Taste of phytocompounds: A better predictor for ethnopharmacological activities of medicinal plants than the phytochemical class?

Journal of Ethnopharmacology Volume 220, 28 June 2018, Pages 129-146 Author links open overlay panelDorinDragosabMarilenaGilcac a Medical Semiology Dept., Faculty of General Medicine, Carol Davila University of Medicine and Pharmacy, B-dul Eroilor Sanitari nr.8, 050471 Bucharest, Romania b Nephrology Clinic, University Emergency Hospital Bucharest, Bucharest, Romania c Biochemistry Dept., Faculty of General Medicine, Carol Davila University of Medicine and Pharmacy, B-dul Eroilor Sanitari nr.8, 050471 Bucharest, Romania Received 12 July 2017, Revised 26 March 2018, Accepted 26 March 2018, Available online 29 March 2018. crossmark-logo https://doi.org/10.1016/j.jep.2018.03.034 Get rights and content Referred to by Dorin Dragos, Marilena Gilca PhytoMolecularTasteDB: An integrative database on the “molecular taste” of Indian medicinal plants Data in Brief, Volume 19, August 2018, Pages 1237-1241 Download PDF Abstract Ethnopharmacological relevance Understanding the patterns that shape traditional medical knowledge is essential for accelerating ethnopharmacological progress. According to Ayurveda, medicinal plants that belong to different taxa, but which have similar taste, may display similar (ethno)pharmacological activities (EPAs) (Bhishagratna, 1998; Sharma and Dash, 2006). Aim of the study To understand the patterns that govern the distribution of herbal EPAs in Ayurveda and to evaluate the potential concordance between chemical class or taste of the constituent phytocompounds and EPAs. Material and methods A mixed database (PhytoMolecularTasteDB) was constructed for Ayurvedic medicinal plants by integrating modern data (medicinal plant composition, phytochemical taste) with traditional data (ethnopharmacological activities of plant). PhytoMolecularTasteDB contains 431 Ayurvedic medicinal plants, 94 EPAs, 223 chemical classes of phytocompounds and 438 herbal tastants. Potential global or individual associations between chemical classes/taste of the phytoconstituents and EPAs were statistically analyzed. Results There was no global statistical correlation between the various chemical classes of phytocompounds and EPAs, although there were several individual correlations. The results suggest the existence of a global statistical correlation (besides several individual correlations) between the plant “molecular taste” (various taste-based classes of phytocompounds) and EPAs. Conclusions These results suggest that phytochemical taste may be more relevant than chemical class for EPAs prediction. Graphical abstract fx1 Download high-res image (182KB)Download full-size image Previous article in issue Next article in issue Abbreviations 95% CL95% confidence limits Comp/Comp-scategories of compounds EPA/ EPAsethnopharmacological action/s EV/ EVsexpected value/s EVTEVs table dfdegrees of freedom ODTobserved data table LT5EVsless than 5 EVs OROdds Ratio SElnORstandard error of the natural logarithm of OR TASRtaste receptors #number of Keywords Taste Rasa Medicinal plants Phytochemical Ayurveda Ethnopharmacology 1. Introduction Building a solid theoretical ethnopharmacology is “the most demanding work for the future” of this field (Etkin and Elisabetsky, 2005). Ethnopharmacology "strives for a more holistic, theory-driven, and culture- and context sensitive study” of the therapeutic and nutritive potential of botanical species used by indigenous peoples (Etkin and Elisabetsky, 2005). Hence, the interest of ethnopharmacologists and medical anthropologists has shifted in recent decades from random screening of traditional herbs or data collecting to studies aimed at a deeper understanding of traditional medical epistemology. In particular, this includes an assessment of the criteria used for herb selection in various communities, and an understanding of the relationship of people to plants in a cultural context (Heinrich, 2008, Pieroni and Torry, 2007, Reyes-Garcia, 2010). Revitalization of traditional knowledge within the context of modern science (e.g. bioprospection or predictive science), requires the integration of historical (e.g. traditional use of plants) and modern data (e.g. medicinal chemistry, botanical information) in an appropriately designed ethnopharmacological database that facilitates data mining process (Marilena Gilca and Barbulescu, 2015a, Gilca and Barbulescu, 2015b; Leonti and Verpoorte, 2017; Paolini et al., 2006; Schuffenhauer and Jacoby, 2004). 1.1. Ethnopharmacology and the sensory ecology approach The sensory ecology approach is an anthropological model proposed by Shephard in 2004, which is centered on sensation as a biocultural process deeply rooted in physiology. The process is shaped by individual experience, cultural and environmental factors. According to this model, sensation mediates the interactions among plants, animals, humans, and spirits, forming an organic network (Shepard JR, 2004). This interdisciplinary approach suggests application to a broad range of anthropologic and ethnopharmacological questions that may trigger important progress. Within the framework of the sensory ecology model, sensorial similarity refers to similarity of interactions (similar herbal therapeutic activities) and sensorial dissimilarity refers to dissimilar interactions (dissimilar herbal therapeutic activities). Analysing the role of taste in the exploration of the chemosensory environment and study of basic sensorial information translates culture into knowledge. This translation consolidates the theoretical foundation for ethnopharmacology and offers explanations for the use of herbs as medicines (Brett and Heinrich, 1998). In certain ethnomedical systems, sensation plays a role in the cultural construction of an illness (e.g. in Ayurveda there is an alternative, taste-related, classification of pathology in diseases caused by excessive sweetness, saltyness, sourness, bitterness, astringency or pungency) (Bhishagratna, 1998, Murthy, 1994, Sharma and Dash, 2006). Beginning with Descartes and continuing with Cartesian scientists, sensations were regarded with disdain and mistrust (Shepard JR, 2004). It is our purpose to revive interest in sensation, within the fields of ethnopharmacology, anthropology and the chemical senses. Brett, Heinrich and Ankli evidentiated since '90 s herb patterns with particular organoleptic properties used for the treatment of particular types of pathology (Ankli et al., 1999, Brett, 1994, Heinrich, 1994, Weimann and Heinrich, 1998). Since then, several teams of scientists confirmed the importance of this sensorial approach in various ethnomedical systems (de Medeiros et al., 2015; Marilena Gilca and Barbulescu, 2015a, Gilca and Barbulescu, 2015b; Joshi et al., 2007; Leonti et al., 2002; Molares and Ladio, 2014). Chemosensory input (such as herbal taste and smell) is first cognitively structured, named, and assigned medicinal utility, and then systematically organized and assembled into a sensory based pharmacopoeia, which often has a biological basis (Gollin, 2004). Moreover, plant efficacy is a cultural construct, sometimes based on taste or smell (Gollin, 2004, Ortiz de Montellano, 1986, Snow, 2016). In East Kalimantan (Borneo), the intensity of herbal taste is associated with the healing strength of the plant (e.g. some intensely bitter tasting plants have powerful emetic activity, strong enough to “expel the disease from the body”) (Gollin, 2004). The smell of Nekemagi aromatic plants (New Guinea) “carries” the herbal healing power, “circulates inside the body and drives out the sickness” (Johannes, 1986, Johannes, 1975). In European herbal medicine it is also stated “insipid plants and such as have no smell have rarely any medicinal virtue”(Milne, 1805). In Yucatec Maya traditional medicine, absence of taste or smell in a plant is indicative of a lack of therapeutic potential (Ankli et al., 1999), while in Ayurveda it is stated “There is no potency (virya) without a taste (rasa) (…).” (Sushruta Samhita, Sutrasthana XL.11–15) (potency meaning therapeutic potency) (Bhishagratna, 1998). The hot-cold concept has also been used in plant selection, and either criticized (e.g. considered too narrow to explain plant use) (Brett and Heinrich, 1998) or more recently, emphasized (e.g. considered the key predictor of therapeutic uses) (Geck et al., 2017). 1.2. Ayur-taste: a complex perceptual modality According to Ayurveda, taste (rasa) is an important ethnopharmacological descriptor of the substance (dravya) and a criterion for medicinal plant selection (Sharma and Dash, 2006). Ancient Indian physicians recognized six basic rasas or ayur-tastes: sweet, salty, sour, pungent, bitter, and astringent (Sharma and Dash, 2006). Even though pungency and astringency are not considered primary taste modalities by modern science, they are trigeminal orosensations (Jiang et al., 2014, Macpherson et al., 2005, Roper, 2014, Schiffman et al., 1992, Schobel et al., 2014), and part of the complex significance of rasa (Gilca and Dragos, 2017). Rasa or the ayur-taste includes all the three basic chemosensations (taste, odor, detection of irritation), best translated as “flavor”. This is a rather global phenomenon, noted in many other ethnomedical systems, such as the Amazonian (Shepard JR, 2004), Mesoamerican (Geck et al., 2017), Chinese (Bensky et al., 2004) and South Asian traditional medical systems (Gollin, 2004). For example, in Traditional Chinese Medicine, taste (Chinese wèi) “describes the perception of the herb in the mouth” (Bensky et al., 2004). Therefore, the ethno-taste is a complex perceptual modality rather than a simple sensorial modality. This description is quite similar to a modern scientific definition of flavor (Auvray and Spence, 2008, Spence, 2015, Stevenson, 2014). In Ayurveda, this feature is best exemplified by katu rasa (ayur-pungency). Unlike the English “spicy-hot” that emphasizes thermal qualities, the Ayurveda term for pungency (katu), has a more versatile significance, referring to pungent, sharp, penetrating, hot, caustic, acrid¸ but also to strong-scented, fragrant, exhaling strong odor, and stimulating (about smell) (Monier-Williams, 2002). This explains why aromatic plants, which are rich in volatile oils (even those inducing a cooling sensation, such as mint) are considered in Ayurvedic texts to have katu rasa (“pungency”) (Pandey, 2005). Interestingly, modern science considers that a pungent sensation of a higher or lesser intensity is present, whenever the essential oil of an aromatic plant is tasted. The intensity of this sensation depends on the concentration of the essential oil (Govindarajan, 1979, Hirasa and Takemasa, 1998). In order to remain faithful to Ayurvedic concepts, all the volatile compounds, even those which induce cooling (menthol) (Peier et al., 2002), numbing (eugenol) (Klein et al., 2013), or tingling (sanshool) (Kuroki et al., 2016) sensations should also be considered katu rasa compounds. 1.3. Taste and humoral quality corelations in Ayurveda The Ayurveda system, like many other ethnomedical systems, is centered around the ‘‘humoral theory’’ of the body: a system of correspondences in which physical and mental health are considered the result of a complex dynamic balance of opposing principles. Disease stems from a disturbance of this equilibrium. Treatment aims to restore balance through therapies based on the principle of ‘‘humoral opposition’’. The humoral Ayurvedic principles or the three doshas (Vata, Pitta, Kapha) are characterized by clusters of humoral qualities (Sanskrit guna), each dosha distinguishing itself from the other two by a specific quality (Pitta is hot while the other two doshas are cold, Vata is dry while the other two doshas are wet, Kapha is heavy while the other two doshas are light). The humoral qualities are important since they are organized like an universal code that allows “translation” of a category into another category within the tridosha based Ayurvedic system (e.g. clinical signs are “translated” into doshas, doshas are “translated” into rasas, etc.). This allows corespondence between diseases and tastes and the selection of taste-based treatments which are in ‘‘humoral opposition’’to the disease (see Fig. 1). Fig. 1 Download high-res image (248KB)Download full-size image Fig. 1. The “code” of humoral qualities and dosha-guna-rasa correspondences in Ayurveda philosophy. For instance, Pitta type diseases (characterized by an increase or aggravation of Pitta) are essentially identified by an excess of hot quality (Sanskr. ushna guna), and secondary by an excess of lightness (laghu) and/or wetness-oiliness (Sanskr. snighda). Hence, treatment is essentially cold (Sanskr. shita) in nature, based on cold tastes (bitter, astringent, sweet). Sour taste, being similar to Pitta in terms of “humoral qualities”, should be avoided, as well as the other two hot tastes (pungent, and salty). Excessive sourness produces mainly Pitta type diseases. Vata type diseases are distinguished by an excess of dry quality (Sanskr. ruksha guna) with treatment essentially oily-wet (snigdha) in nature, being based on onctuos tastes (salty, sour, sweet). Bitter taste is similar to Vata in terms of “humoral qualities” and should be avoided, as well as the other two dry tastes (pungent and astringent), which are not entirely opposite to Vata. Kapha type diseases are essentially characterized by an excess of heavy quality (Sanskr. guru guna) and treatment should be essentialy light (laghu), based on light tastes (pungent, bitter, sour). Sweet taste, being similar to Kapha in terms of “humoral qualities”, should be avoided, as well as the other two heavy tastes (salty and astringent). The “code” of humoral qualities (Sanskr. guna), or at least the hot-cold dichotomy, is a “common denominator” for many medical traditions around the globe and rather universal (Geck et al., 2017, Ma et al., 2010, Mane et al., 2010, Messer, 1987). However, one of the Ayurveda distinguishing features is the complexity of the humoral qualities system, which is based not on a single (hot-cold), two (hot-cold, dry-wet) or three (hot-cold, dry-wet, light-heavy) dichotomies like Mesoamerican, European or Chinese Traditional Medicine. Rather it is based on 10 dichotomies (hot-cold, dry-wet, light-heavy, mobile- stable, clear- sticky, gross- subtle, dense- flowing, dull- sharp, soft- hard, smooth- rough) (Sharma and Dash, 2006). This is, as far as we know, unique in the world of ethnomedicine. When a quality (guna) acts powerfully, it is mentioned as virya (Sharma, 1998). Therapeutic action cannot be expected without virya (Sharma, 2002). For instance, sour taste (amla rasa) has an intense hot quality (ushna virya), and therefore, acts as dipana (digestive tonic, which increases digestive fire or Agni) (Astanga Samhita, Sutrasthana X.10) (Stefan, 2005). Radish has emollient potency (snigdha virya) and increases Kapha, while honey has drying potency (ruksha virya) and decreases Kapha (Sushruta Samhita, Sutrasthana XL.4) (Bhishagratna, 1998). This very power represents the differentiating factor between guna and virya. Of the types of virya, two (cold-hot or shita-ushna) are essential. Scholars suggest that the explanation resides in the role played by Agni/Sun/Fire and Soma/Moon/Water in the Indian cosmogony (Sharma, 2002). Quite interestingly, Geck et al. recently highlighted the interrelationships between humoral and chemosensory properties of herbals in a study performed with a community of Zoque healers, in Chiapas, Mexico (Geck et al., 2017). They demonstrated that the hot-cold humoral qualities are not simple post hoc attributions to a certain plant, used to validate the treatment. Rather this dichotomy acts as a cultural filter, which mediates links between empirically perceived properties and plant therapeutic indications. The relationships between chemosensory and humoral qualities have not received sufficient attention from ethnopharmacologists, and they have never been systematically re-assessed in Ayurveda. Focusing our attention on this topic, a few Sanskrit verses from an ancient Ayurveda text awakened our interest. “There is no potency without a taste, and taste without a substance is an absurdity. Hence, substance (dravya) is the greatest of all [our note: all means dravya (substance), guna (qualities), rasa (taste), virya (potency), vipaka (postdigestive effect)]” (Sushruta Samhita, Sutrasthana XL.11–15) (Bhishagratna, 1998). Another reason for the supremacy of substance (dravya) over rasa and virya is that a substance (dravya) is the “receptacle of the attributes” of taste (rasa), qualities (guna), potency (virya), etc, while the latter are the “things contained” (Sushruta Samhita, Sutrasthana XL.2) (Bhishagratna, 1998). “Taste (rasa) and substance (dravya) are correlative categories from the time of their origin, like a body and an embodied self in the plane of organic existence.” (Sushruta Samhita, Sutrasthana XL.16–18) (Bhishagratna, 1998). Taking into account the following three facts: 1) the supremacy of substance (dravya) 2) the inseparable connection (samavaya) between taste (rasa) and substance (dravya), and 3) the molecular structure of a substance (dravya), the integrative concept of tastant molecules or “molecular taste”, is an important key to understanding the distributions of EPAs among Ayurvedic herbs. We identified a correlation between the ayur-tastes (rasa) traditionally assigned to a plant and its acknowledged EPAs, traditionally known as karman (Marilena Gilca and Barbulescu, 2015a, Gilca and Barbulescu, 2015b). Therefore, the basis for this rasa-EPA association was evaluated and chemical categories seemed to be the most obvious candidates responsible for this association. 