin, apigenin, Chrysoeriol, kaempferol, (S)-naringenin, luteolin have good antibacterial and antiviral effects.[114]^29^,[115]^30^,[116]^38 Emodin and chrysophanic acid have hypoglycemic and lipid-reducing effects.[117]^44^,[118]^45 Baicalin has antibacterial and antiviral effects.[119]^46 Eriodictyol can improve glycolipid metabolism.[120]^47 The analysis results showed that the “Cold” can regulate the body's glycolipid and energy metabolism and also has the effect of eliminating pathogens through antibacterial and antiviral effects to eliminate pathogens and relieve fever symptoms, which exerting its “Cold” properties. Fig. 6. [121]Fig. 6 [122]Open in a new tab Classification of chemical compositions of the “Cold”. 4.5. GO enrichment and KEGG pathway analysis of “natures and flavors” related targets The results of “Pungent” targets analysis are shown in [123]Fig. S1. The spontaneous aggregation of β-amyloid proteins results in formation numerous toxic oligomers, which can cause neuronal death in the brain and lead to cognitive impairment.[124]^48 The regulation of β-amyloid proteins by “Pungent” reflects its potential enlightening effect. The active regulation of myocardial cell contraction reflects the blood-activating effect of “Pungent”. The type I interferon (IFN–I) signaling pathway is a key element of the innate immune signaling pathway,[125]^49 and its positive-regulation suggests that “Pungent” may have an immune regulatory effect. The KEGG analysis results showed that “Pungent” may affect the folate antagonist resistance metabolism. Folic acid inhibitors can inhibit biological processes such as cell division, DNA and RNA synthesis, as well as protein synthesis. Theey are frequently used in tumor chemotherapy.[126]^50 “Pungent” show the potential anticancer activity. The results of “Sweet” related targets analysis are shown in [127]Fig. S2. The GO enrichment analysis indicate that the regulation of biological processes such as cell stress response, nutrient level response and cell proliferation, which reflect the “Pungent” plays an important role in supplementing the body's deficiencies. At the molecular function level, threonine is related to protein serine/Threonine kinase activity. Threonine is an essential amino acid that acts as a nutritional enhancer. It has been shown to alleviate fatigue and promote growth and development in humans. Its regulation reflects “Sweet” nourishing effect. The results of “Bitter” related targets analysis are shown in [128]Fig. S3. Similar to the results of the “Bitter” composition analysis, no significant “Bitter” characteristics were found. Some study have reported that “Bitter-Neutral” TCM have many pharmacological effects, including dispelling wind and dampness, reducing blood pressure, improving blood circulation and hemostasis.[129]^40 Therefore, when using the “Bitter-Neutral” TCM to summarize the characteristics of “Bitter”, The “Bitter” characteristics are not clear. The results of “Warm” related targets analysis are shown in [130]Fig. S4. The GO enrichment analysis showed that the “Warm” intervention may be reflected in the regulation of cardiac contraction, the process of cellular protein catabolism, the process of Messenger RNA metabolism, the process of mRNA catabolism, the positive regulation of protein hydrolysis and other biological processes. The impacts of “Warm” on these catabolic biological processes are related to the promotion of material and energy metabolism in the body. The regulation of cardiac contraction suggests that “Warm” may play a role in regulating the circulatory system. At the cellular component level, β-Endorphins have a strong analgesic effect and can promote the release of Glucagon[131]^51 and α1-adrenergic receptors are mainly distributed in vascular smooth muscle and can cause vasoconstriction when stimulated. The KEGG metabolic analysis results showed that “Warm” is closely related to the pancreatic hyperglycemia signaling pathway, as well as the linoleic acid metabolism and fatty acid degradation. Glucagon is secreted by islet A cell. Its primary function is to stimulate the conversion of the glycogenolysis and non-glucose substances into glucose, thereby increasing blood glucose levels. Fatty acids can undergo oxidation in liver cells to produce linoleic acid. Linoleic acid is bound to the glycoester metabolism-related kinases in the cytoplasm, which activate the process of glucose and lipid metabolism, and release energy to improve blood glucose levels.[132]^52 “Warm” may promote the metabolism of substances and thermogenesis in the body through these pathways. The results of “Cold” related targets analysis are shown in [133]Fig. S5. The GO enrichment analysis indicated that both “Cold” and “Warm” have impact on the breakdown and metabolism of various proteins and fatty acid in the body. It is speculated that “Warm” promotes metabolism, while “Cold” inhibits metabolism. At the same time, “Cold” is also related to the Hippo signaling pathway. The Hippo signaling pathway is a signaling pathway that inhibits cell growth. The Hippo signaling pathway also plays an crucial role in cancer development, tissue regeneration and the regulation of stem cell function.[134]^53 In molecular function analysis, “Cold” is associated with functions such as ATP dependent activity and ATPase-coupled ion transmembrane transport protein activity. It is speculated that the possible mechanism of “Cold” is that it can directly inhibit the functional regions related to energy metabolism.[135]^54^,[136]^55 The KEGG enrichment analysis revealed that “Cold” may directly inhibit the metabolism of fatty acids and other substances, leading to reduction in energy release and heat production. In general, upon analyzing the selected related “Natures and Flavors” targets, there are still numerous unexplainable results. However, some certain characteristics of “Natures and Flavors” have been identified, which providing the theoretical basis for the contemporary comprehension of TCM's characterization of “Natures and Flavors”. 5. Conclusion The theory of TCM properties involves multiple-dimensions. It is difficult to control variables and obtain expected multi-dimensional correlation results when conducting synchronous research at high dimensions. Although network pharmacology can not reflect the dose-effect relationship between compounds and pharmacological actions, it provides a new research model from the whole perspective in TCM modernization initial stage. At present, “Natures and Flavors” is not clear, but we has revealed some scientific and regularity of them. In this study, the “Pungent-Neutral”, “Sweet-Neutral”, “Bitter-Neutral”, “Pungent-Warm”, “Bitter-Cold” TCMs were used to screen the “Natures” and “Flavors” characterized compositions and targets using network pharmacology integrated strategy. In the “Pungent” group, flavonoids are the most. “Pungent” may have immune-regulatory effects and potential anticancer activity. In the “Sweet” group, isoflavones are the most. “Sweet” are related to effectively invigorate health. Fatty acids in the “Warm” group are the most. Flavonoids in the “Cold’ group are far more than other components. “Warm” and “Cold” are both related to fatty acid and energy metabolism.The “Pungent”, “Sweet”, “Bitter”, “Warm” and “Cold” characterized compositions and targets were identified, which providing the foundation for further understanding of the rules and identification of the “Natures” and “Flavors”. Credit author statement Yu Zhiguo: Conceptualization, Methodology Zhao Yunli: Conceptualization Wang Hong: Methodology, Formal analysis, Investigation, Writing - Original Draft Wei Wenfeng: Investigation Liu Jing: Investigation Zhang Shuang: Investigation. Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Acknowledgements