Impact statement

Globally, there has been a significant shift in healthcare from a reactive to a proactive approach with an emphasis on P4 (Preventive, Predictive, Personalized, and Participatory) and precision medicine. Ayurveda, an ancient system of medicine, has been practicing this approach for thousands of years, but its integration with modern medicine in clinical settings is still limited. Currently, individuals who want to avail of both systems do so independently without an evidence-based informed choice. The two medical streams have different approaches to diagnosis, treatment, and therapy. Modern medicine focuses on treating diseases in an organ-specific manner, while Ayurveda takes a holistic perspective, treating the individual in a personalized manner based on their inherent constitution type “Prakriti.” Ayurvedic medicines consist of complex herbal formulations, different from modern medicine drugs that often prioritize active principles. Both systems have their merits, but they currently do not have much cross-communication or collaboration. A dialogue between Ayurveda and modern medicine can offer patients more options for managing their health. The emerging field of Ayurgenomics shows promise in bridging the two contrasting disciplines of medicine. Research in this area over the last two decades has provided a unifying ontological framework to explore the molecular basis of principles and practices. These have provided (a) a molecular basis of inter-individual differences between Prakriti that govern their differential health and disease trajectories, (b) methods for phenotype-based non-invasive methods of stratification of healthy individuals, (c) biomarkers and targets for early actionable interventions, and (d) platforms for evidence-based usage of Ayurvedic medicines. By combining the strengths of Ayurveda and modern medicine through Ayurgenomics, there is potential to enhance patient care and promote a more precise, integrative, and effective healthcare system that embraces preventive, predictive, personalized, and participatory aspects.

Human genetic individuality, health baselines, and disease trajectory: Modern medicine versus Ayurveda

Genomic advancements are blurring the boundaries between the specialties of modern medicine and emphasis has now shifted to a system-based network medicine (Loscalzo et al., Reference Loscalzo, Kohane and Barabasi2007; Loscalzo and Barabasi, Reference Loscalzo and Barabasi2011; Chen and Snyder, Reference Chen and Snyder2012; Noell et al., Reference Noell, Faner and Agustí2018). This has primarily been propelled by the large-scale genome and phenome-wide association studies (GWAS and PHEWAS) that have revealed extensive pleiotropy, as well as overlap of genes and networks between disparate diseases. Shared intermediate patho-phenotypes are observed between diseases and different endophenotypes within the same disease sometimes govern differential outcomes (Loscalzo et al., Reference Loscalzo, Kohane and Barabasi2007). The endophenotypes include inflammation, fibrosis, wound healing, immune response, cell proliferation, and apopotosis/necrosis that govern outcomes, such as stroke, hemolytic crisis, pain, and so on. The specificity for a disease could thus be a cumulative consequence of innate predisposition, modifier genes, and local environment. For example, in sickle cell anemia, with the same primary mutations a patient can have pain, hemolytic crisis, or thrombosis due to variability in modifier genes and environment such as dehydration, hypoxia, and infection and presently it is difficult to predict trajectories from primary mutations.

Evidence of extensive inter-individual variability is also apparent from observations of the numbers needed to treat (NNT) and numbers needed to harm (NNH) from a large number of randomized control trials (https://www.thennt.com/). This has clearly highlighted the need for stratification of population for most effective outcomes (Cook and Sackett, Reference Cook and Sackett1995). A large-scale study on millions of patients on three common diseases, diabetes, hypertension, and depression, have reported extensive heterogeneity in treatment practices of drugs with a trial-and-error approach (Hripcsak et al., Reference Hripcsak, Ryan, Duke, Shah, Park, Huser, Suchard, Schuemie, De Falco, Perotte, Banda, Reich, Schilling, Matheny, Meeker, Pratt and Madigan2016). These are changing the clinical trial design paradigm from conventional randomized controlled trials (RCTs) to n-of-1 trials wherein interventions are planned through a double-blinded randomized crossover trial conducted in a single patient (Lillie et al., Reference Lillie, Patay, Diamant, Issell, Topol and Schork2011).