1.4. Aim of the study The general purpose was to better understand the taste-based patterns that determine the distribution of EPAs among Ayurvedic herbs. The chemosensory property of a herbal drug is determined by the constituent phytocompounds. The phytocompounds may be characterized by two different attributes/variables: chemical class and taste (Fig. 1) The two attributes/variables are not deduced from one another, since same chemical class does not indicate same taste, nor does the same taste indicate the same chemical class. (Fig. 2) Fig. 2 Download high-res image (155KB)Download full-size image Fig. 2. Main objectives of the present study. Taking these aspects into account, the specific objectives of this research were to: • Explore the molecular basis for taste based-patterns that rule the distribution of herbal EPAs among Ayurvedic herbs. • Evaluate the potential concordance among the chemical classes of the constituent phytocompounds and the EPAs • Evaluate the potential links between taste of the phytochemicals and EPAs • Determine which attribute of the phytocompounds (either the chemical class or the taste) has a higher predictive EPAs power or is a better predictive tool • Propose explanations for observed associations (if they exists). 2. Material and methods An interdisciplinary chemosensorial approach was used. We have build, to the best of our knowledge, the first mixed database (identified as PhytoMolecularTasteDB) for Ayurvedic medicinal plants, by integrating modern phytochemical and chemosensorial data with traditional Ayurvedic data (Dragos and Gilca, n.d.). PhytoMolecularTasteDB has two parts: 1) a list of plant-derived tastants found in Indian medicinal plants included in this study; 2) a list of the “molecular taste” of Ayurvedic medicinal plants (the “molecular taste” of a plant is the combination of tastes resultant from all the principal/major phytoconstituents which are tastants or taste provoking). For a given plant, the following constituents were considered as principal/major: 1) phytocompounds mentioned in Ayurveda/Indian Materia Medica or Pharmacopoea in the rubrics regarding herbal chemical composition (see Data in Brief) (Dragos and Gilca, n.d.) 2) phytocompounds designated as “main”, “principal”, “major” in at least one of the scientific references (see Data in Brief, Appendix B) (Dragos and Gilca, n.d.) 3) phytocompounds to which one of the scientifically recognized pharmacodynamic actions of that plant is currently attributed, as revealed by the 1077 scientific references (see Data in Brief, Appendix B) (Dragos and Gilca, n.d.). PhytoMolecularTasteDB contains 431 ayurvedic medicinal plants, 94 EPAs, 223 chemical classes of phytocompounds and 438 herbal tastants. For only 394 plants a taste could be defined based on the chemical constituents (see Data in Brief, Table A.6 for a list of phytocompound taste, and Table A.7 for a a list of medicinal plants included in the study) (Dragos and Gilca, n.d.). In PhytoMolecularTasteDB the global and individual relationships between various chemical classes of phytocompounds and EPAs, and between various taste-based classes of phytocoumpounds and EPAs were explored. For the distribution of medicinal plants, which are included in this database, among the various botanical families, the reader is referred to Data in Brief (Dragos and Gilca, n.d.). The first null hypothesis (H0) tested in this study was: there is NO correlation between the active principles found in plants and their EPAs. It was tested against the alternative hypothesis (H1) that there is a correlation between the active principles found in plants and their EPAs. Chi-squared test was used to sort this out. For reasons detailed in Results section a second null hypothesis (H0) was afterwards put forward: there is NO correlation between the tastes of the active principles found in plants and their EPAs. This null hyptothesis was tested against the alternative hypothesis (H1): there is a correlation between the tastes of the active principles found in plants and their EPAs. Again Chi-squared test was used to clarify this issue. The results obtained while testing these two hypotheses raised a third work hypothesis: chemical categories with the same taste should have similar spectra of EPAs, while chemical categories with different tastes should have different spectra of EPAs. Yet again Chi-squared test was applied to pairs of chemical classes having either different, or similar tastes. For the sake of statistical correctness, the pairs of chemical categories should ideally fulfill the following four conditions: 1. Each category in the pair should have a definite taste (such as astringent for tannins). 2. Each category in the pair should be as narrow and as chemically-well-defined as possible (broad categories such as alkaloids, monoterpenes, sugars or lipids can hardly be accepted). 3. Each category in the pair should be sufficiently prevalent among the studied plants to make the statistical calculations relevant. 4. The two categories in the pair should not be associated among the studied plants, either positively (the plants containing compounds from one category frequently/ usually contain also compounds from the other category) or negatively (the plants containing compounds from one category frequently/ usually do not contain compounds from the other category) (actually factor analysis may be a way around this problem, albeit an involved one). Practically no pair of chemical categories fulfills all these criteria mainly because the narrower a chemical category, the less prevalent, criteria 2 and 3 being inherently antithetical. 2.1. Global association between active compounds and EPAs Regarding the EPAs, there are about 1000 of sanskrit words used in Pandey's book to describe the EPAs, many of them being synonyms with one another. This number was cut down (to a manageable one of about 100) using two criteria: prevalence and synonymy. First the words were ordered in a list (the EPAs list) according to their prevalence (i.e. the number of plants having the respective EPA). The EPAs list was then taken term by term (i.e. EPA by EPA) in a top to bottom order, picking for the term in each given position (the current term) the other terms in the list in a relation of perfect synonymy (from a linguistic and/or ayurvedic/practical point of view) with the current one. These synonyms situated lower in the list were moved up to the position of the current term (the highest placed one, hence the most prevalent one). At this stage the EPAs list contained in many positions not a single term but a class of synonymous terms. For each given class of synonymy the most prevalent term was used to replace, in the database, the other terms in the given synonymy class. The replaced terms were eliminated from the EPAs list. Then the prevalences were calculated again and the EPAs list was reordered based on the new prevalences. Only the EPAs with a prevalence of at least 10 (i.e. the given EPA was attributed to at least 10 plants) were considered for the statistical analysis (Table 1). Wherever there was a class of synonymous terms, the total prevalence for the entire class (i.e. the number of plants having at least one of the EPA in the synonymy class) was considered. Table 1. Ayurvedic ethnopharmacological actions and their prevalences (the number of plants to which that EPA is attributed in Pandey's Dravyaguna Vijnana). Ethnopharmacological action – Sanskrit term [prevalence] English translation aksepahara [16], aksepasamaka [3], aksepasamana [4] anticonvulsant amapacana [27], amapacaka [2] detoxifying by digesting/metabolising toxins anulomana [121], vatanulomana [34] carminative and aperient arsoghna [50] antihemorrhoidal artavajanana [53] emmenagogue asmaribhedana [14], asmarinasana [8], asmarihara [2], asmarighna [3] litothriptic balya [148], balavardhaka [2] body strengthening, tonic bhedana [27] breaking up obstructions in general, piercing, disintegrating dosas/humors and malas bhutaghna [5], bhutahara [1], bhutanasana [1] psychiatric reliever brmhana [51] nourishing, weight promotion caksusya [51] beneficial for eyes, promoter of eyesight chardinigrahana [34], chardighna [4] antiemetic chedana [11] scarificant (“removing the adherent humors and wastes from their places”) (Murthy, 2006) dahaprasamana [102], dahahara [2], dahaprasadana [1], dahasamaka [7] relieves burning sensation (which is associated or not with high body temperature), refrigerant dantya [13] beneficial for teeth dipana [211], agnidipana [2], agneya [1], agnivardhana [1] kindling Agni (digestive fire), digestive tonic dourgandhyahara [13], durgandhahara [2], durgandhanasana [12] removing bad smell garbhasayasankocaka [25] induces uterine contractions or miscarriage garbhasayottejaka [18] uterine stimulant, activates the uterine functions and metabolism grahi [108], sangrahi [12], sangrahaka [1], sangrahika [1] antidiarrhoeal hikkanigrahana [18] anti-hiccups hrdayottejaka [40] cardiostimulant, anti-bradycardic hrdya [109], hrdaya [7] beneficial for heart, harmonises the cardiac functions and metabolism jantughna [47], jivanunisudana [6] antimicrobial, vermicide jvaraghna [200], jvarahara [5], jvaranasaka [1], jvarapratisedhaka [1], jvaraprativandhaka [1] febrifuge, antipyretic kandughna [72] anti-pruritic kanthya [23] beneficial for throat (e.g. local anti-inflammatory) kaphaghna [94], kaphahara [2], kaphanasana [1], slesmaghna [1], slesmahara [17] antikapha (kapha is a body humour corresponding to mucus), antiphlegmatic, antimucus kaphanihsaraka [77], kaphanihsarka [1], kaphanissaraka [1], slesmanihsaraka [2] expectorant kasahara [37], kasaghna [32], kasagara [1] antitussive, anti-cough katupaustika [49], katupoustika [10] tonic (pungent), invigorating kesya [32], keshya [1], kesavardhana [3] beneficial for hair, hair tonic kothaprasamana [12] relieving a type of skin rash krmighna [172], krmihara [2] antimicrobial, antiparasitic, vermicide kusthaghna [159] antidermatosis lekhana [55] scraping (“removing/expelling/dissolving the tissues and wastes after drying up their moisture” (Murthy, 2006), e.g. anti-obesity) madaka [12], madakara [1], madakari [6] narcotic mastiskabalya [12] brain tonic, central nervous tonic or nervine mastiskasamaka [13] sedative/calming (central mechanism) medhya [47], medya [1] nootropic, memory enhancer, intellect enhancer medohara [24], medonasaka [1], medodhatuksayakara [1], medososana [1], medasosana [1], medapanayana [1], karsana [3] anti-obesity mrdurecana [8], mrduvirecana [3], sukhavirecana [4] mild/light cathartic mutrala [128], mutravirecaniya [12], mutrajanana [60], mutrapravrtaka [1], mutrasravottejaka [1] promoting diuresis, diuretic mutrasangrahaniya [29], mutrastambhana [1] antidiuretic nadibalya [43] peripheral nervous tonic or nervine nadisamaka [5], nadisamana [1], naditantusamaka [1] sedative/ calming (peripheral mechanism) nadyuttejaka [11] nervous stimulant nidrajanana [13] hypnotic, sleep promoting pacana [117], pachan [1], pacaka [2] digestive stimulant pittasamaka [25], pittaprasamana [3] Pitta pacifying/ balancing (Note. Pitta is a humoral entity or dosa, corresponding to the bile.) (Sharma and Dash, 2006) pittasaraka [38], pittavirecana [2], pittarecaka [1], pittasamsravaka [1], pittasravajanana [2], pittasravakara [1] causing Pitta to flow, choleretic-cholagogue prajasthapana [20], garbhasthapana [14] embrio stabilizer pramehaghna [23], pramehahara [2], mehaghna [7], mehahara [4] anti-prameha (prameha = urinary diseases with polyuria, pollakiuria and turbid urination, including diabetes mellitus) putihara [9], putighna [1] antiseptic, removing pus raktapittasamaka [45], raktapittaprasamana [1], raktapittanasaka [1], raktapittahara [10] raktapitta (bleeding-provoking diseases) pacifying/ balancing/ reducing/curing raktaprasadana [40] antianemia raktasodhaka [116], raktasodhana [8] blood-purifying raktastambhana [64], raktastambhaka [6], raktarodhaka [18], raktarodhana [1], raktasravarodhaka [1], raktasravahara [1], raktavahinisankocaka [1], sonitasthapana [21] hemostatic rasayana [63], rasanyana [1], vayahsthapana [6] antiageing recana [28], recaka [14], virecaka [2], virecana [8] cathartics (“which makes the feces watery and expel them out forcibly either formed or not formed into a mass”)(Murthy, 2006) rocana [74], rocaka [1], rucikrt [1], rucikara [7], ruciprada [1], rucya [8], rucivardhaka [8], rucivardhana [5] anti-anorexia, increases appetite samaka [18], samana [2], prasadana [2], prasamana [3] calming, pacifying, balancing sandhaniya [30] unifies the broken parts, causing to grow together (e.g. broken bones, wounds), fracture healing saraka [19] purgative/laxative sirovirecana [11] errhine, promote sneezing and increased discharges from the nose, clears toxins from the head sitajvarahara [1] febrifuge (in cold fever or sitajvara) sitaprasamana [18] relieves cold sensation snehana [38] oleating, lubrifying sodhana [7], sodhaka [7], sodhita [2] purification, elimination of toxin, detoxifying sothahara [205], sothaghna [4], sothanasana [1], sothaprasamana [1], sothasamaka [2] anti-inflammatory, antiswelling soumanasyajanana [10], tarpana [6], prinana [1], pustiprada [1] produces goodness in mind, satisfying, satiating, pleasing, nourishing sramahara [10], sramaha [1], sramaghna [1] relieves exhaustion and fatigue, antiasthenia stambhana [74], stambhaka [4] restraining, arresting/blocking the excessive flow (e.g. exudate, heavy periods, diarrhea, etc), preventing the leakage of fluids/blood or mobile components of the body (e.g. feces) stanyajanana [32] galactogogue stanyasodhana [11], stanyasodhaka [5] milk purifier sukrajanana [14], sukrala [13], sukrakara [1], sukravardhaka [2], sukravardhana [2] spermatopoetic, sperm producing sukrasodhana [10], sukrasodhaka [1] sperm purifier, sperm antiseptic sukrastambhana [13], sukrastambhaka [3] antiejaculation, delaying ejaculation, including anti-nocturnal emissions sulaprasamana [48], sulahara [14], sulaghna [1] alleviates colicky pain, antispasmodic svarya [13] promoter of good voice, good for throat svasahara [53], svasagara [1], svasaghna [3] antiasthmatic svedajanana [71], svedala [1], svedana [2] diaphoretic svedapanayana [10] antidiaphoretic tiksnavirecana [1], tivrarecana [1], tivravirecaka [1] very strong cathartic trsnanigrahana [53], trsnaprasamana [4], trsaprasamana [2], trsnahara [2], trsahara [4], trsnasamaka [7] antithirst, antipolydipsia tvacya [47] beneficial for skin tvagdosahara [31] reduces the humors (sanskr. dosas) accumulated in the skin uttejaka [41] stimulant vajikarana [45], bajikarana [8], vajikara [1] aphrodisiac which increases sexual drive, sexual pleasure (in both sexes), erectile performance vamaka [37], vamana [1], vantikara [1], vantikrta [1], chardikara [1] emetic varnya [56], varnakara [1], varnasanjanana [3] giving skin colour, promoting complexion vataghna [28], vatahara [16] anti-vata (vata is a humour corresponding with bodily air or gases) vatasamaka [14] vata pacifying/ balancing vedanasthapana [150], vedanisthapana [1], vedanahara [4], vedanasamaka [3] pain reliever, antalgic, anodyne vidahi [12] induce burning sensation, irritation visaghna [149], visahara [1] antitoxic visamajvaraghna [12], visamajvarahara [1], visamajvaraprati [2], visamajvarapratibandhaka [1], visamajvaraprativandhaka [2] febrifuge in periodic fever, reducing antimalarial fever vranaropana [100], ropana [18], vranahara [1], vranarohana [1] unifies the broken parts, causing to grow together- subcategorie, wound healer, antiulcer vranasodhana [37], vranasodhaka [1], vranavisodhana [1] wound purifier, wound antiseptic, wound cleansing vrsya [92] aphrodisiac which increases fertility (increases the amount and quality of sperm and ovum) yakrduttejaka [68] hepatostimulant Note. The numbers in the square brackets indicate the prevalences, i.e. the number of plants to which that EPA is attributed in Pandey's Dravyaguna Vijnana. The terms in each cell of the first column are (if more than one) synonymous from a linguistic and/or ayurvedic point of view (Marilena Gilca and Barbulescu, 2015a, Gilca and Barbulescu, 2015b; Monier-Williams, 2002; Murthy, 2006, Murthy, 2004; Pandey, 2005; Sharma and Chandola, 2011; Sharma and Dash, 2006). For each class of synonyms, the term with the highest prevalence is bold-typed. Using Excel, a database was constructed, consisting in a table with the names of the plants as row titles and several columns including one containing the EPAs and one containing the active principles. The vast number of active principles (more than 1400) would have made a statistical analysis, for all practical purposes, impossible. Therefore this number was cut down by using categories of active principles (223 in number – see Appendix A, Table A.1) such as flavones, pentacyclic triterpenes, C6-C3 compounds, monoterpene aldehydes, etc. A further column was constructed including these categories of active compounds. This column and the column containing the EPAs were used to build a contingency table having the EPAs as row titles and the categories of compounds as column titles (Table 2). Table 2. Ethnopharmacological actions by categories of compounds contingency table (observed data table). ChemCateg1 ChemCateg2 … ChemCategj … ChemCategn RMS EPA1 o11 o12 … o1j … o1n RMS1 EPA2 o21 o22 … o2j … o2n RMS2 … … … … EPAi oi1 oi2 … oij … oin RMSi … … … … … … … … EPAm om1 om2 … omj … omn RMSm CMS CMS1 CMS2 … CMSj … CMSn TS Legend: EPA = ethnopharmacological action, ChemCateg = category of compounds, RMS = row marginal sum, CMS = column marginal sum, oij = the number of plants containing ChemCategj that exerted EPAi. However at this stage there was a large degree of overlap between the various categories of active principles and one of the conditions that must be met in order to apply Chi-squared test is that the columns should be independent of each other (and also the rows). Consequently a drastic elimination operation was performed at the end of which the number of categories of compounds dropped to less than half: 109 (see Appendix A, Table A.2.). The above mentioned contingency table (Table 2) contained m rows (EPA1 to EPAm) and n columns (ChemCateg1 to ChemCategn). An entry oij in a given cell (i,j) (i.e. the cell situated at the intersection of row i and column j) of this table signified the number of plants that both exerted the EPA representing the title of the row i (EPAi) and contained a compound in the category representing the title of the column j (ChemCategj). The letter o is used to denote these entries because these are the observed values, by contrast to the expected values (EVs) (denoted by the letter e) which are calculated on the assumption that there is no correlation between the EPAs and the active principles (H0). A marginal sum (MS) was calculated for each row (RMSi for the ith row) and for each column (CMSj for the jth column) equal to the sum of all the elements in the corresponding row or column, respectively. The total sum (TS) is calculated by summing up all the RMS-s (or all the CMS-s). (1) (2) (3) The aim was to establish whether the numbers in the cells of this contingency table are either randomly distributed (which would mean that there is no correlation between the titles of the rows and the titles of the columns – H0) or they follow a certain pattern (tipping the balance in favour of the alternative hypothesis, H1). A random distribution would have meant that most of the entries in the cells of the table were approximately proportional to the marginal sums [the entry oij in the cell (i,j) would have been proportional both to the sum of the entries in the row i (row i marginal sum = RMSi) and to the sum of the entries in the column j (column j marginal sum = CMSj) and therefore to the product of these sums]. In order to decide among these two hypotheses it was necessary to compare the observed data table (ODT) with an EVs table (EVT) built upon the hypothesis of a complete lack of correlation between the EPAs and the ChemCateg-s (Table 3). Consequently the entry in the cell (i,j) in these EVT was the product of marginal sums divided by the total sum (4) Table 3. Expected values table. ChemCateg1 ChemCateg2 … ChemCategj … ChemCategn EPA1 e11 e12 … e1j … e1n EPA2 e21 e22 … e2j … e2n … … … EPAi ei1 ei2 … eij … ein … … … … … … … EPAm em1 em2 … emj … emn Legend: EPA = ethnopharmacological action, ChemCateg = category of compounds, eij = the expected number of plants containing ChemCategj that exerted EPAi; given there were no association between the ChemCateg-s and EPAs. The Chi-squared test calculates how far the observed values oij are from the EVs eij: the further they are, the less probable the randomness hypothesis is. The distance between a given observed value (oij) and its corresponding EV (eij) is estimated by the difference of the two (oij−eij), but as this difference may be positive or negative, it is squared: (oij−eij)2. It is also divided by the observed value [(oij−eij)2/oij] as the magnitude of a certain difference is relevant only by comparing it to the value from which it marks the difference. Actually the Chi-squared statistics (χ2) is a weighted sum: the sum of the squared differences of the expected and observed values weighted by the inverse of the observed values (Altman, 1991): (5) The value of the Chi-squared statistics is compared against the number of degrees of freedom (df) (the number of values that can vary independently once the marginal sums are fixed), calculated as the number of rows minus 1 (m−1) times the number of columns minus 1 (6) Traditionally, the significance level is usually established at 5% (α = 0.05), meaning that in order to be statistically significant χ2 should be greater than the 95th percentile of the Chi-squared distribution for the given df. Should this be the case, H0 should be rejected and H1 accepted. Otherwise H1 is to be rejected and H0 accepted. However, given the surprising and counter-intuitive nature of the main conclusion of this paper, we decided to lower the significance level by one order of magnitude (to 0.005), in compliance with the recommendations in a recent paper (Benjamin et al., 2018). Chi-squared test can be applied provided no more than 20% of the EVs are less than 5 and no individual EV is less than 1 (Yates et al., 1999) – these two conditions were strictly observed while performing the statistical calculations. All the computations were done with Excel. Specifically the Chi-squared test was performed by means of the statistical function CHISQ.TEST which provides practically unlimited precision in calculating the probability of error p (p value). The following table (Table 4) summarizes in a flowchart the stages of these calculations. Table 4. Table summarizing the stages of calculating Chi-squared statistics. process equations employed (if any) results obtained For each pair of EPA-chemical category (ChemCateg), count (in our database) the plants that both exerted the EPA and contained a compound in ChemCateg. The resulting numbers are called the observed values. observed values (oij) Image 2 Put the observed values in a table with EPAs as row titles and ChemCateg-s as column titles observed data table (Table 2) Image 2 calculate the row- and column-marginal sums (RMS and CMS, respectively) Eqs. (1), (2) RMS-s, CMS-s in Table 2 Image 2 calculate the total sum (TS) Eq. (3) Image 2 calculate the expected values Eq. (4) expected values (eij) Image 2 Put the expected values in a table with EPAs as row titles and ChemCateg-s as column titles expected data table (Table 3) Image 2 Calculate chi-squared statistics Eq. (5) χ2 Image 2 Calculate the degrees of freedom Eq. (6) df Image 2 Compare χ2 with the 95th percentile of the Chi-squared distribution for the given df (can be found in tables or can be calculated with various programs such as Excel). Accepting or rejecting the null hypothesis 2.2. Individual associations between categories of active principles and EPAs The method used for calculating these associations is detailed below, in the Section 2.3. The very large number of calculations (94 EPAs × 109 categories of active principles = 10246) imposed a large correction factor (in accordance with Bonferroni criterion, as explained in Section 2.3.). Consequently the level of statistical significance was lowered to 0.05/10246 ≈ 5E-06. The statistical significance was assessed using Fisher's exact test. The instruments used for performing the statistical calculations are described in Section 2.3. 2.3. Global association between tastes and EPAs The hypothesis that EPAs may be associated with the herbal “molecular taste” was tested. The herbal “molecular taste” represents the combination of tastes derived from all the principal constituent tastants of a certain medicinal plant. The analysis was performed for the six rasas or ayur-tastes (the oral sensations recognized by Ayurveda): sweet (madhura), salty (lavana), sour (amla), bitter (tikta), pungent (katu), astringent (kashaya). Consequently a new work hypothesis (H1) was formulated: there is a correlation between the tastes of the active principles found in plants and their EPAs. This alternative hypothesis (H1) was tested against the null hypothesis (H0) that there is NO correlation between the tastes of the active principles found in plants and their EPAs. Given the lack of relevant data in literature, in PhytoMolecularTasteDB, a taste/orosensation was assigned to 438 active principles among the total number of more than 1400. As for the leftovers, 8 of them were found in only 3 plants, 101 in only 2 plants, and the rest in only one plant. Because the present study explores the ayurvedic knowledge from a modern scientific perspective, only the six rasas or ayur-tastes (sweet/ madhura, sour/ amla, bitter/ tikta, astringent/ kashaya, salty/ lavana, pungent/ katu) recognized in Ayurveda were taken into consideration. Consequently, volatile substances were considered ayur-pungent (katu) (since all the aromatic plants are pungent in ayurvedic texts) (Gogte, 2000, Pandey, 2005), together with the active principles known to impart a burning orosensation (such as those found in pepper, mustard, chili peppers, etc.). These tastes/orosensations were afterwards attributed to the plants containing the corresponding active principles (if a plant P contained an active compound C having the taste T, than the taste T was attributed to the plant P) and investigated whether there is a correlation between the tastes of the plants and their EPAs by means of Chi-squared test. In the first stage, a 94 × 6 contingency table was built with 94 EPAs as row titles and the 6 ayurvedic rasas (ayur-tastes) as column titles (ODT). Following a similar course of action as previously described, the row and column marginal sums (RMS and CMS, respectively) were calculated using Eqs. (1), (2), the total sum (TS) was calculated using Eq. (3), and then the EVs were calculated using Eq. (4). The resultant EVs were used to build the EVT. At this stage, the proportion of less than 5 EVs (LT5EVs) was 25%. In order to decrease this proportion, one strategy would have been to eliminate the rows with too many LT5EVs applying a decreasing threshold, namely to progressively eliminate the rows with 5, 4, 3, etc. LT5EVs until a less than 20% proportion of LT5EVs would have been reached, while losing maybe important EPAs in the process. Nonetheless a quick glance at the EVT revealed that the CMS for the salty taste was far less than the CMSs for the other tastes, pointing out the salty taste as probably the main culprit for the high proportion of LT5EVs. Indeed, there were 74 (out of 94) LT5EVs in the salty taste column, compared to 30, 19, 9, 9, in the sour, astringent, ayur-pungent, ayur-sweet, and bitter columns, respectively. Therefore the salty taste column was eliminated and the proportion of LT5EVs shrunk to 14.25% enabling the calculation of the Chi-squared statistics by applying Eq. (5). 