A few aspects are now being appreciated in recent times (1) There is an enormous inter-individual variability that governs variable outcomes in health, disease as well as during management and we need to evolve methods for risk stratification (Khera et al., Reference Khera, Chaffin, Aragam, Haas, Roselli, Choi, Natarajan, Lander, Lubitz, Ellinor and Kathiresan2018; Gibson, Reference Gibson2019a, Reference Gibson2019b; Nagpal et al., Reference Nagpal, Tandon and Gibson2022). (2) Health cannot be defined by an absence of disease state and we need to evolve measures for defining health and individuality (Olson, Reference Olson2012; Topol, Reference Topol2019). (3) Yield of biomarker discovery and association studies could increase if there were methods to reduce phenotypic heterogeneity in reference controls (Olson, Reference Olson2012; Prasher et al., Reference Prasher, Gibson and Mukerji2016; Noell et al., Reference Noell, Faner and Agustí2018; Juyal et al., Reference Juyal, Pandey, Garcia, Negi, Gupta, Kumar, Bhat, Juyal and Thelma2022).

Ayurveda, in contrast, has a clear definition of health baselines that determines individuality, is scale-free (like waist to hip ratio), and govern inter-individual differences in response, susceptibility as well as disease and therapeutic outcomes (extensively reviewed in Prasher et al., Reference Prasher, Gibson and Mukerji2016). Individuals in any population can be stratified to seven broad constitution groups “Prakriti” that are fixed very early in life, remain invariant throughout lifetime (Sharma, Reference Sharma1981, Reference Sharma1999). These differences arise from a unifying organizing principle that determines an individual’s thresholds in health and perturbations during disease, predict risks for different diseases and allows optimization of therapy. Drugs or therapy are mapped to specific imbalances to restore homeostasis. All this is contextualized in the geo-spatio-temporal context keeping ethnicity and heritability into perspective. The understanding of the linkage between cause, features, and therapy through the common organizing principle forms the basis of its translation from principle to practice in an individualized perspective. This resonates with the current approach of precision and networked medicine (Prasher et al., Reference Prasher, Gibson and Mukerji2016) (see Box 1).

Ayurgenomics: A framework for probing ontological links between modern medicine and Ayurveda

Ayurgenomics is an emerging discipline that seeks to bridge and build an operational framework for integrating the principles of Ayurveda with modern medicine for translational applications (Mukerji and Prasher, Reference Mukerji and Prasher2011; Sethi et al., Reference Sethi, Prasher and Mukerji2011; Sreedevi et al., Reference Sreedevi, Sayed, Harikumar and Tripathy2016; Prasher et al., Reference Prasher, Varma, Kumar, Khuntia, Pandey, Narang, Tiwari, Kutum, Guin, Kukreti, Dash and Mukerji2017; Singh et al., Reference Singh, Gehlot and Agrawal2019; Wallace, Reference Wallace2020; Venkanna Rao, Reference Venkanna Rao2022). This uses the tools of genomic and related omics to explore and elucidate the molecular correlates of “tridosha,” the common organizing principle in Ayurveda.