2.4. Individual associations between tastes and EPAs A 2 × 2 contingency table was built for each EPA-taste pair and the strength of the association between the two (say taste TASi and EPA EPAj) was evaluated using Odds Ratio (OR), calculated by means of the formula (Altman, 1991): a = the number of plants having TASi and exerting EPAj; b = the number of plants having TASi but NOT exerting EPAj; c = the number of plants NOT having TASi but exerting EPAj; d = the number of plants NOT having TASi and NOT exerting EPAj; An OR > 1 (positive ln OR) meant that the plants having TASi are more likely to exert EPAj. An OR < 1 (negative ln OR) meant that the plants having TASi are less likely to exert EPAj. The problem of zero count cells was solved by adding a fixed value of 0.5 to each of these cells (Higgins and Green, 2011). This solution was preferred because the method recommended by Sweeting et al. for studies with unbalanced arm sizes [to include a correction proportional to the reciprocal of the size of the contrasting study arm (J. Sweeting et al., 2004)] yielded huge, unrealistic values for OR. The 95% confidence limits for OR (95% CL) were calculated according to Eqs. (7), (8) (Altman, 1991): (7) (8) where standard error of the natural logarithm of OR (9) A further criterion for statistical significance is that the 95% CL should not include unity (Altman, 1991). A search was performed for associations between the five remaining rasas (ayur-sweet, ayur-pungent, bitter, sour, astringent) and the EPAs, which meant a total of 470 calculations (94 EPAs × 5 ayurvedic tastes). Given the large number of calculations (470) made, the significance level had to be lowered according to Bonferroni criterion (Fleiss, 1986) to αcorr = 0.05/470 ≈ 0,0001 = 1E-4. Consequently the significance threshold was set at 1E-4. However, an inspection of the associations with a p value between this 1E-4 threshold and the commonly employed 0.05 threshold pointed out many traditionally recognized taste-EPA associations. Therefore, for fear of throwing the baby out together with the bathwater (i.e. making a type II error) (Altman, 1991), all the associations with a p value less than 0.05 were presented in the results section, highlighting those observing Bonferroni criterion. In order to assess the statistical significance of these associations an online available spreadsheet was used [http://www.biostathandbook.com/fishers.html in the online version of (McDonald, 2014)], which calculates the Fisher's exact test for 2 × 2 tables. The spreadsheet may be downloaded and allows automated data entry – consequently the necessary calculations were rapidly performed, granting, a good idea about the level of statistical significance. The results were afterwards ordered ascendingly according to the p-values and those with a p-value below 0.05 were checked by means of GraphPad InStat software package (GraphPad InStat software package, 2017) (GraphPad Software Inc., 2017, https://graphpad.com/quickcalcs/contingency2/). As GraphPad InStat software package has a limitation in the precision with which it calculates the p value (below 0.0001 it does not calculate it precisely, but only states that the p value is < 0.0001) the calculations were double-checked with Fisher's exact test Calculator in the Social Sciences Statistics software package (Social Sciences Statistics software package, 2017) (http://www.socscistatistics.com/tests/fisher/Default2.aspx) which precisely calculates the p value down to 1E-6. Fortunately this threshold was enough for the purposes of this study, as it is below 1E-4, the accepted significance level for αcorr. The results provided by these two software packages were in perfect agreement. For the sake of statistical safety a third software package was also used: VassarStats software package (http://vassarstats.net) Website for Statistical Computation, specifically two-tailed Fisher Exact Probability Test on the page “For a 2 × 2 Contingency Table: ” (http://vassarstats.net/table2×2.html) (Lowry, 2017) which allows for practically unlimited precision in the calculation of p value. As the p values calculated with McDonald's spreadsheet were in perfect agreement with those calculated with GraphPad and Fisher's exact test Calculator in the Social Sciences Statistics software package down to 0.0001 and 1E-6, respectively, only the p values below 1E-6 were triple-checked (with VassarStats software package) and an almost perfect agreement was found with McDonald's spreadsheet regarding the order of magnitude, with some differences in the precise digits in the various decimal places. Nevertheless these differences bore no influence on p-value's order of magnitude and in most cases were wiped out by rounding to the first significant figure. Rouding p value to the first non-zero digit is not only legitimate but also encouraged, as spurious precision in reporting statistical results conveys no real information and is generally not recommended (Altman, 1991). Finally only the associations with a less than 5% p value (as calculated by all three software packages) were taken into account. The following tables are meant to ease the understanding of the way the statistical calculations were performed, starting with a statistics overview table summarizing the various statistical calculations performed in this study (Table 5), and continuing with several tables (Table 6, Table 7, Table 8, Table 9) concretely exemplifying the tables on which the statistical calculations were done. Table 5. Overview table summarizing the various statistical calculations. Section/title Objective/aim Method Titles of the rows Titles of the columns significance of the number in a given cell of the table 2.1. Global association between categories of active principles and EPAs Is there an association between chemical classes of the active principles in the medicinal plants and the EPAs of these plants? one chi-square test performed on a table having the EPAs as rows and the chemical classes as columns EPAs chemical classes number of plants exerting the corresponding EPA (the title of the row) and containing phytocompounds from the corresponding chemical class (the title of the column) 2.2. Individual associations between categories of active principles and EPAs Is there an association between a given chemical class of active principles in the medicinal plants and a given EPA of these plants? 10246 (94 ×109) Fisher's exact tests, each performed on a table with 2 rows and 2 columns chemical class present, chemical class absent EPA present, EPAabsent number of plants exerting/ not exerting the corresponding EPA and containing/not containing phytocompounds from the corresponding chemical class 2.3. Global association between tastes and EPAs Is there an association between the EPAs of the medicinal plants and the tastes of the active principles in these plants? one chi-square test performed on a 94 × 6 table having the 94 EPAs as rows and the 6 tastes as columns EPAs tastes number of plants exerting the corresponding EPA (the title of the row) and containing phytocompounds with the corresponding taste (the title of the column) 2.4. Individual associations between tastes and EPAs Is there an association between a given taste of the active principles in the medicinal plants and a given EPA of these plants? 470 (5 ×94) Fisher's exact tests, each performed on a 2 × 2 Table (5 refers to the number of tastes after eliminating salty/ lavana for reasons explained in the text of the article) taste present, taste absent EPA present, EPA absent number of plants exerting/ not exerting the corresponding EPA and having/ not having the corresponding taste Table 6. The table used in Section 2.1 in order to establish (by chi-squared test) whether there is a global association between EPAs and chemical classes. obsValues acridoneAlkaloid aliphAldehyde … withanolide xanthone sumOnKarma #LT5EV aksepahara 1 0 … 0 1 84 107 amapacana 0 0 … 0 2 98 107 … … … … … … … … vrsya 0 0 … 0 3 326 89 yakrduttejaka 0 0 … 0 4 256 96 sumOnChemCateg 14 28 … 31 194 18569 Legend. #LT5EV- number of less than 5 expected values. Table 7. Example of table used to determine (by Fisher's exact test) whether there is an (individual) association between stambhana (an EPA) and the content in tannins. obsValues #Plants exerting stambhana #Plants not exerting stambhana #plants containing tannins 37 65 #plants not containing tannins 37 292 Legend. #plants- number of plants. Table 8. The table used in Section 2.3 in order to establish (by chi-squared test) whether there is a global association between EPAs and tastes. obsValues Sour Bitter Astringent Sweet ay_katu sumOnKarma #LT5EV aksepahara 5 19 7 11 17 59 0 amapacana 7 26 7 9 10 59 0 … … … … … … … … vrsya 20 74 26 43 28 191 0 yakrduttejaka 14 58 17 25 33 147 0 sumOnTaste 1039 4540 1455 2110 2114 11258 Legend. #LT5EV- number of less than 5 expected values. Table 9. Example of table used to determine (by Fisher's exact test) whether there is an (individual) association between dourgandhyahara (an EPA) and the katu rasa. obsValues #Plants exerting dourgandhyahara #Plants not exerting dourgandhyahara #plants tasting katu 67 117 #plants not tasting katu 2 250 Legend. #plants- number of plants. The next table (Table 7) is one of the 10,246 tables used in Section 2.2 in order to establish whether there are associations between the individual EPAs and the individual chemical classes. The next table (Table 9) is one of the 470 tables used in Section 2.4 in order to establish whether there are associations between the individual EPAs and the individual tastes. 3. Results 3.1. Global association between active compounds and EPAs Initially, 94 EPAs and 109 categories of active compounds were assessed. At this stage the proportion of less-than-5-expected-values (LT5EVs) was 91.4% (in that many compounds were present in only a few plants), far above the acceptable 20% limit. There are two strategies to decrease the proportion of LT5EVs: pooling together columns (or rows) or eliminating columns (or rows). Applying the first strategy to rows would result in pooling together different EPAs (which is not justified from a medical standpoint), while application to columns is warranted only if pooled together chemical categories have similar EPAs spectra – an arbitrary supposition undermining the very aim of this study, which is to determine whether an association between categories of active principles and EPAs exists. Hence, the second strategy (that eliminated the need for injudicious presumptions) was used. This second strategy may be applied in three variants: 1) deleting mostly columns (i.e. chemical categories), or 2) deleting mostly rows (i.e. EPAs), or 3) something in between, a middle path. All three were assessed. 1. Initially the less prevalent categories of compounds were eliminated. This was accomplished by establishing a threshold for the column marginal sums (CMS): all the columns with a CMS less than the established threshold were eliminated. A threshold of 100 was initially employed and then successively increased to 300 and to 500, while the proportion of LT5EVs decreased to 79.7%, 59.63%, and 41.76%, respectively. Meanwhile the number of categories of active principles decreased to 46 (see Appendix A, Table A.3), 18 (see Appendix A, Tables A.4), and 8 (see Appendix A, Table A.5), respectively. As the risk arose of running out of biochemical categories by following this course of action, the EPAs were then considered, as some of them were also scantily represented among the plants. This time, the elimination criterion was the number of LT5EVs in the corresponding row in the EVT. A threshold of 7 (out of 8 possible) was intially employed: all the rows with more than 7 LT5EVs (in the corresponding row in the EVT) were eliminated. A promising but, alas, insufficient 22,89% proportion of LT5EVs was obtained. As there were no rows in the EVT with 7 LT5EVs, the threshold was lowered to 5, where a comfortable 14.34% was reached (at which stage there were still 61 EPAs left), allowing the application of Chi-squared test. Unfortunately, a χ2 = 448.7 was obtained, which was less than 468,7 the value necessary to obtain a 5% significance level with df = (61 −1)× (8 −1) = 420 (actually the p value corresponding to χ2 = 448.7 for df = 420 is 0.16). 2. Next, rows were considered and the departure point was the table with 49 chemical categories (obtained after the elimination of the columns with a CMS less than 100). As above, the elimination criterion was the number of LT5EVs. A threshold of 40 (out of 46 posssible) was initially established, and afterwards lowered progressively to 25, 15, and 14, while the proportion of LT5EVs slowly decreased to 62.0%, 36.7%, 20.8%, and finally to 16.3%. The number of EPAs also decreased to 46, 17, 7, and 6 – a drastic and unacceptable decrease. Nevertheless the Chi-squared statistics was calculated, yielding a χ2 = 122.73, which is far less than 260.9 the value necessary to obtain a 5% significance level with df = (46−1) × (6−1) = 225 (actually the p value corresponding to χ2 = 122.73 for df = 225 is 1). 3. For the middle path, initially 18 chemical categories (the result of eliminating the columns with a CMS of less than 300) were considered, followed by deletion of rows using the number of LT5EVs as the elimination criterion. A threshold of 15 (out of 18 posssible) was initially employed, and then lowered to 10, while the proportion of LT5EVs decreased to 37.95% and 19.23%, respectively and the number of EPAs shrunk to 59 and 39, respectively. Applying the Chi-squared test a χ2 = 673.47 was obtained, which is less than 706.2 the value necessary to obtain a 5% significance level with df = (39−1) × (18−1) = 646 (actually the p value corresponding to χ2 = 673.47 for df = 646 is 0.22). 4. No matter which course of action was chosen, the result was the same: the chi-squared test failed to reach statistical significance. Therefore the null hypothesis had to be accepted: there is no correlation between the scientifically documented chemical categories of active principles and the the EPAs. 3.2. Individual associations between active principles and EPAs Although the calculations failed to prove a global association between the chemical classes of active principles and EPAs, a number of individual associations were identified by means of Fisher's exact test (Table 10). Table 10. Associations between chemical classes of active principles and the ethnopharmacological activities fulfilling Bonferroni criterion for statistical significance. chemical class EPA a/b/c/d p OR ln OR 95% CL tannin stambhana 37/65/37/292 4E- 08 4,49 1,5 2,65-7,63 tannin raktastambhana 44/58/58/271 5E- 07 3,54 1,27 2,19-5,75 tannin mutrasangrahaniya 18/84/13/316 2E- 05 5,21 1,65 2,45-11,06 monoterpene hydrocarbons dipana 26/3/193/209 9E- 06 9,39 2,24 2,8–31,51 monoterpene hydrocarbons dourgandhyahara 9/20/19/383 2E- 05 9,07 2,21 3,65-22,57 monoterpene alcohols dourgandhyahara 8/13/20/390 1E- 05 12 2,48 4,46-32,25 monoterpene phenols dourgandhyahara 7/10/21/393 3E- 05 13,1 2,57 4,53-37,85 short aliphatic acids rocana 19/14/82/316 1E- 05 5,23 1,65 2,52-10,87 glucosinolates vidahi 4/5/8/414 4E- 05 41,4 3,72 9,34-183,58 gums vatasamaka 6/18/10/397 9E- 05 13,23 2,58 4,33-40,43 Legend. For a given category of compounds ChemCateg and a given EPA: a = the number of plants containing ChemCateg and exerting EPA; b = the number of plants containing ChemCateg but NOT exerting EPA; c = the number of plants NOT containing ChemCateg but exerting EPA; d = the number of plants NOT containing ChemCateg and NOT exerting EPA; OR = odds ratio; ln OR = natural logarithm of OR; 95% CL = 95% confidence limits for OR; p = p value calculated with VassarStats (Lowry, 2017). 3.3. Global association between tastes and EPAs When exploring the hypothesized association between the six (actually five after eliminating salty/lavana) rasas (ayur-tastes) and the EPAs by means of Chi-squared test, a χ2 = 442.00 was obtained which is more than 417.97 the value necessary to obtain a 0.05 significance level with df = (94 − 1) × (5 − 1) = 372, but less than 446.00 the value demanded for a 0.005 significance level. Actually the probability of error was 0.007 and consequently, in compliance with the recent recommendations (Benjamin et al., 2018), the result obtained should not be called significant, but only suggestive for the conclusion that there is a correlation between the ayurvedic tastes of the active principles found in plants and their EPAs. 3.4. Individual associations between tastes and EPAs Table 11, Table 12, Table 13, Table 14, Table 15 contain the associations between tastes/orosensations and EPAs with a p value at most 0.05 (as calculated by all three software packages). The associations that also met the Bonferroni criterion for statistical significance in the case of multiple comparisons (i.e. a p value <1E-4) are bold-typed to prominence. The associations significant in accordance with the recent recommendations (Benjamin et al., 2018) are italicized. For the vast majority of the associations selected on the basis of p value, the 95% CL did not include unity. The Appendix B contains some supplementary material for each of the tastes/ orosensations, including a table with the English translation of the sanskrit words denoting the EPAs positively associated with the respective taste, followed by a list of the herbs containing compounds with the respective taste. Table 11. Associations between sour (amla) phytochemicals and the ethnopharmacological activities. EPA a/b /c/d p OR ln OR 95% CL POSITIVE ASSOCIATIONS mastiskabalya 9/70/6/309 0,0006 6,62 1,89 2,28-19,21 dahaprasamana* * 30/49/69/246 0,005 2,18 0,78 1,29-3,7 sramahara 6/73/5/310 0,01 5,1 1,63 1,51-17,16 rocana 28/51/68/247 0,01 1,99 0,69 1,17-3,4 snehana 13/66/23/292 0,02 2,5 0,92 1,2-5,19 svarya 6/73/7/308 0,03 3,62 1,29 1,18-11,08 kesya* 11/68/20/295 0,03 2,39 0,87 1,09-5,21 stambhana 19/60/45/270 0,04 1,9 0,64 1,04-3,48 vatasamaka 6/73/8/307 0,04 3,15 1,15 1,06-9,37 trsnanigrahana 20/59/48/267 0,04 1,89 0,63 1,04-3,41 brmhana 15/64/33/282 0,05 2 0,69 1,03-3,91 NEGATIVE ASSOCIATIONS vamaka 2/77/31/284 0,04 0,24 − 1,44 0,06-1,02 recana 4/75/41/274 0,05 0,36 − 1,03 0,12-1,03 stanyasodhana* 0/79/16/299 0,05 0,12 − 2,13 0,01-2 Legend. a = the number of plants having sour tasting phytochemicals and exerting the corresponding ethnopharmacological activity (EPA); b = the number of plants having sour tasting phytochemicals but NOT exerting the corresponding EPA; c = the number of plants NOT having sour tasting phytochemicals but exerting the corresponding EPA; d = the number of plants NOT having sour tasting phytochemicals and NOT exerting the corresponding EPA; p = probability of error; OR = odds ratio; ln OR = natural logarithm of OR; 95% CL = 95% confidence limits for OR; * - association not mentioned in the traditional sources, * *- association in opposition with traditional sources. There was no association abiding by the Bonferroni criterion for statistical significance. The associations significant in accordance with the recent recommendations (Benjamin et al., 2018) are italicized. See also Appendix B, Table B.1 for the English translation of the sanskrit words denoting the positively associated EPAs, followed by a list of the herbs containing sour tasting compounds. Table 12. Associations between bitter (katu) phytochemicals and the ethnopharmacological activities. EPA a/b /c/d p OR ln OR 95% CL POSITIVE ASSOCIATIONS sothahara 177/174/13/30 0,01 2,35 0,85 1,18-4,65 recana 44/307/1/42 0,04 6,02 1,8 0,81-44,85 NEGATIVE ASSOCIATIONS trsnanigrahana 53/298/15/28 0,003 0,33 − 1,1 0,17-0,66 kothaprasamana* 6/345/5/38 0,003 0,13 − 2,02 0,04-0,45 vidahi 6/345/4/39 0,02 0,17 − 1,77 0,05-0,63 uttejaka* 32/319/9/34 0,03 0,38 − 0,97 0,17-0,86 Legend. a = the number of plants having bitter tasting phytochemicals and exerting the corresponding ethnopharmacological activity (EPA); b = the number of plants having bitter tasting phytochemicals but NOT exerting the corresponding EPA; c = the number of plants NOT having bitter tasting phytochemicals but exerting the corresponding EPA; d = the number of plants NOT having bitter tasting phytochemicals and NOT exerting the corresponding EPA; p = probability of error; OR = odds ratio; ln OR = natural logarithm of OR; 95% CL = 95% confidence limits for OR; *- association not mentioned in the traditional sources. None of the associations is statistically significant if Bonferroni criterion is taken into account. The associations significant in accordance with the recent recommendations (Benjamin et al., 2018) are italicized. See also Appendix B, Table B.2 for the English translation of the sanskrit words denoting the positively associated EPAs, followed by a list of the herbs containing bitter tasting compounds. Table 13. Associations between astringent (kashaya) phytochemicals and the ethnopharmacological activities. EPA a/b /c/d p OR ln OR 95% CL POSITIVE ASSOCIATIONS stambhana 31/82/33/248 0,0003 2,84 1,04 1,64-4,93 mutrasangrahaniya 16/97/12/269 0,002 3,7 1,31 1,69-8,1 snehana* * 16/97/20/261 0,03 2,15 0,77 1,07-4,32 raktastambhana 34/79/56/225 0,03 1,73 0,55 1,05-2,84 mastiskabalya* 8/105/7/274 0,04 2,98 1,09 1,06-8,43 kaphanihsaraka* * 31/82/51/230 0,05 1,7 0,53 1,02-2,85 NEGATIVE ASSOCIATIONS svedajanana 11/102/61/220 0,006 0,39 − 0,94 0,2-0,77 stanyasodhana 0/113/16/265 0,008 0,07 − 2,61 0-1,23 dipana 49/64/154/127 0,05 0,63 − 0,46 0,41-0,98 Legend. a = the number of plants having astringent phytochemicals and exerting the corresponding ethnopharmacological activity (EPA); b = the number of plants having astringent phytochemicals but NOT exerting the corresponding EPA; c = the number of plants NOT having astringent phytochemicals but exerting the corresponding EPA; d = the number of plants NOT having astringent phytochemicals and NOT exerting the corresponding EPA; p = probability of error; OR = odds ratio; ln OR = natural logarithm of OR; 95% CL = 95% confidence limits for OR; *- association not mentioned in the traditional sources, **- association in opposition with traditional sources. There was no association abiding by the Bonferroni criterion for statistical significance. None of the associations is statistically significant if Bonferroni criterion is taken into account. The associations significant in accordance with the recent recommendations (Benjamin et al., 2018) are italicized. See also Appendix B, Table B.3 for the English translation of the sanskrit words denoting the positively associated EPAs, followed by a list of the herbs containing astringent tasting compounds. Table 14. Associations between sweet (madhura) phytochemicals and the ethnopharmacological activities. EPA a/b /c/d p OR ln OR 95% CL POSITIVE ASSOCIATIONS dahaprasamana 56/103/43/192 0,0002 2,43 0,89 1,53-3,86 anulomana 74/85/75/160 0,004 1,86 0,62 1,23-2,81 hrdya* 55/104/51/184 0,005 1,91 0,65 1,22-2,99 snehana 22/137/14/221 0,01 2,53 0,93 1,25-5,12 vajikarana 31/128/25/210 0,02 2,03 0,71 1,15-3,6 sukrastambhana 10/149/4/231 0,02 3,88 1,35 1,19-12,58 soumanasyajanana 12/147/6/229 0,03 3,12 1,14 1,14-8,48 mutrala 86/73/101/134 0,03 1,56 0,45 1,04-2,34 brmhana 26/133/22/213 0,04 1,89 0,64 1,03-3,48 raktapittasamaka 26/133/22/213 0,04 1,89 0,64 1,03-3,48 vrsya 43/116/43/192 0,05 1,66 0,5 1,02-2,68 NEGATIVE ASSOCIATIONS kusthaghna 43/116/103/132 0,0009 0,48 − 0,74 0,31-0,73 krmighna 50/109/113/122 0,001 0,5 − 0,7 0,32-0,75 svedajanana* 20/139/52/183 0,02 0,51 − 0,68 0,29-0,89 pacana 37/122/81/154 0,02 0,58 − 0,55 0,37-0,91 sandhaniya* * 4/155/19/216 0,03 0,29 − 1,23 0,1-0,88 sothahara 66/93/124/111 0,03 0,64 − 0,45 0,42-0,95 rasayana* * 16/143/41/194 0,04 0,53 − 0,64 0,29-0,98 pramehaghna 8/151/26/209 0,04 0,43 − 0,85 0,19-0,97 recana 12/147/33/202 0,05 0,5 − 0,69 0,25-1 Note. Ayur-sweet tasting phytochemicals include sweet tasting, fatty tasting, umami tasting, and tasteless phytochemicals. Legend. a = the number of plants having ayur-sweet tasting phytochemicals and exerting the corresponding ethnopharmacological activity (EPA); b = the number of plants having ayur-sweet tasting phytochemicals but NOT exerting the corresponding EPA; c = the number of plants NOT having ayur-sweet tasting phytochemicals but exerting the corresponding EPA; d = the number of plants NOT having ayur-sweet tasting phytochemicals and NOT exerting the corresponding EPA; p = probability of error; OR = odds ratio; ln OR = natural logarithm of OR; 95% CL = 95% confidence limits for OR; *- association not mentioned in the traditional sources, **- association in opposition with traditional sources. None of the associations is statistically significant if Bonferroni criterion is taken into account. The associations significant in accordance with the recent recommendations (Benjamin et al., 2018) are italicized.. See also Appendix B, Table B.4 for the English translation of the sanskrit words denoting the positively associated EPAs, followed by a list of the herbs containing ayur-sweet tasting compounds. Table 15. Associations between the ayur-pungent (katu) phytochemicals and the ethnopharmacological activities. EPA a/ b/ c/ d p OR ln OR 95% CL POSITIVE ASSOCIATIONS pacana 67/63/55/234 3E- 11 4,52 1,51 2,88-7,11 dourgandhyahara 24/106/3/286 2E- 10 21,58 3,07 6,37-73,17 dipana 94/36/118/171 2E- 09 3,78 1,33 2,41-5,93 uttejaka 29/101/12/277 4E- 08 6,63 1,89 3,26-13,48 kaphaghna 56/74/57/232 1E- 06 3,08 1,12 1,96-4,84 anulomana 70/60/85/204 2E- 06 2,80 1,03 1,83-4,29 vedanasthapana 69/61/88/201 1E- 05 2,58 0,95 1,69-3,96 jantughna 29/101/20/269 2E- 05 3,86 1,35 2,09-7,14 artavajanana 31/99/23/266 3E- 05 3,62 1,29 2,01-6,51 hrdayottejaka 26/104/17/272 3E- 05 4 1,39 2,08-7,68 chedana 13/117/4/285 0,0001 7,92 2,07 2,53-24,78 mutrala* 79/51/117/172 0,0001 2,28 0,82 1,49-3,48 hikkanigrahana 13/117/5/284 0,0003 6,31 1,84 2,2–18,1 svedajanana 36/94/39/250 0,0008 2,45 0,9 1,47-4,09 krmighna 69/61/103/186 0,0009 2,04 0,71 1,34-3,11 vajikarana 29/101/29/260 0,001 2,57 0,95 1,47-4,52 garbhasayasankocaka 15/115/10/279 0,003 3,64 1,29 1,59-8,34 sulaprasamana 30/100/33/256 0,003 2,33 0,84 1,35-4,02 sothahara 75/55/124/165 0,006 1,81 0,6 1,19-2,76 aksepahara 13/117/10/279 0,01 3,10 1,13 1,32-7,27 sitaprasamana 10/120/6/283 0,01 3,93 1,37 1,4-11,06 vidahi 7/123/3/286 0,01 5,43 1,69 1,38-21,33 svasahara 26/104/30/259 0,01 2,16 0,77 1,22-3,83 dantya 9/121/5/284 0,01 4,22 1,44 1,39-12,87 yakrduttejaka 30/100/38/251 0,01 1,98 0,68 1,16-3,37 lekhana 24/106/27/262 0,01 2,2 0,79 1,21-3,98 katupaustika 28/102/36/253 0,02 1,93 0,66 1,12-3,33 medohara 15/115/15/274 0,02 2,38 0,87 1,13-5,03 nadyuttejaka 6/124/3/286 0,03 4,61 1,53 1,14-18,74 vataghna 19/111/22/267 0,03 2,08 0,73 1,08-3,99 rocana 39/91/58/231 0,03 1,71 0,53 1,06-2,74 kaphanihsaraka 35/95/51/238 0,04 1,72 0,54 1,05-2,81 kothaprasamana 7/123/4/285 0,04 4,05 1,4 1,17-14,1 vatasamaka 9/121/7/282 0,05 3 1,1 1,09-8,23 NEGATIVE ASSOCIATIONS dahaprasamana 20/110/85/204 0,002 0,44 − 0,83 0,25-0,75 brmhana 7/123/42/247 0,008 0,33 − 1,09 0,15-0,77 bhedana* * 2/128/22/267 0,01 0,19 − 1,66 0,04-0,82 raktapittasamaka 9/121/46/243 0,01 0,39 − 0,93 0,19-0,83 Note. Katu (ayur-pungent) phytochemicals include pungent-, burning- and aromatic- orosensation inducing phytochemicals. Legend. a = the number of plants containing ayur-pungent phytochemicals and exerting the corresponding ethnopharmacological activity (EPA); b = the number of plants containing ayur-pungent phytochemicals but NOT exerting the corresponding EPA; c = the number of plants NOT containing ayur-pungent phytochemicals but exerting the corresponding EPA; d = the number of plants NOT containing ayur-pungent phytochemicals and NOT exerting the corresponding EPA; p = probability of error; ln OR = natural logarithm of OR; 95% CL = 95% confidence limits for OR; *- association not mentioned in the traditional sources, **- association in opposition with traditional sources. The associations abiding by the Bonferroni criterion for statistical significance are highlighted by bold typing. Those significant in accordance with the recent recommendations (Benjamin et al., 2018) are italicized. See also Appendix B, Table B.6 for the English translation of the sanskrit words denoting the positively associated EPAs, followed by a list of the herbs containing ayur-pungent tasting compounds. 3.5. Comparisons between pairs of chemical classes This hypothesis was tested on several pairs of chemical categories with either different or similar tastes (see Table 16). Table 16. Comparisons between the EPA spectra of pairs of chemical categories with either different or similar tastes. Chemical category 1 Taste of chemical category 1 Chemical category 2 Taste of chemical category 2 p SIMILAR TASTES flavonoids bitter, astringent triterpenes bitter, astringent NS (0.98) tannins astringent triterpenes bitter, astringent NS (0.98) flavonoids bitter, astringent tannins astringent NS (0.84) tannins astringent anthocyans astringent NS (0.98) lactones bitter alkaloids bitter NS (0.99) DIFFERENT TASTES tannins astringent alkaloids bitter HS (4E-08) essential oils ayur-pungent alkaloids bitter HS (4E-09) Legend. p = probability of error; NS = not statistically significant (there is no statistically significant difference between the EPA spectra of chemical categories 1 and 2); HS = highly statistically significant (there is a statistically significant difference between the EPA spectra of chemical categories 1 and 2). Unfortunately almost none of these chemical categories (maybe with the exception of the tannins) satisfies the very criteria stated in the Material and methods section. 4. Discussions Initial investigation of the database provided an unexpected result: there was no global statistical correlation between the various chemical classes of phytocompounds and EPAs, although there were several individual correlations. The second investigation suggested a statistically significant global correlation and also several individual correlations among the various taste-based classes of phytocompounds and EPAs. These results suggest that, among these two attributes of medicinal substances, the phytochemical taste may be more relevant than the chemical class in predicting the EPAs of that medicinal substance. 4.1. Are the results compatible with the bioscientific paradigm? The correlations between organoleptic properties and therapeutic indications of medicinal plants are usually accepted as important mnemonic aids. However more and more scientists consider that these “may not be as fully incompatible with bioscientific paradigms as previously criticized” (Geck et al., 2017). Taste and taste receptors may play a “dominant role” not only in the evaluation, but also in the biological activities of medicinal plants (Behrens et al., 2018). Recent studies confirm also our ethnopharmacology- based and statistically- confirmed hypothesis that there is an association between the taste of phytochemicals and their biological activities. For example, a positive association was found between bitter molecular taste (sanskr. tikta rasa) and anti-inflammatory activity (sanskr. sothahara) (Table 12). Bitter food has been shown to have beneficial effects in inflammatory diseases not only in Ayurveda, but also in Traditional Chinese Medicine (Lin and Lin, 2011). Bitter compounds belonging to different chemical classes (e.g. alkaloids- aloperine, berberine, crotaline, chloroquine, quinine; cyanogenic glucosides- amygdalin; flavanones- naringenin, nitro compounds- aristolochic acid, terpenes- humulone, etc.) have anti-inflammatory effects (e.g. modulatory effects on cytokine secretion) in various in vitro and in vivo models. These effects may be in part mediated by the bitter taste receptors TAS2R (Ano et al., 2017, Camoretti-Mercado et al., 2015, Lin and Lin, 2011, Sharma et al., 2017, Yang et al., 2007). Another example is the positive association between ayur-pungency (sanskr. katu rasa) and antiinfectious activity (Sanskr. krimighna) (Table 15). Extensive evidence has demonstrated antimicrobial properties for essential oils and their constituents that belong to the katu rasa category (Chouhan et al., 2017, Dhifi et al., 2016, Orchard and van Vuuren, 2017, Pandey and Singh, 2017, Swamy et al., 2016, Uma et al., 2017). The positive association between astringency (Sanskr. kashaya rasa) and antihemorrhagic activity (Sanskr. raktastambhana) is supported by several clinical studies showing that various tannins accelerate blood clotting and hemostasis (Chung et al., 1998, Kim et al., 2016, Lim, 2012). There are several new and interesting associations, not traditionally acknowledged, that were revealed by statistical analysis: e.g. PA astringent- mastikabalya (brain tonic) (Table 13) and PA sour- kesya (hair tonic) (Table 5). Interestingly, many members of tannin class, which induce astringent orosensation, elicited neuroprotective effects in several in vitro and in vivo studies (Braidy et al., 2017, Chandrasekhar et al., 2017, Mendonca et al., 2017, Tejada et al., 2017, Tong et al., 2016, Wu et al., 2015). 