Amongst the seven broad Prakriti types the extreme/predominant Vata, Pitta, Kapha are most readily distinguishable (Hankey, Reference Hankey2005; Prasher et al., Reference Prasher, Gibson and Mukerji2016). These display contrasting phenotyping attributes comprise 10% of the population and have predominance of either of doshas. Prakriti of individuals is assessed using a compendium of composite phenotypes that encompass anatomical, physiological, response, and activity-associated attributes. Nearly two decades of exploratory research carried out by different groups have provided molecular correlates of Prakriti from different functional hierarchies – phenomic, genetic, genomic, epigenomic, transcriptomic, metagenomic as well as at biochemical, immunological, and cellular levels, and most recently, at the physiological level (Bhushan et al., Reference Bhushan, Kalpana and Arvind2005; Prasher et al., Reference Prasher, Aggarwal, Mandal, Sethi, Deshmukh, Purohit, Sengupta, Khanna, Mohammad, Garg, Brahmachari and Mukerji2008; Aggarwal et al., Reference Aggarwal, Negi, Jha, Singh, Stobdan, Pasha, Ghosh, Agrawal, Prasher and Mukerji2010, Reference Aggarwal, Gheware, Agrawal, Ghosh, Prasher and Mukerji2015; Joshi et al., Reference Joshi, Ghodke and Patwardhan2011; Rotti et al., Reference Rotti, Raval, Anchan, Bellampalli, Bhale, Bharadwaj, Bhat, Dedge, Dhumal, Gangadharan, Girijakumari, Gopinath, Govindaraj, Halder, Joshi, Kabekkodu, Kamath, Kondaiah, Kukreja, Kumar, Nair, Nair, Nayak, Prasanna, Rashmishree, Sharanprasad, Thangaraj, Patwardhan, Satyamoorthy and Valiathan2014, Reference Rotti, Mallya, Kabekkodu, Chakrabarty, Bhale, Bharadwaj, Bhat, Dedge, Dhumal, Gangadharan, Gopinath, Govindaraj, Joshi, Kondaiah, Nair, Nair, Nayak, Prasanna, Shintre, Sule, Thangaraj, Patwardhan, Valiathan and Satyamoorthy2015; Satyamoorthy et al., Reference Satyamoorthy, Guruprasad, Nayak, Kabekkodu, Kukreja, Mallya, Nayak, Bhradwaj, Gangadharan, Prasanna, Raval, Kamath, Gopinath, Kondaiah and Satyamoorthy2014; Govindaraj et al., Reference Govindaraj, Nizamuddin, Sharath, Jyothi, Rotti, Raval, Nayak, Bhat, Prasanna, Shintre, Sule, Joshi, Dedge, Bharadwaj, Gangadharan, Nair, Gopinath, Patwardhan, Kondaiah, Satyamoorthy, Valiathan and Thangaraj2015; Tiwari et al., Reference Tiwari, Kutum, Sethi, Shrivastava, Girase, Aggarwal, Patil, Agarwal, Gautam, Agrawal, Dash, Ghosh, Juvekar, Mukerji and Prasher2017; Chauhan et al., Reference Chauhan, Pandey, Mondal, Gupta, Verma, Jain, Ahmed, Patil, Agarwal, Girase, Shrivastava, Mobeen, Sharma, Srivastava, Juvekar, Prasher, Mukerji and Dash2018; Chaudhari et al., Reference Chaudhari, Dhotre, Agarwal, Gondhali, Nagarkar, Lad, Patil, Juvekar, Sinkar and Shouche2019; Abbas et al., Reference Abbas, Kutum, Pandey, Dakle, Narang, Manchanda, Patil, Aggarwal, Bansal, Sharma, Chaturvedi, Girase, Srivastava, Juvekar, Dash, Prasher and Mukerji2020, Reference Abbas, Chaturvedi, Prakrithi, Pathak, Kutum, Dakle, Narang, Manchanda, Patil, Aggarwal, Girase, Srivastava, Kapoor, Gupta, Pandey, Juvekar, Dash, Mukerji and Prasher2022; Mobeen et al., Reference Mobeen, Sharma and Prakash2020; Chakraborty et al., Reference Chakraborty, Singhmar, Singh, Maulik, Patil, Agrawal, Mishra, Ghazi, Vats, Natarajan, Juvekar, Prasher and Mukerji2021; Shalini et al., Reference Shalini, Jnana, Sriranjini, Tanwar, Brand, Murali, Satyamoorthy and Gangadharan2021; Rani et al., Reference Rani, Rengarajan, Sethi, Khuntia, Kumar, Punera, Singh, Girase, Shrivastava, Juvekar, Pesala, Mukerji, Deepak and Prasher2022). These studies have been primarily conducted on healthy individuals of extreme and contrasting constitution types that are most easily discernible at the level of composite phenotypes, and who also have contrasting responses and vulnerability (Prasher et al., Reference Prasher, Varma, Kumar, Khuntia, Pandey, Narang, Tiwari, Kutum, Guin, Kukreti, Dash and Mukerji2017). There have been novel discoveries, iterative developments and insights as well as opportunities for translation into integrative medicine settings (Ahmad et al., Reference Ahmad, Kumar, Mabalirajan, Pattnaik, Aggarwal, Singh, Singh, Mukerji, Ghosh and Agrawal2012; Juyal et al., Reference Juyal, Negi, Wakhode, Bhat, Bhat and Thelma2012, Reference Juyal, Pandey, Garcia, Negi, Gupta, Kumar, Bhat, Juyal and Thelma2022; Gheware et al., Reference Gheware, Dholakia, Kannan, Panda, Rani, Pattnaik, Jain, Parekh, Enayathullah, Bokara, Subramanian, Mukerji, Agrawal and Prasher2021a, Reference Gheware, Panda, Khanna, Bhatraju, Jain, Sagar, Kumar, Singh, Kannan, Subramanian, Mukerji, Agrawal and Prasher2021b; Haider et al., Reference Haider, Dholakia, Panwar, Garg, Gheware, Singh, Singhal, Burse, Kumari, Sharma, Ray, Medigeshi, Sharma, Prasher and Mukerji2021, Reference Haider, Anand, Enayathullah, Parekh, Ram, Kumari, Anmol, Shukla, Dholakia, Ray, Bhattacharyya, Sharma, Bokara, Prasher and Mukerji2022; Table 1).

Table 1. Studies of predominant Prakriti types in healthy individuals at different omic levels

Note: These studies have been carried out on healthy individuals of predominant and contrasting Prakriti types – Vata (V), Pitta (P), and Kapha (K) from diverse cohorts. The number of subjects, population details, and gender information are included. A brief summary of key results with few examples in each of the studies has been highlighted. The references for the studies are also included. *A few candidate gene-based studies are included. eQTL, expression quantitative trait loci; HAPE, high altitude pulmonary edema; HGDP, human genome diversity panel; IGV, Indian Genome Variation Consortium Project. Up and dn indicates the direction of expression/levels towards positive and negative, respectively.

Ayurgenomics-based studies have provided the following insights for application (1) a compendium of composite system-level non-invasive phenotypes that could be assessed in healthy individuals to predict inherent risk for different diseases, exposures, and also for monitoring health trajectories, (2) molecular correlates of “dosha” mapping to intermediate patho-phenotypes that could enable setting of individualized baselines in health, (3) a method to enrich for informative variations and molecular axes that could enable development of marker panels as well as be useful for computation of individual risk scores for early and stratified interventions, (4) development of preclinical disease models for molecular delineation of dosha specific interventions, (5) operational frameworks for network medicine based on trisutra principles and leads, and (6) Evidence-based solutions for Ayurveda drug usage, actionable targets and novel molecules for drug development and repurposing. The application space of Ayurgenomics and the existing needs for technology developments is illustrated with specific examples in the subsequent sections (Figure 1).

Figure 1. Summary of research highlights and potential of Ayurgenomics. Molecular correlates of Tridosha, the common organizing principle in Ayurveda have been explored through studies on healthy individual of three extreme Prakriti constitution types that have predominance of either Vata, Pitta, or Kapha doshas as well as have been validated in model systems. These have been studied at various levels of functional hierarchy from molecular (genetic and epigenetic) to expression and response at cellular (transcriptomic, metagenomic, biochemical) and organ – physiome level and also with respect to multisystem phenotypic attributes in diverse populations. These have been validated through cell-based studies, in LCL and disease models (cell-based and mouse), and machine learning approaches. The studies have highlighted the application of Ayurgenomics for elucidating PhenotypetoPhenotype linkages in an individual, understanding the genetic and phenotypic basis of human individuality, identification of different baselines of intermediate patho-phenotypes that govern outcomes in health, genetic variations, and biomarkers for risk and response stratification in healthy individuals, novel therapeutic targets, systems’ understanding and drug repurposing through trisutra framework. These could be useful for noninvasive risk stratification of healthy individual, endophenotype identification in diseases, evolving polygenic risk scores from GWAS studies for more effective risk stratification, identification of actionable targets and genes with pleiotropic effects from PHEWAS cohorts, development of preclinical models for exploring and generating evidence-based usage Ayurveda formulations and in developing endophenotype based precision and networked medicine.

Molecular correlates from studies on extreme Prakriti types could help define baselines of health

There is a common underlying theme to all diseases. In response to diverse environmental triggers that could include heat, hypoxia, Ultraviolet, infection, and dehydration, individuals are either resilient or could present with variable outcomes, such as thrombosis, bleeding, fibrosis, and pain, which are mainly restricted to a specific organ (Loscalzo et al., Reference Loscalzo, Kohane and Barabasi2007; Noell et al., Reference Noell, Faner and Agustí2018). These diseases manifest through intermediate endophenotypes, such as immune response, inflammation, cell adhesion, cell proliferation, apoptosis, necrosis, and so on. A disease in an individual is thus a cumulative effect of rare and common variations in highly penetrant, as well as, modifier genes along with epigenetic effects. Threading the genotypes to variable phenotypic outcome through the different levels of functional hierarchy in an individualized manner is a major aim of precision medicine.