4.2. Apparent contradictions in the results yielded by the statistical calculations There are two apparent contradictions: 1. There is a global association among tastes and EPAs but not among chemical categories and EPAs. 2. There is no global association between chemical categories and EPAs, but there are some individual associations. The first (apparent) contradiction. The existence of a correlation between taste and chemical class seems self-evident. It was the purpose of this investigation to demonstrate that chemical categories underlie the correlation between rasas/tastes and EPAs (M. Gilca and Barbulescu, 2015a, Gilca and Barbulescu, 2015b), but this was not the case (see Section 3.1). However, it is worth noting that: – A given taste is not restricted to a certain chemical category [astringent are tannins, but so are many flavonoids and triterpenes; bitter are not only lactones and iridoids, but also many flavonoids and triterpenes, as well as most of the alkaloids (a vast category including many chemical classes); pungent/aromatic is the taste of most of the volatile compounds [a class which is defined by a physical characteristic (volatility), which includes many different chemical categories] – see Annex A in Data in Brief (Dragos and Gilca, n.d.)]; – Reciprocally, a given chemical category may include substances with several different tastes (triterpenes may be bitter or astringent, flavonoids may be bitter or astringent or even sweet, etc. – see Annex A in Data in Brief (Dragos and Gilca, n.d.); nonetheless, there are many chemical categories with one definite taste, e.g. tannins). Hence, there may exist a global correlation between EPAs and tastes (M. Gilca and Barbulescu, 2015a, Gilca and Barbulescu, 2015b) in the absence of a global correlation between EPAs and chemical classes (see Section 3.1). The second (apparent) contradiction. There is no global association between chemical categories and EPAs, but there are some individual associations and those associations may be due to statistical coincidence. Howevere, statistical rigor was maintained by inclusion of the Bonferroni correction (see Section 3.2, Table 10). The fact that several subcategories of a larger category have a certain characteristic does not mean that the whole category should necessarily have those characteristics. Reciprocally, the fact that a large category does not have a certain feature does not necessarily mean that all its subcategories should not have that feature. Furthermore, few individual correlations were found with phytocompounds that have a very definite taste (astringent for tannins, pungent/aromatic for monoterpene hydrocarbons, monoterpene alcohols, and monoterpene phenols, sour for short aliphatic acids, pungent for glucosinolates; not even the gums are a true exception, as they are also definitely tasteless). It is true that these phytocompounds also pertain to a well defined chemical category. Therefore these few individual associations might be the consequence of the correlation between tastes and EPAs, as well as between chemical categories and EPAs. But for the former there are many others arguments (see Section 3.4), while for the latter there are none. 4.3. Implications of the results of the present study and applicability in taste science This study suggests that taste is potentially more indicative of pharmacological activity than chemical class. A possible explanation may be that the geometrical shape may be more relevant for taste determinism than the chemical category (Kortagere et al., 2009). Molecular shape is a critical determinant for ligand-receptor interaction which constitutes the substrate for both the biological activity and the taste. Molecules of the same chemical class may have significantly (or even completely) different shapes, while molecules with different chemical classes may have very similar shapes. Consequently, taste (read “molecular shape”, therefore “molecular structure”, but not necessarily “chemical category/class”) may be a more relevant determinant of the pharmacodynamic effect of a natural compound than its chemical affiliation. Another interesting scientific finding that supports association between tastes and EPAs is the recently discovered diffuse extraoral pan-organ/ pan-tissue distribution of gustative receptors (Behrens and Meyerhof, 2011, Gilca and Dragos, 2017, Laffitte et al., 2014, Manson et al., 2014). Surprisingly, during the last two decades, taste receptors (TASR) have been discovered not only on the tongue, but also in many extragustative organs and tissues with apparent ubiquitous distribution (e.g. stomach, intestines, liver, pancreas, heart, respiratory system, brain, kidney, urinary bladder, testes, ovaries, bones, and joints) (Dehkordi et al., 2012, Foster et al., 2014, Henquin, 2012, Li, 2013, Lund et al., 2013, Raybould, 1998, Ren et al., 2009, Trujillo et al., 1999, Voigt et al., 2012, Yamamoto and Ishimaru, 2013). Taste receptors are now considered emerging targets for new pharmacological agents (e.g. bronchodilation agents) (Behrens and Meyerhof, 2015, Qadri et al., 2012). Tastants may have broad applications in the food and therapeutic industries. As with Ayurveda theory, each taste (sanskrit rasa) has specific systemic pharmacological effects (Sharma and Dash, 2006). Taking into account the recent discovery of extragustative locations and physiological roles for taste receptors, the results herein suggest a potential mode of action for traditional medicines, which requires more experimental and clinical investigation. 4.4. Potential application of the results of the present study in finding a substitute remedy A millennial old procedure for drug substitution based on the similarities of ethnopharmacological descriptors and other criteria (e.g. similar morphological characteristics) dates back to Caraka, one of the most prominent ancient Ayurveda physicians (400–200 B.C.) (Nagarajan et al., 2015a, Sharma and Dash, 2006). The method was for the first time systematically documented in Bhavaprakasha, a traditional text dating from the 16th century AD (Joshi et al., 2012, Nagarajan et al., 2015b). According to Bhavaprakasha, a substitute drug (sanskr. pratinidhi dravya) should be used if the most appropriate drug is unavailable (Sanskr. abhava-pratinidhi dravya) or the patient is intolerant (Purva Khanda VI.I.159–168) (Murthy, 2004). Since many Indian medicinal plants are not available to Western people, the use of non-Indian substitute plants (pratinidhi dravya) represents a solution recommended by Ayurveda. This substitution method was suggested for rare endangered species, in order to avoid their extinction (Nagarajan et al., 2015b). The pratinidhi dravya methodical approach is clearly described in Ayurveda: “(…) if a remedy is not available, any other remedy which is similar in terms of rasa (taste), virya (energetic nature) and vipaka (postdigestive effect) should be selected by the physician and used.” (Bhavaprakasha, Purva Khanda VI.I.159–168) (Murthy, 2004). Nonetheless, a substitute drug in a polyherbal formula should be limited to the role of auxiliary or accessory constituent, meaning that a substitute medicinal plant should not be used as the chief ingredient (sanskr. pradhana dravya) of the polyherbal formula (Murthy, 2004). The validity of the pratinidhi dravya method was already assessed in a few phytochemical (Nagarajan et al., 2015b), pharmacological (Nagarajan et al., 2015a) or animal (Venkatasubramanian et al., 2010) studies. Scientists concluded that unrelated species could share similar Ayurvedic profile and biological activity, despite phytochemical or morphological dissimilarities. This is the basis for a pharmaco-taxonomic or functional taxonomic classification of medicinal plants (Nagarajan et al., 2015a, Venkatasubramanian et al., 2010). 4.5. Factors susceptible to bias the statistical calculations 1. The most important is the process of eliminating rows and columns while performing the Chi square test in order to decrease the proportion of less-than-5-expected-values (LT5EVs) to below the threshold of 20%. The rationale for doing this is stated in the Material and methods section. The only other strategy to achieve the same objective would be to pool rows and columns. Pooling together EPAs is quite arbitrary and medically nonsensical. Pooling together chemical categories can be done (and it was done) to a point, but the main limitation of this process is that it inevitably leads to intersecting (non-disjunct) categories (which are not admissible while performing Chi-square test) and/or chemically heteroclite categories which often are more similar by their taste than by their chemical structure (alkaloids and volatile compounds are good examples, almost all being bitter and ayur-pungent, respectively, but including many substance classes with vastly diverse chemical structures. Hence, the statistical association of such classes would be an argument for the taste-EPAs association, not for the chemical class-EPAs association. On the other hand, the elimination of rows and columns was justified because this approach allowed the null hypothesis to be tested on samples of chemical categories and samples of EPAs established by means of an objective process. Indeed, the process of elimination was blinded with respect to the actual EPAs and chemical categories – the only criterion was prevalence, reflected in the magnitude of the marginal sums, and therefore the magnitude of the expected values and the proportion of LT5EVs. This proportion was the obvious elimination criterion as the magnitude of this proportion was the one precluding the application of Chi-square test. 2. The chemical categories may not be completely disjunctive as required by the Chi-square test. However, only one of these chemical categories was applied to each phytocompound. So, for example none of the phytocompounds was simultaneously a monoterpene alcohol and an aliphatic amine. Of course all monoterpene alcohols are also monoterpenes, and assuredly terpenes. But broad categories such as monoterpenes and terpenes were not considered in the analysis. 3. The EPAs may not be completely disjunctive as the Chi-square test requires. In other words some of the EPAs may be synonymous enough to be considered one and the same thing. To avoid this confound, all of the synonymous Sanskrit terms employed for designating each EPA were identified (see Section 2.1). However, there were still several EPAs left with somewhat similar meanings. Hence, an analysis employing enlarged EPAs [such as abortive (garbhasayasankocaka + garbhasravakara) or analgesic (sulaprasamana + vedanasthapana)] was performed – the results were still statistically suggestive, but less so than the results obtained by employing the narrow EPAs (actually the traditional EPAs). The reduction in statistical power after such a merging operation generally indicates that entities with significantly different association patterns (and hence significantly different meanings) have been merged, resulting in a blunted statistical significance. 4.6. Limitations of the present study Circular reasoning is a serious issue regarding the relation between traditional rasa-s and traditional EPAs. Centuries or even millennia ago healers may have noticed that certain tastes (rasa-s) were associated with certain effects. These associations entered into the collective consciousness and consequently specific tastes/rasa-s may have become memes for specific pharmacological actions, subsequently guiding the selection of new plants and being integrated into Ayurvedic theory. However, this study is a step out of this self-affirming circle as one of the two terms of our correlation is not a tradition-based one – on the contrary, it is based on modern scientifically recognized data. The plants’ tastes considered in this study were not the (more or less subjectively established) traditional rasa-s, but the tastes imparted by the main phytocompounds. For each given plant, this tastes were derived by means of an objective, modern references-based, three-stage process: 1. Establish, for each given plant, the main/major phytocompounds (defined in the Methods Section); 2. Establish the tastes of these major phytocompounds; 3. Attribute these tastes to the corresponding plant. The relevant references are mentioned in Data in Brief (Dragos and Gilca, n.d.). There are other several limitations, mostly due to data availability and to cultural differences: □ Not all the phytochemicals had a corresponding taste in the database, and only 394 medicinal plants from the initial 431 were defined in terms of their “molecular taste” (phytochemical tastes). There were objective difficulties in collecting data regarding the taste of various phytochemicals (see Section 2. Materials and methods). Further pharmacological receptor profile were needed to identify the exact “molecular taste” of phytochemicals. □ Only the major phytochemicals isolated from each plant were included in the study. The rationale was that, despite the huge variety of secondary metabolites with differing tastes, only the predominant tastes of each plant (correlated with the predominant constituent phytochemicals) are relevant from an ethnopharmacological point of view. □ The phytochemical taste was identified using heterogenous types of studies: ○ Direct evidence- sensory analytical methods, receptor assay studies ○ Indirect evidence- studies on the compounds contributing to the taste of various food items identified by the combined application of chemical and sensory analytical methods. □ This was an objective limitation, due to the scarcity of studies that evaluated organoleptic phytochemicals. In the future, advancement in taste science will allow for a database with herbal “molecular taste” exclusively on receptor assay studies, which are not currently available. With that availability, the influence of specific cultural factors on the “molecular taste” will be eliminated. □ Certain phytochemicals may induce different oral sensations depending on their concentration. For instance, N-Feruloyl-3-methoxytyramine, a natural compound found in Beta vulgaris, was described as weakly pungent at 10 ppm and bitter at 50 ppm and 100 ppm by trained consensus panel (Backes et al., 2015). Concentration was not considered in this study. □ There is no perfect overlap between the traditional rasas or ayur-tastes (sweet, bitter, salty, sour, pungent, astringent) and the modern tastes (sweet, bitter, salty, sour, umami). It was problematic to classify certain phytochemicals as pungent or astringent, since these orosensations are still not well defined by modern science (des Gachons et al., 2012, Jiang et al., 2014, Roper, 2014). Based on Ayurveda theory only, aromatic and burning sensations were classified as belonging to pungency (katu rasa) (Gilca and Dragos, 2017). Western culture has several confusing words that describe taste, like tart or acrid, which may signify bitter, sour or even pungent orosensations, each corresponding to a distinct specific rasa in the Indian culture. In the Western culture, there is a tendency to qualify as “bitter” any unpleasant intense taste – by contrast, in Ayurveda a strong disagreeable taste is more susceptible to be considered katu (ayur-pungent) in Ayurveda (as katu is also translated as displeasing) (Monier-Williams, 2002). The reliability of sensory examinations is sometimes difficult to assess, especially when the words used may not be the most appropriate description of the stimulus (e.g. bitter for horseradish) (Fenwick et al., 1990). □ Chemical data are unavailable for the majority of plant species (Saslis-Lagoudakis et al., 2011), therefore the “molecular taste” can be expensive to identify (as it relies on the chemical analysis of the plants). Nevertheless, the perceptible taste identified by a panel of tasters may be an alternative and inexpensive tool used for identification of medicinal plants (or even phytochemicals) with a specific therapeutic potential. Taste assessment by electronic-tongue, which mimics taste perception in humans, was recently proposed to evaluate rasa in Ayurvedic pharmacology (Jayasundar and Ghatak, 2016). Moreover, electronic taste sensors have the advantage of non-specificity, and are able to detect the overall gustatory impression of a multicomponent mixture (e.g. herbal extract) instead of analysing the isolated individual phytochemicals (Eckert et al., 2013). □ Another limitation of the chemosensorial approach is the extreme inter- and intraindividual variability of taste perception. Genetic, physiological, hormonal, sexual, dietary and cultural factors may result in differences in taste perception (Bartoshuk, 2000, Behrens and Meyerhof, 2006, Bufe et al., 2005, Duffy et al., 2004, Leonti, 2011, Martin and Sollars, 2017, Prutkin et al., 2000, Reinberger, 2006, World Health Organisation, 2011). Particular sensations may have different symbolic values in various cultural contexts, and hence may be associated with different ethnomedical practices/uses (Classen, 1997). For instance, bitter is cold in Ayurveda, therefore used for the treatment of “hot” diseases (e.g. fever, biliousness) (Sharma and Dash, 2006, Stefan, 2005). On the contrary, in Mesoamerican ethnomedicine, there is a considerable cross-cultural consensus that bitter is hot, therefore it is used for symptoms of excessive coldness (e.g. post-partum ailments and musculoskeletal pain) (Geck et al., 2017). □ Taste is not the single ethnopharmacological descriptor that is used for selection of medicinal plants in ethnomedicine. Other organoleptic properties (olfactive, visual and tactile) should be evaluated when analysing the traditional medicinal potential of plants (Geck et al., 2017, Leonti et al., 2002). □ Perception and organoleptic properties can change over time (e.g. modern people versus their predecessors), being shaped by the cultural conditioning specific to a certain period (Classen, 1997, Shepard JR, 2004, Sorokowska et al., 2014). In this database, the phytochemical tastes were based on sensory analytical methods collected from different sources over a long period of time, starting with 1894 (Sohn, 1894) and ending in the present. Organoleptic re-evaluation of these phytochemicals by a panel of tasters represents a mandatory task for future research programs. □ In its present state the database is insufficient to allow statistical analysis of the third working hypothesis. 