There has been a massive deluge of variation data from whole genome and exome sequencing, as well as genome-wide association studies from diverse populations. Methods are now being evolved to develop polygenic risk scores (PRS) that take into account the cumulative involvement of risk/protective alleles for more effective stratification in precision medicine (Khera et al., Reference Khera, Chaffin, Aragam, Haas, Roselli, Choi, Natarajan, Lander, Lubitz, Ellinor and Kathiresan2018; Gibson, Reference Gibson2019b; Nagpal et al., Reference Nagpal, Tandon and Gibson2022). Genome-wide expression studies on healthy individuals have also revealed high inter-individual variance within a population (Montgomery and Dermitzakis, Reference Montgomery and Dermitzakis2011; Martin et al., Reference Martin, Costa, Lappalainen, Henn, Kidd, Yee, Grubert, Cann, Snyder, Montgomery and Bustamante2014). A major challenge, however, is threading genotypes to phenotypes through enormous genetic heterogeneity within populations. The effect of common genetic variants associated with many diverse diseases in GWAS studies is often masked in this heterogeneity as there are no adequate methods to comprehensively define and distinguish healthy individuals within a population (Manolio et al., Reference Manolio, Collins, Cox, Goldstein, Hindorff, Hunter, MI, Ramos, Cardon, Chakravarti, Cho, Guttmacher, Kong, Kruglyak, Mardis, Rotimi, Slatkin, Valle, Whittemore, Boehnke, Clark, Eichler, Gibson, Haines, Mackay, SA and Visscher2009; Eichler et al., Reference Eichler, Flint, Gibson, Kong, Leal, Moore and Nadeau2010). Using millions of electronic health records that are available in cohorts on whom genome-wide association studies have been conducted for specific traits, phenotype–phenotype associations are being derived through phenome-wide association studies (Denny et al., Reference Denny, Ritchie, Basford, Pulley, Bastarache, Brown-Gentry, Wang, Masys, Roden and Crawford2010). The presence of overlapping phenotypes that are captured in health records has enabled the discovery of pleiotropic variants that connect different diseases and phenotypes through shared variants.

Genome-wide expression, genetic variation (microarray based as well as exome), and epigenome (global methylation and array based) studies on extreme Prakriti healthy individuals that have predominance of one of the doshas, have provided molecular correlates of tridosha (Prasher et al., Reference Prasher, Aggarwal, Mandal, Sethi, Deshmukh, Purohit, Sengupta, Khanna, Mohammad, Garg, Brahmachari and Mukerji2008; Govindaraj et al., Reference Govindaraj, Nizamuddin, Sharath, Jyothi, Rotti, Raval, Nayak, Bhat, Prasanna, Shintre, Sule, Joshi, Dedge, Bharadwaj, Gangadharan, Nair, Gopinath, Patwardhan, Kondaiah, Satyamoorthy, Valiathan and Thangaraj2015; Rotti et al., Reference Rotti, Mallya, Kabekkodu, Chakrabarty, Bhale, Bharadwaj, Bhat, Dedge, Dhumal, Gangadharan, Gopinath, Govindaraj, Joshi, Kondaiah, Nair, Nair, Nayak, Prasanna, Shintre, Sule, Thangaraj, Patwardhan, Valiathan and Satyamoorthy2015; Abbas et al., Reference Abbas, Kutum, Pandey, Dakle, Narang, Manchanda, Patil, Aggarwal, Bansal, Sharma, Chaturvedi, Girase, Srivastava, Juvekar, Dash, Prasher and Mukerji2020, Reference Abbas, Chaturvedi, Prakrithi, Pathak, Kutum, Dakle, Narang, Manchanda, Patil, Aggarwal, Girase, Srivastava, Kapoor, Gupta, Pandey, Juvekar, Dash, Mukerji and Prasher2022). Some of the highlights of these studies have been provided in Table 1. Contrasting doshas exhibit differences with respect to gene-ontology enrichments. In the first of such studies carried out on healthy individuals the authors reported ontological differences between the individuals of contrasting constitution types at genome-wide expression and biochemical levels (Prasher et al., Reference Prasher, Aggarwal, Mandal, Sethi, Deshmukh, Purohit, Sengupta, Khanna, Mohammad, Garg, Brahmachari and Mukerji2008). Corroborating observations of differences in ontological enrichments between different constitution types have been reported from multiple studies that have been carried out at different functional hierarchies and in different population cohorts. For instance cell proliferation and DNA damage response differentiates Vata Prakriti, T-cell mediated immunity, elevated metabolism and inflammation in Pitta Prakriti, and lipid profiles and BMI-associated correlates differentiate Kapha have been observed from expression, epigenetic as well as genetic studies (Prasher et al., Reference Prasher, Aggarwal, Mandal, Sethi, Deshmukh, Purohit, Sengupta, Khanna, Mohammad, Garg, Brahmachari and Mukerji2008; Rotti et al., Reference Rotti, Mallya, Kabekkodu, Chakrabarty, Bhale, Bharadwaj, Bhat, Dedge, Dhumal, Gangadharan, Gopinath, Govindaraj, Joshi, Kondaiah, Nair, Nair, Nayak, Prasanna, Shintre, Sule, Thangaraj, Patwardhan, Valiathan and Satyamoorthy2015; Abbas et al., Reference Abbas, Kutum, Pandey, Dakle, Narang, Manchanda, Patil, Aggarwal, Bansal, Sharma, Chaturvedi, Girase, Srivastava, Juvekar, Dash, Prasher and Mukerji2020, Reference Abbas, Chaturvedi, Prakrithi, Pathak, Kutum, Dakle, Narang, Manchanda, Patil, Aggarwal, Girase, Srivastava, Kapoor, Gupta, Pandey, Juvekar, Dash, Mukerji and Prasher2022). Besides these, there are differences in different cellular and physiological axes that are associated with development, cell adhesion, signaling, and transport functions as well as processes that govern circadian rhythm, olfaction, and so on. Differences in gut microbiome that are correlated with many disease predispositions have also been reported amongst Prakriti type (Chauhan et al., Reference Chauhan, Pandey, Mondal, Gupta, Verma, Jain, Ahmed, Patil, Agarwal, Girase, Shrivastava, Mobeen, Sharma, Srivastava, Juvekar, Prasher, Mukerji and Dash2018; Chaudhari et al., Reference Chaudhari, Dhotre, Agarwal, Gondhali, Nagarkar, Lad, Patil, Juvekar, Sinkar and Shouche2019; Mobeen et al., Reference Mobeen, Sharma and Prakash2020; Shalini et al., Reference Shalini, Jnana, Sriranjini, Tanwar, Brand, Murali, Satyamoorthy and Gangadharan2021). Recently a study on two cohorts has also reported physiological difference in response to orthostatic stress that was evaluated by heart rate variability (Rani et al., Reference Rani, Rengarajan, Sethi, Khuntia, Kumar, Punera, Singh, Girase, Shrivastava, Juvekar, Pesala, Mukerji, Deepak and Prasher2022). All these studies substantiate that the assessment of extreme Prakriti can be used to assess the proportions of different “doshas,” their dynamics in a healthy individual, and predict the associated risk and trajectories. Noteworthy, the healthy individuals displayed baseline differences in the processes linked to intermediate patho-phenotypes that are often perturbed in diseases (Prasher et al., Reference Prasher, Aggarwal, Mandal, Sethi, Deshmukh, Purohit, Sengupta, Khanna, Mohammad, Garg, Brahmachari and Mukerji2008).