5. Conclusions and future directions For a given medicinal plant, the tastes of the constituent phytocompounds seems to be more closely linked to its EPAs than are the chemical classes of the phytocompounds (but not their physicochemical structure). This conclusion is not surprising if the connection between taste and molecular shape, hence structure, is considered/ taken into consideration. There are a number of possible considerations. For example, a potential challenge to our hypothesis (beside that already described in Section 3.5. Comparisons between pairs of chemical classes) would be to develop two predictive tools for the EPAs (one based on the traditionally ascribed rasas and another based on “molecular taste”) and to test the efficiency of these predictive tools on the plants not considered herein. Fundamental ethnopharmacological questions, such as the rationale behind the use of medicinal plants for specific therapeutic indications in traditional medicine, continue to motivate further development of the PhytoMolecularTasteDB in connection with the chemosensorial approach. Future extensions may include additional information about the intensity of taste (e.g. slightly bitter, intensely bitter), assigning the plant-derived tastants to their cognate taste receptors or orosensation transducers, as well as reference to the tastant biological activities, and the implementation of visualization tools that may help in the exploration of 2D or 3D structure similarities. Acknowledgments This research received no specific grant from any funding agency in the public, commercial or not-for-profit sectors. References Altman, 1991 D.G. Altman Practical Statistics for Medical Research Chapman and Hall (1991) Ankli et al., 1999 A. Ankli, O. Sticher, M. Heinrich Yucatec Maya Medicinal plants versus nonmedicinal plants: indigenous characterization and selection Hum. Ecol., 27 (1999), pp. 557-580, 10.1023/A:1018791927215 CrossRefView Record in Scopus Ano et al., 2017 Y. Ano, A. Dohata, Y. Taniguchi, A. Hoshi, K. Uchida, A. Takashima, H. Nakayama Iso-alpha-acids, bitter components of beer, prevent inflammation and cognitive decline induced in a mouse model of Alzheimer's disease J. Biol. Chem., 292 (2017), pp. 3720-3728, 10.1074/jbc.M116.763813 CrossRefView Record in Scopus Higgins and Green, 2011 AnonHiggins, J.P.T., Green, S., (editors)., 2011. Section 16.9.2. Studies with zero-cell counts in Cochrane Handbook for Systematic Reviews of Interventions Version 5.1.0 [updated March 2011]. The Cochrane Collaboration, 2011. Available from 〈www.handbook.cochrane.org〉. [WWW Document]. Auvray and Spence, 2008 M. Auvray, C. Spence The multisensory perception of flavor Conscious. Cogn., 17 (2008), pp. 1016-1031, 10.1016/j.concog.2007.06.005 ArticleDownload PDFView Record in Scopus Backes et al., 2015 M. Backes, K. Obst, J. Bojahr, A. Thorhauer, N. Roudnitzky, S. Paetz, K.V. Reichelt, G.E. Krammer, W. Meyerhof, J.P. Ley Rubemamine and Rubescenamine, two naturally occurring N-Cinnamoyl Phenethylamines with umami-taste-modulating properties J. Agric. Food Chem., 63 (2015), pp. 8694-8704, 10.1021/acs.jafc.5b04402 CrossRefView Record in Scopus Bartoshuk, 2000 L.M. Bartoshuk Comparing sensory experiences across individuals: recent psychophysical advances illuminate genetic variation in taste perception Chem. Senses, 25 (2000), pp. 447-460 CrossRefView Record in Scopus Behrens and Meyerhof, 2006 M. Behrens, W. Meyerhof Bitter taste receptors and human bitter taste perception (in:) Cell. Mol. Life Sci. (2006), pp. 1501-1509, 10.1007/s00018-006-6113-8 CrossRefView Record in Scopus Behrens and Meyerhof, 2011 M. Behrens, W. Meyerhof Gustatory and extragustatory functions of mammalian taste receptors Physiol. Behav., 105 (2011), pp. 4-13, 10.1016/j.physbeh.2011.02.010 ArticleDownload PDFView Record in Scopus Behrens and Meyerhof, 2015 M. Behrens, W. Meyerhof Flavour Development, Analysis and Perception in Food and Beverages, Flavour Development, Analysis and Perception in Food and Beverages Elsevier (2015), 10.1016/B978-1-78242-103-0.00014-X Behrens et al., 2018 M. Behrens, M. Gu, S. Fan, C. Huang, W. Meyerhof Bitter substances from plants used in traditional Chinese medicine exert biased activation of human bitter taste receptors Chem. Biol. Drug Des., 91 (2018), pp. 422-433, 10.1111/cbdd.13089 View Record in Scopus Benjamin et al., 2018 D.J. Benjamin, J.O. Berger, M. Johannesson, B.A. Nosek, E.-J. Wagenmakers, R. Berk, K.A. Bollen, B. Brembs, L. Brown, C. Camerer, D. Cesarini, C.D. Chambers, M. Clyde, T.D. Cook, P. De Boeck, Z. Dienes, A. Dreber, K. Easwaran, C. Efferson, E. Fehr, F. Fidler, A.P. Field, M. Forster, E.I. George, R. Gonzalez, S. Goodman, E. Green, D.P. Green, A.G. Greenwald, J.D. Hadfield, L.V. Hedges, L. Held, T. Hua Ho, H. Hoijtink, D.J. Hruschka, K. Imai, G. Imbens, J.P.A. Ioannidis, M. Jeon, J.H. Jones, M. Kirchler, D. Laibson, J. List, R. Little, A. Lupia, E. Machery, S.E. Maxwell, M. McCarthy, D.A. Moore, S.L. Morgan, M. Munafó, S. Nakagawa, B. Nyhan, T.H. Parker, L. Pericchi, M. Perugini, J. Rouder, J. Rousseau, V. Savalei, F.D. Schönbrodt, T. Sellke, B. Sinclair, D. Tingley, T. Van Zandt, S. Vazire, D.J. Watts, C. Winship, R.L. Wolpert, Y. Xie, C. Young, J. Zinman, V.E. Johnson Redefine statistical significance Nat. Hum. Behav, 2 (2018), pp. 6-10, 10.1038/s41562-017-0189-z CrossRefView Record in Scopus Bensky et al., 2004 D. Bensky, S. Clavey, E. Stoger Chinese Herbal Medicine Materia Medica (3rd ed.), Eastland Press, Seattle, USA (2004) Bhishagratna, 1998 K.K.L. Bhishagratna Susruta Samhita (First. ed.), Chowkhamba Sanskrit Series Office, Varanasi (1998) Braidy et al., 2017 N. Braidy, B.-E. Jugder, A. Poljak, T. Jayasena, S.M. Nabavi, P. Sachdev, R. Grant Molecular targets of tannic acid in Alzheimer's disease Curr. Alzheimer Res., 14 (2017), pp. 861-869, 10.2174/1567205014666170206163158 View Record in Scopus Brett, 1994 J.A. Brett Medical Plant Selection Criteria Among the Tzeltal Maya of Highland Chiapas, Mexico (Ph.D. Dissertation) Univ. California, San Francisco (1994) Brett and Heinrich, 1998 J.A. Brett, M. Heinrich Culture, perception and the environment the role of chemosensory perception Angew. Bot., 72 (1998), pp. 67-69 View Record in Scopus Bufe et al., 2005 B. Bufe, P.A.S. Breslin, C. Kuhn, D.R. Reed, C.D. Tharp, J.P. Slack, U.K. Kim, D. Drayna, W. Meyerhof The molecular basis of individual differences in phenylthiocarbamide and propylthiouracil bitterness perception Curr. Biol., 15 (2005), pp. 322-327, 10.1016/j.cub.2005.01.047 ArticleDownload PDFView Record in Scopus Camoretti-Mercado et al., 2015 B. Camoretti-Mercado, S.H. Pauer, H.M. Yong, D.C. Smith, D.A. Deshpande, S.S. An, S.B. Liggett Pleiotropic effects of bitter taste receptors on [Ca2+]i mobilization, hyperpolarization, and relaxation of human airway smooth muscle cells PLoS One, 10 (2015), p. e0131582, 10.1371/journal.pone.0131582 CrossRef Chandrasekhar et al., 2017 Y. Chandrasekhar, E.M. Ramya, K. Navya, G. Phani Kumar, K.R. Anilakumar Antidepressant like effects of hydrolysable tannins of Terminalia catappa leaf extract via modulation of hippocampal plasticity and regulation of monoamine neurotransmitters subjected to chronic mild stress (CMS) Biomed. Pharmacother., 86 (2017), pp. 414-425, 10.1016/j.biopha.2016.12.031 ArticleDownload PDFView Record in Scopus Chouhan et al., 2017 S. Chouhan, K. Sharma, S. Guleria Antimicrobial activity of Some essential oils-present status and future perspectives Medicines (Basel, Switz.), 4 (2017), 10.3390/medicines4030058 Chung et al., 1998 K.-T. Chung, T.Y. Wong, C.-I. Wei, Y.-W. Huang, Y. Lin Tannins and human health: a review Crit. Rev. Food Sci. Nutr. (1998), 10.1080/10408699891274273 Classen, 1997 C. Classen Foundations for an anthropology of the senses Int. Soc. Sci. J., 49 (1997), pp. 401-412, 10.1111/j.1468-2451.1997.tb00032.x View Record in Scopus Dehkordi et al., 2012 O. Dehkordi, J.E. Rose, M. Fatemi, J.S. Allard, K.V. Balan, J.K. Young, S. Fatima, R.M. Millis, A. Jayam-Trouth Neuronal expression of bitter taste receptors and downstream signaling molecules in the rat brainstem Brain Res., 1475 (2012), pp. 1-10, 10.1016/j.brainres.2012.07.038 ArticleDownload PDFView Record in Scopus des Gachons et al., 2012 C.P. des Gachons, E. Mura, C. Speziale, C.J. Favreau, G.F. Dubreuil, P.A.S. Breslin Opponency of astringent and fat sensations Curr. Biol. (2012), 10.1016/j.cub.2012.08.017 Dhifi et al., 2016 W. Dhifi, S. Bellili, S. Jazi, N. Bahloul, W. Mnif Essential oils' chemical characterization and investigation of some biological activities: a critical review Med. (Basel, Switz.), 3 (2016), 10.3390/medicines3040025 Dragos and Gilca, n.d. Dragos, D., Gilca, M., n.d.. PhytoMolTasteDB: an integrative database with tastants derived from Indian medicinal plants (Submitted for publication). Data Br. Duffy et al., 2004 V.B. Duffy, J.M. Peterson, L.M. Bartoshuk Associations between taste genetics, oral sensation and alcohol intake Physiol. Behav., 82 (2004), pp. 435-445, 10.1016/j.physbeh.2004.04.060 ArticleDownload PDFView Record in Scopus Eckert et al., 2013 C. Eckert, M. Pein, J. Reimann, J. Breitkreutz Taste evaluation of multicomponent mixtures using a human taste panel, electronic taste sensing systems and HPLC Sens. Actuators B Chem., 182 (2013), pp. 294-299, 10.1016/j.snb.2013.03.012 ArticleDownload PDFView Record in Scopus Etkin and Elisabetsky, 2005 N.L. Etkin, E. Elisabetsky Seeking a transdisciplinary and culturally germane science: the future of ethnopharmacology J. Ethnopharmacol., 100 (2005), pp. 23-26, 10.1016/j.jep.2005.05.025 ArticleDownload PDFView Record in Scopus Fenwick et al., 1990 G.R. Fenwick, C.L. Curl, N.M. Griffiths, R.K. Heaney, K.R. Price Bitter principles in food plants R.L. Rouseff (Ed.), Bitternes in Foods and Beverages. Developments in Food Science, 25, Elsevier Science Publishers B.V, Amsterdam (1990), pp. 205-250 Fleiss, 1986 Fleiss, J.L., 1986. The Design and Analysis of Clinical ExperimentsNo Title. Foster et al., 2014 S.R. Foster, K. Blank, L.E. See Hoe, M. Behrens, W. Meyerhof, J.N. Peart, W.G. Thomas Bitter taste receptor agonists elicit G-protein-dependent negative inotropy in the murine heart FASEB J. Off. Publ. Fed. Am. Soc. Exp. Biol., 28 (2014), pp. 4497-4508, 10.1096/fj.14-256305 CrossRefView Record in Scopus Geck et al., 2017 M.S. Geck, S. Cabras, L. Casu, A.J. Reyes Garcia, M. Leonti The taste of heat: how humoral qualities act as a cultural filter for chemosensory properties guiding herbal medicine J. Ethnopharmacol. (2017), 10.1016/j.jep.2017.01.027 Gilca and Barbulescu, 2015a M. Gilca, A. Barbulescu Taste of medicinal plants: a potential tool in predicting ethnopharmacological activities? J. Ethnopharmacol., 174 (2015), pp. 464-473, 10.1016/j.jep.2015.08.040 ArticleDownload PDFView Record in Scopus Gilca and Barbulescu, 2015b M. Gilca, A. Barbulescu Taste of medicinal plants: a potential tool in predicting ethnopharmacological activities? J. Ethnopharmacol. (2015), p. 174, 10.1016/j.jep.2015.08.040 Gilca and Dragos, 2017 M. Gilca, D. Dragos Extraoral Taste Receptor Discovery: new Light on Ayurvedic Pharmacology. Evidence-Based Complement Altern. Med., 2017 (2017), pp. 1-30, 10.1155/2017/5435831 CrossRef Gogte, 2000 V.M. Gogte Ayurvedic Pharmacology and Therapeutic Uses of Medicinal Plants Bhavan's Book University, Bharatiya Vidya Bhavan's SPARC, Mumbai, India (2000) Gollin, 2004 L.X. Gollin Subtle and profound sensory attributes of medicinal plants among the Kenyah Leppo’ Ke of East Kalimantan Borneo. J. Ethnobiol., 24 (2004), pp. 173-201 View Record in Scopus Govindarajan, 1979 V.S. Govindarajan Pungency: the stimuli and their evaluation J.C. Boudreau (Ed.), Food Taste Chemistry, ACS Symposium Series 115, American Chemical Society (1979), pp. 3-53, 10.1021/bk-1979-0115.ch003 GraphPad InStat software package, 2017 GraphPad InStat software package, 2017. (〈https://graphpad.com/quickcalcs/contingency2/〉) [WWW Document]. Heinrich, 1994 M. Heinrich Herbal and symbolic medicines of the lowland Mixe (Oaxaca, Mexico): disease concepts, healers’role and plant use Anthropos, 89 (1994), pp. 73-83 Heinrich, 2008 M. Heinrich Ethnopharmacy and natural product research-Multidisciplinary opportunities for research in the metabolomic age Phytochem. Lett., 1 (2008), pp. 1-5, 10.1016/j.phytol.2007.11.003 ArticleDownload PDFView Record in Scopus Henquin, 2012 J.-C. Henquin Do pancreatic β cells “taste” nutrients to secrete insulin? Sci. Signal., 5 (2012), p. pe36, 10.1126/scisignal.2003325 CrossRef Hirasa and Takemasa, 1998 K. Hirasa, M. Takemasa Spice Science and Technology CRC Press (1998) J. Sweeting et al., 2004 M. J. Sweeting, A. J. Sutton, P. C. Lambert What to add to nothing? Use and avoidance of continuity corrections in meta-analysis of sparse data Stat. Med., 23 (2004), pp. 1351-1375, 10.1002/sim.1761 CrossRef Jayasundar and Ghatak, 2016 R. Jayasundar, S. Ghatak Spectroscopic and E-tongue evaluation of medicinal plants: a taste of how rasa can be studied J. Ayurveda Integr. Med., 7 (2016), pp. 191-197, 10.1016/j.jaim.2016.09.003 ArticleDownload PDFView Record in Scopus Jiang et al., 2014 Y. Jiang, N.N. Gong, H. Matsunami Astringency: a more stringent definition Chem. Senses, 39 (2014), pp. 467-469, 10.1093/chemse/bju021 CrossRef Johannes, 1975 A. Johannes Medicinal plants of the nekematigi of the eastern highlands of New Guinea Econ. Bot., 29 (1975), pp. 268-277, 10.1007/BF02873177 CrossRefView Record in Scopus Johannes, 1986 A. Johannes Medicinal plants of the New Guinea highlands: an ethnopharmacologic and phytochemical update N. Etkin (Ed.), Plants in Indigenous Medicine & Diet: Biobehavioral Approaches, Redgrave Publishing Company, New York (1986), pp. 187-210 Joshi et al., 2007 K. Joshi, A. Hankey, B. Patwardhan Traditional Phytochemistry: identification of Drug by “Taste” Evid. Based Complement. Altern. Med., 4 (2007), pp. 145-148, 10.1093/ecam/nel064 CrossRefView Record in Scopus Joshi et al., 2012 P.R. Joshi, B.R. Patel, V.J. Shukla An overview of the causes of current practices in Pratinidhi Dravyas (substitution of drugs) in Ayurveda including newer techniques for their evaluation Ayu, 33 (2012), pp. 481-485, 10.4103/0974-8520.110518 CrossRefView Record in Scopus Kim et al., 2016 K. Kim, H. Lee, S. Hong TAPE: a biodegradable hemostatic glue inspired by a ubiquitous compound in plants for surgical application J. Vis. Exp. (2016), 10.3791/53930 Klein et al., 2013 A.H. Klein, M.I. Carstens, E. Carstens Eugenol and carvacrol induce temporally desensitizing patterns of oral irritation and enhance innocuous warmth and noxious heat sensation on the tongue PAIN®, 154 (2013), pp. 2078-2087, 10.1016/j.pain.2013.06.025 ArticleDownload PDFCrossRefView Record in Scopus Kortagere et al., 2009 S. Kortagere, M.D. Krasowski, S. Ekins The importance of discerning shape in molecular pharmacology Trends Pharmacol. Sci., 30 (2009), pp. 138-147, 10.1016/j.tips.2008.12.001 ArticleDownload PDFView Record in Scopus Kuroki et al., 2016 S. Kuroki, N. Hagura, S. Nishida, P. Haggard, J. Watanabe Sanshool on the fingertip interferes with vibration detection in a rapidly-adapting (RA) tactile channel PLoS One, 11 (2016), p. e0165842, 10.1371/journal.pone.0165842 CrossRef Laffitte et al., 2014 A. Laffitte, F. Neiers, L. Briand Functional roles of the sweet taste receptor in oral and extraoral tissues Curr. Opin. Clin. Nutr. Metab. Care, 17 (2014), pp. 379-385, 10.1097/MCO.0000000000000058 CrossRefView Record in Scopus Leonti, 2011 M. Leonti The future is written: impact of scripts on the cognition, selection, knowledge and transmission of medicinal plant use and its implications for ethnobotany and ethnopharmacology J. Ethnopharmacol., 134 (2011), pp. 542-555, 10.1016/j.jep.2011.01.017 ArticleDownload PDFView Record in Scopus Leonti and Verpoorte, 2017 M. Leonti, R. Verpoorte Traditional Mediterranean and European herbal medicines J. Ethnopharmacol. (2017), 10.1016/j.jep.2017.01.052 Leonti et al., 2002 M. Leonti, O. Sticher, M. Heinrich Medicinal plants of the Popoluca, Mexico: organoleptic properties as indigenous selection criteria J. Ethnopharmacol., 81 (2002), pp. 307-315, 10.1016/S0378-8741(02)00078-8 ArticleDownload PDFView Record in Scopus Li, 2013 F. Li Taste perception: from the tongue to the testis Mol. Hum. Reprod., 19 (2013), pp. 349-360, 10.1093/molehr/gat009 CrossRefView Record in Scopus Lim, 2012 S.W. Lim Aluminum potassium sulfate and tannic acid injection for hemorrhoids J. Korean Soc. Coloproctol., 28 (2012), pp. 73-77, 10.3393/jksc.2012.28.2.73 CrossRefView Record in Scopus Lin and Lin, 2011 W.