It seems probable that differences in baselines of intermediate phenotypes of healthy individuals could be governed by a need to maintain a homeostatic state in an individual and subvert the deleterious consequences of an epistatic variant. For example, in one of our studies, we demonstrated that the elevated bleeding risk in Pitta could be a protective phenotype to subvert the epistatic effect of a variation in oxygen sensor gene EGLN1 that contributes to higher baseline levels of hypoxia responsiveness (Aggarwal et al., Reference Aggarwal, Negi, Jha, Singh, Stobdan, Pasha, Ghosh, Agrawal, Prasher and Mukerji2010, Reference Aggarwal, Gheware, Agrawal, Ghosh, Prasher and Mukerji2015). Hypoxic conditions lead to increase in angiogenesis and enhanced formation of platelet glycoprotein as a protective phenotype. However, in a chronic state of hypoxia this could put an individual to enhanced risk of thrombosis. In individuals where baseline levels of hypoxia are higher, a bleeding-linked state could thus be physiologically favorable. In another study, we demonstrate how elevated baseline levels of cell proliferation Vata Prakriti could be advantageous if the individuals have an inherent sensitivity to DNA-damaging conditions (Chakraborty et al., Reference Chakraborty, Singhmar, Singh, Maulik, Patil, Agrawal, Mishra, Ghazi, Vats, Natarajan, Juvekar, Prasher and Mukerji2021). An inherent susceptibility to DNA-damaging agents at the embryonic stage might have led to a genetic rewiring that ensures higher baseline states of cell proliferation for more effective recovery. This, however, later in life might contribute to differences in outcomes due to the involvement of the cell proliferation axes.

The above observations suggest if interventions are targeted toward elevated levels of intermediate patho-phenotypes without taking cognizance of an individual’s baselines, this could lead to ineffective outcomes in diseases or even drastic consequences. For instance, ischemic consequences in hypoxia might differ amongst individuals based on their constitution types and would have different treatment and dose calibration requirements. Also, DNA-damaging agents in cancer conditions, if calibrated on the basis of thresholds of DNA damage sensitivity of cancerous tissues without a cognizance of inherently baseline proliferation rates in response to such conditions, could result in higher recurrence in some individuals. Could different constitution types govern such differences? Interestingly, according to textual descriptions there are many intermediate pathophenotype that are described to be Prakriti specific. For instance, amongst many distinguishing features, pain-related outcomes are more prevalent in individuals with heightened Vata, bleeding in Pitta, and thrombosis in Kapha Prakriti. Of interest, there are also textual descriptions that describe certain biological processes being enriched in particular doshas, for example, enhanced cell proliferation in Vata and inflammation in Pitta and adhesion in Kapha. Many of the molecular observations resonate with descriptions of Ayurveda (Tables 1 and 2).

Table 2. Prakriti (dosha) associated differential functional outcomes in baseline physiology, stress response, and diseases

Since we observe hypoxia responsiveness as an axis that differs between Pitta and Kapha and also govern differences in bleeding versus thrombotic risks, it is testable whether we might be able to predict trajectories of sickle cell patients based on their Prakriti types. There are a number of modalities through lifestyle, dietary as well as drugs through which doshic imbalances are managed. It might therefore be possible to manage the quality of life through dosha modulating formulations once the Prakriti phenotypes are assessed, and is testable in retrospective cohorts. The relevance of this is borne out by the most recent observations across the world where there was a wide variability in hypoxia response and inflammatory consequences even between related family members. Could these be governed by constitutional types?

Extreme Prakriti genomics enables the enrichment of genetic markers associated with pleiotropic effects, differential susceptibilities, drug, and environmental response