-C. Lin, J.-Y. Lin Five bitter compounds display different anti-inflammatory effects through modulating cytokine secretion using mouse primary splenocytes in vitro J. Agric. Food Chem., 59 (2011), pp. 184-192, 10.1021/jf103581r CrossRefView Record in Scopus Lowry, n.d. Lowry, R., 2017. VassarStats: Web Site for Statistical Computation 〈[http://vassarstats.net]〉, 1998–2017. Two-tailed Fisher Exact Probability Test on the page For a 2×2 Contingency Table: [〈http://vassarstats.net/table2×2.html〉]. Lund et al., 2013 T.C. Lund, A.J. Kobs, A. Kramer, M. Nyquist, M.T. Kuroki, J. Osborn, D.S. Lidke, S.T. Low-Nam, B.R. Blazar, J. Tolar Bone marrow stromal and vascular smooth muscle cells have chemosensory capacity via bitter taste receptor expression PLoS One, 8 (2013), p. e58945, 10.1371/journal.pone.0058945 CrossRef Ma et al., 2010 T. Ma, C. Tan, H. Zhang, M. Wang, W. Ding, S. Li Bridging the gap between traditional Chinese medicine and systems biology: the connection of Cold Syndrome and NEI network Mol. Biosyst., 6 (2010), pp. 613-619, 10.1039/b914024g CrossRefView Record in Scopus Macpherson et al., 2005 L.J. Macpherson, B.H. Geierstanger, V. Viswanath, M. Bandell, S.R. Eid, S. Hwang, A. Patapoutian The pungency of garlic: activation of TRPA1 and TRPV1 in response to allicin Curr. Biol., 15 (2005), pp. 929-934, 10.1016/j.cub.2005.04.018 ArticleDownload PDFView Record in Scopus Mane et al., 2010 S. Mane, M. Vyas, B. Ravishankar, R.R. Dwivedi Evaluation of some objective parameters for Ushna and Sheeta Gunas based on pharmacological study Ayu, 31 (2010), pp. 147-152, 10.4103/0974-8520.72368 CrossRefView Record in Scopus Manson et al., 2014 M.L. Manson, J. Safholm, M. Al-Ameri, P. Bergman, A.-C. Orre, K. Sward, A. James, S.-E. Dahlen, M. Adner Bitter taste receptor agonists mediate relaxation of human and rodent vascular smooth muscle Eur. J. Pharmacol., 740 (2014), pp. 302-311, 10.1016/j.ejphar.2014.07.005 ArticleDownload PDFView Record in Scopus Martin and Sollars, 2017 L.J. Martin, S.I. Sollars Contributory role of sex differences in the variations of gustatory function J. Neurosci. Res., 95 (2017), pp. 594-603, 10.1002/jnr.23819 CrossRefView Record in Scopus McDonald, 2014 J.H. McDonald Handbook of Biological Statistics (3rd ed.), Sparky House Publishing, Baltimore, Maryland (2014) de Medeiros et al., 2015 P.M. de Medeiros, B.L. Santos Pinto, V.T. do Nascimento Can organoleptic properties explain the differential use of medicinal plants? Evidence from Northeastern Brazil J. Ethnopharmacol., 159 (2015), pp. 43-48, 10.1016/j.jep.2014.11.001 ArticleDownload PDF Mendonca et al., 2017 P. Mendonca, E. Taka, D. Bauer, M. Cobourne-Duval, K.F.A. Soliman The attenuating effects of 1,2,3,4,6 penta-O-galloyl-beta-d-glucose on inflammatory cytokines release from activated BV-2 microglial cells J. Neuroimmunol., 305 (2017), pp. 9-15, 10.1016/j.jneuroim.2017.01.011 ArticleDownload PDFView Record in Scopus Messer, 1987 E. Messer The hot and cold in Mesoamerican indigenous and hispanicized thought Soc. Sci. Med., 25 (1987), pp. 339-346 ArticleDownload PDFView Record in Scopus Milne, 1805 C. Milne A Botanical Dictionary. Elements of Systematic and Philosophical Botany H.D.Symonds (1805) Molares and Ladio, 2014 S. Molares, A. Ladio Medicinal plants in the cultural landscape of a Mapuche-Tehuelche community in arid Argentine Patagonia: an eco-sensorial approach J. Ethnobiol. Ethnomed., 10 (2014), p. 61, 10.1186/1746-4269-10-61 CrossRef Monier-Williams, 2002 M. Monier-Williams A Sanskrit-English Dictionary (15th ed.), Motilal Banarsidass, Delhi (2002) Murthy, 1994 S.K.R. Murthy Vagbhata's Astanga Hrdaya (Second. ed.), Krishnadas Academy, Varanasi (1994) Murthy, 2004 S.K.R. Murthy Bhavaprakasa ofBhavamisra (Reprint, 2. ed.), Chowkhamba Krishnadas Academy, Varanasi (2004) Murthy, 2006 S.K.R. Murthy Sarngadhara Samhita. A Treatise on Ayurveda (Sixth. ed.), Chaukhamba Sanskrit Pratishthan, Varanasi (2006) Nagarajan et al., 2015a M. Nagarajan, G.R. Kuruvilla, K.S. Kumar, P. Venkatasubramanian Pharmacology of Ativisha, Musta and their substitutes J. Ayurveda Integr. Med., 6 (2015), pp. 121-133, 10.4103/0975-9476.146551 View Record in Scopus Nagarajan et al., 2015b M. Nagarajan, G.R. Kuruvilla, K.S. Kumar, P. Venkatasubramanian Abhava pratinidhi dravya: a comparative phytochemistry of Ativisha, Musta and related species J. Ayurveda Integr. Med., 6 (2015), pp. 53-63, 10.4103/0975-9476.146550 View Record in Scopus Orchard and van Vuuren, 2017 A. Orchard, S. van Vuuren Commercial essential oils as potential antimicrobials to treat skin diseases Evid. Based Complement. Altern. Med., 2017 (2017), p. 4517971, 10.1155/2017/4517971 Ortiz de Montellano, 1986 B.R. Ortiz de Montellano Aztec medicinal herbs: evaluation of therapeutic effectiveness N.L. Etkin (Ed.), Plants in Indigenous Medicine & Diet: Biobehavioral Approaches, Redgrave Publishing Company, New York (1986), pp. 113-128 View Record in Scopus Pandey and Singh, 2017 A.K. Pandey, P. Singh The genus Artemisia: a 2012–2017 literature review on chemical composition, antimicrobial, Insecticidal and antioxidant activities of essential oils Medicines (Basel, Switz.), 4 (2017), 10.3390/medicines4030068 Pandey, 2005 G. Pandey Dravyaguna Vij nana (Third. ed.), Chowkhamba Krishnadas Academy, Varanasi (2005) Paolini et al., 2006 G.V. Paolini, R.H.B. Shapland, W.P. van Hoorn, J.S. Mason, A.L. Hopkins Global mapping of pharmacological space Nat. Biotechnol., 24 (2006), pp. 805-815, 10.1038/nbt1228 CrossRefView Record in Scopus Peier et al., 2002 A.M. Peier, A. Moqrich, A.C. Hergarden, A.J. Reeve, D.A. Andersson, G.M. Story, T.J. Earley, I. Dragoni, P. McIntyre, S. Bevan, A. Patapoutian A TRP channel that senses cold stimuli and menthol Cell, 108 (2002), pp. 705-715 ArticleDownload PDFView Record in Scopus Pieroni and Torry, 2007 A. Pieroni, B. Torry Does the taste matter? Taste and medicinal perceptions associated with five selected herbal drugs among three ethnic groups in West Yorkshire, Northern England J. Ethnobiol. Ethnomed., 3 (2007), p. 21, 10.1186/1746-4269-3-21 CrossRef Prutkin et al., 2000 J. Prutkin, V.B. Duffy, L. Etter, K. Fast, E. Gardner, L.A. Lucchina, D.J. Snyder, K. Tie, J. Weiffenbach, L.M. Bartoshuk Genetic variation and inferences about perceived taste intensity in mice and men Physiol. Behav. (2000), 10.1016/S0031-9384(00)00199-2 Qadri et al., 2012 Y.J. Qadri, A.K. Rooj, C.M. Fuller ENaCs and ASICs as therapeutic targets Am. J. Physiol. - Cell Physiol., 302 (2012), pp. C943-C965, 10.1152/ajpcell.00019.2012 CrossRefView Record in Scopus Raybould, 1998 H.E. Raybould Does your gut taste? Sensory transduction in the gastrointestinal tract News Physiol. Sci., 13 (1998), pp. 275-280 CrossRefView Record in Scopus Reinberger, 2006 S. Reinberger Bitter could be better Sci. Am. Mind. (2006), 10.1038/scientificamericanmind0606-56 Ren et al., 2009 X. Ren, L. Zhou, R. Terwilliger, S.S. Newton, I.E. de Araujo Sweet taste signaling functions as a hypothalamic glucose sensor Front. Integr. Neurosci., 3 (2009), p. 12, 10.3389/neuro.07.012.2009 CrossRefView Record in Scopus Reyes-Garcia, 2010 V. Reyes-Garcia The relevance of traditional knowledge systems for ethnopharmacological research: theoretical and methodological contributions J. Ethnobiol. Ethnomed., 6 (2010), p. 32, 10.1186/1746-4269-6-32 CrossRef Roper, 2014 S.D. Roper TRPs in taste and chemesthesis Handb. Exp. Pharmacol., 223 (2014), pp. 827-871, 10.1007/978-3-319-05161-1_5 CrossRefView Record in Scopus Saslis-Lagoudakis et al., 2011 C.H. Saslis-Lagoudakis, B.B. Klitgaard, F. Forest, L. Francis, V. Savolainen, E.M. Williamson, J.A. Hawkins The use of phylogeny to interpret cross-cultural patterns in plant use and guide medicinal plant discovery: an example from Pterocarpus (Leguminosae) PLoS One, 6 (2011), p. e22275, 10.1371/journal.pone.0022275 CrossRef Schiffman et al., 1992 S.S. Schiffman, M.S. Suggs, A.L. Sostman, S.A. Simon Chorda tympani and lingual nerve responses to astringent compounds in rodents Physiol. Behav., 51 (1992), pp. 55-63 ArticleDownload PDF Schobel et al., 2014 N. Schobel, D. Radtke, J. Kyereme, N. Wollmann, A. Cichy, K. Obst, K. Kallweit, O. Kletke, A. Minovi, S. Dazert, C.H. Wetzel, A. Vogt-Eisele, G. Gisselmann, J.P. Ley, L.M. Bartoshuk, J. Spehr, T. Hofmann, H. Hatt Astringency is a trigeminal sensation that involves the activation of G protein-coupled signaling by phenolic compounds Chem. Senses, 39 (2014), pp. 471-487, 10.1093/chemse/bju014 CrossRef Schuffenhauer and Jacoby, 2004 A. Schuffenhauer, E. Jacoby Annotating and mining the ligand-target chemogenomics knowledge space Drug Discov. Today BIOSILICO, 2 (2004), pp. 190-200, 10.1016/S1741-8364(04)02408-4 ArticleDownload PDFView Record in Scopus Sharma and Chandola, 2011 H. Sharma, H.M. Chandola Prameha in Ayurveda: correlation with obesity, metabolic syndrome, and diabetes mellitus. Part 1-etiology, classification, and pathogenesis J. Altern. Complement. Med., 17 (2011), pp. 491-496, 10.1089/acm.2010.0396 CrossRefView Record in Scopus Sharma, 2002 P. Sharma Dravyagunasutra (2nd ed.), Chaukhambha Sanskrit Sansthan, Varanasi (2002) Sharma et al., 2017 P. Sharma, R. Yi, A.P. Nayak, N. Wang, F. Tang, M.J. Knight, S. Pan, B. Oliver, D.A. Deshpande Bitter taste receptor agonists mitigate features of allergic asthma in mice Sci. Rep., 7 (2017), p. 46166, 10.1038/srep46166 CrossRef Sharma, 1998 P.V. Sharma Cakradatta (Second ed.), Chaukhambha, Varanasi (1998) Sharma and Dash, 2006 R. Sharma, B. Dash Caraka Samhita (2006th ed.), Chowkhamba Sanskrit Series Office, Varanasi (2006) Shepard JR, 2004 G.H. Shepard JR A sensory ecology of medicinal plant therapy in two Amazonian societies Am. Anthropol., 106 (2004), pp. 252-266, 10.1525/aa.2004.106.2.252 Snow, 2016 J. Snow Context effects in Western herbal medicine: fundamental to effectiveness? Explor. J. Sci. Heal., 12 (2016), pp. 55-62, 10.1016/j.explore.2015.10.004 ArticleDownload PDFView Record in Scopus Social Sciences Statistics software package, 2017. Social Sciences Statistics software package, 2017. (〈http://www.socscistatistics.com/tests/fisher/Default2.aspx〉) [WWW Document]. Sohn, 1894 C.E. Sohn Dictionary of the Active Principles of Plants: Alkaloids, Bitter Principles, Glucosides, their Sources, Nature, and Chemical Characteristics, with Tabular Summary, Classification Of Reactions, And Full Botanical and General Indexes Baillière, Tindall and Cox, London (1894) Sorokowska et al., 2014 A. Sorokowska, P. Sorokowski, T. Hummel Cross-cultural administration of an odor discrimination test Chemosens. Percept., 7 (2014), pp. 85-90, 10.1007/s12078-014-9169-0 CrossRefView Record in Scopus Spence, 2015 C. Spence Multisensory Flavor Perception Cell, 161 (2015), pp. 24-35, 10.1016/j.cell.2015.03.007 ArticleDownload PDFView Record in Scopus Stefan, 2005 E. Stefan Astanga Hrdayam/Vagbhata Performantica, Iasi (2005) Stevenson, 2014 R.J. Stevenson Flavor binding: its nature and cause Psychol. Bull., 140 (2014), pp. 487-510, 10.1037/a0033473 CrossRefView Record in Scopus Swamy et al., 2016 M.K. Swamy, M.S. Akhtar, U.R. Sinniah Antimicrobial properties of plant essential oils against human pathogens and their mode of action: an updated review Evid. Based Complement. Altern. Med., 2016 (2016), p. 3012462, 10.1155/2016/3012462 Tejada et al., 2017 S. Tejada, W.N. Setzer, M. Daglia, S.F. Nabavi, A. Sureda, N. Braidy, O. Gortzi, S.M. Nabavi Neuroprotective effects of Ellagitannins: a brief review Curr. Drug Targets, 18 (2017), pp. 1518-1528, 10.2174/1389450117666161005112002 View Record in Scopus Tong et al., 2016 F. Tong, J. Zhang, L. Liu, X. Gao, Q. Cai, C. Wei, J. Dong, Y. Hu, G. Wu, X. Dong Corilagin Attenuates radiation-induced brain injury in mice Mol. Neurobiol., 53 (2016), pp. 6982-6996, 10.1007/s12035-015-9591-6 CrossRefView Record in Scopus Trujillo et al., 1999 E. Trujillo, Alvarez de la Rosa, D. Mobasheri, A. Gonzalez, T. Canessa, C.M. Martin-Vasallo, P Sodium transport systems in human chondrocytes. II. expression of ENaC, Na+/K+/2Cl- cotransporter and Na+/H+ exchangers in healthy and arthritic chondrocytes Histol. Histopathol., 14 (1999), pp. 1023-1031 View Record in Scopus Uma et al., 2017 K. Uma, X. Huang, B.A. Kumar Antifungal effect of plant extract and essential oil Chin. J. Integr. Med., 23 (2017), pp. 233-239, 10.1007/s11655-016-2524-z CrossRefView Record in Scopus Venkatasubramanian et al., 2010 P. Venkatasubramanian, S.K. Kumar, V.S.N. Nair Cyperus rotundus, a substitute for Aconitum heterophyllum: studies on the Ayurvedic concept of Abhava Pratinidhi Dravya (drug substitution) J. Ayurveda Integr. Med., 1 (2010), pp. 33-39, 10.4103/0975-9476.59825 CrossRefView Record in Scopus Voigt et al., 2012 A. Voigt, S. Hübner, K. Lossow, I. Hermans-Borgmeyer, U. Boehm, W. Meyerhof Genetic labeling of Tas1r1 and Tas2r131 taste receptor cells in mice Chem. Senses, 37 (2012), pp. 897-911, 10.1093/chemse/bjs082 CrossRefView Record in Scopus Weimann and Heinrich, 1998 C. Weimann, M. Heinrich Concepts of medicinal plants among the Nahua of the Sierra de Zongolica, Veracruz (Mexico) Angew. Bot., 72 (1998), pp. 87-91 View Record in Scopus World Health Organisation, 2011 World Health Organisation Quality Control Methods for Herbal Materials (Updated ed. ed.), World Health Organization, Geneva (2011) Wu et al., 2015 S. Wu, Y. Yue, J. Li, Z. Li, X. Li, Y. Niu, J. Xiang, H. Ding Procyanidin B2 attenuates neurological deficits and blood-brain barrier disruption in a rat model of cerebral ischemia Mol. Nutr. Food Res., 59 (2015), pp. 1930-1941, 10.1002/mnfr.201500181 CrossRefView Record in Scopus Yamamoto and Ishimaru, 2013 K. Yamamoto, Y. Ishimaru Oral and extra-oral taste perception Semin. Cell Dev. Biol., 24 (2013), pp. 240-246, 10.1016/j.semcdb.2012.08.005 ArticleDownload PDFView Record in Scopus Yang et al., 2007 H.-Y. Yang, H.-K. Chang, J.-W. Lee, Y.-S. Kim, H. Kim, M.-H. Lee, M.-S. Shin, D.-H. Ham, H.-K. Park, H. Lee, C.-J. Kim Amygdalin suppresses lipopolysaccharide-induced expressions of cyclooxygenase-2 and inducible nitric oxide synthase in mouse BV2 microglial cells Neurol. Res., 29 (Suppl 1) (2007), pp. S59-S64, 10.1179/016164107×172248 View Record in Scopus Yates et al., 1999 Yates, D., Moore, D., McCabe, G., 1999. The Practice of Statistics, 1st Editio. ed. New York. © 2018 Elsevier B.V. All rights reserved.