Exome sequencing of healthy individuals of extreme Prakriti types has provided some interesting insights (Abbas et al., Reference Abbas, Kutum, Pandey, Dakle, Narang, Manchanda, Patil, Aggarwal, Bansal, Sharma, Chaturvedi, Girase, Srivastava, Juvekar, Dash, Prasher and Mukerji2020, Reference Abbas, Chaturvedi, Prakrithi, Pathak, Kutum, Dakle, Narang, Manchanda, Patil, Aggarwal, Girase, Srivastava, Kapoor, Gupta, Pandey, Juvekar, Dash, Mukerji and Prasher2022). Extreme Prakriti types in a population are rare and are often the most vulnerable to specific types of diseases. Genetic, genome-wide arrays and sequencing have revealed the enrichment of biological processes in Prakriti specific manner (Table 1). Variations that govern differences in expression (eQTLs) were also enriched in Prakriti specific manner amongst healthy individuals and govern differences with respect to environmental triggers for example hypoxia, UV, DNA damage, and so on. Also, a significant fraction of GWAS-associated variants (that are normally identified from disease associations) differ in frequency between the Prakriti groups. When the Prakriti groups are pooled their frequency gets averaged and assumed to be a similar frequency as the background population. Since variations associated with GWAS traits are common in the population, extreme Prakriti could be individuals who are most predisposed, thus, this method of phenotyping might allow identification of predisposed individuals and associated genes (Table 2). Many of these variants are also reported to have pleiotropic effects as evidenced from their overlap with reported variants in PHEWAS studies and also drug targets. This shows it is possible to identify actionable variations and associated composite phenotypes based on Prakriti for risk stratification. Similar to genome-wide expression, exome sequencing also revealed specific ontological enrichments in differentiating genes for specific Prakriti types. This approach could also be useful in pharmacogenomics settings as the variations that are linked to differences in drug metabolisms differ between Prakriti types (Joshi et al., Reference Joshi, Ghodke and Patwardhan2011; Prasher et al., Reference Prasher, Varma, Kumar, Khuntia, Pandey, Narang, Tiwari, Kutum, Guin, Kukreti, Dash and Mukerji2017). Some of the prominent examples include human leukocyte antigen (HLA) types as well as genes in whom variations have been approved by FDA for optimizing drug dosage, for example, clopidogrel, bupropion, warfarin, abacavir, and so on. Interestingly, some of the variations in genes such as VWF, F2, and SERPINA10 genes associated with specific risks, of bleeding and thrombosis were enriched in Pitta and Kapha Prakriti respectively. Such risks are described for these Prakriti groups in Ayurveda. Sequencing of phenotypic extremes is assuming immense importance in the identification of disease susceptible genes as well as drug targets (Harper et al., Reference Harper, Nayee and Topol2015; Heck et al., Reference Heck, Milnik, Vukojevic, Petrovska, Egli, Singer, Escobar, Sengstag, Coynel, Freytag, Fastenrath, Demougin, Loos, Hartmann, Schicktanz, Bizzini, Vogler, Kolassa, Wilker, Elbert, Schwede, Beisel, Beerenwinkel, de Quervain and Papassotiropoulos2017). Such a sequencing strategy in Prakriti provides an added advantage of linkage with multisystem phenotypes.

Prakriti based noninvasive risk stratification in phenomics studies

Previous studies described emphasize that healthy individuals stratified on the basis of Prakriti type could be useful for early identification of at-risk individuals who might have different management needs. As described earlier, Prakriti types are distinguished on the basis of system-level composite phenotypes. Integration of knowledge of these composite phenotypes that are connected in an individual-specific manner could enable an affordable and noninvasive approach of stratification across large populations. Since Prakriti of an individual remains invariant throughout lifetime, a single assessment of an individual is highly affordable and can also be implemented in prospective cohorts (Juyal et al., Reference Juyal, Negi, Wakhode, Bhat, Bhat and Thelma2012, Reference Juyal, Pandey, Garcia, Negi, Gupta, Kumar, Bhat, Juyal and Thelma2022). Conjoint assessment with health and disease indications could be highly revealing. This may no longer seem incongruous to the modern audience as there is now ample evidence accumulating that demonstrates the association of noninvasive phenotypes with disease risk. For instance, the relatively innocuous measures, like the 2D/4D finger length ratio is highly informative in predicting many diverse outcomes, such as risk for cardiovascular diseases, Attention deficit hyperactivity disorder ADHD, and behavioral traits, and so on (Manning et al., Reference Manning, Scutt, Wilson and Lewis-Jones1998; Manning et al., Reference Manning, Baron-Cohen, Wheelwright and Sanders2001; Manning et al., Reference Manning, Morris and Caswell2007; Coates et al., Reference Coates, Gurnell and Rustichini2009; Fischer Pedersen et al., Reference Fischer Pedersen, Klimek, Galbarczyk, Nenko, Sobocki, Christensen and Jasienska2021).

The composite nature of these phenotypes in an individual-specific manner has also been recapitulated by unsupervised machine learning and advanced statistics approaches in a study on two cohorts (Tiwari et al., Reference Tiwari, Kutum, Sethi, Shrivastava, Girase, Aggarwal, Patil, Agarwal, Gautam, Agrawal, Dash, Ghosh, Juvekar, Mukerji and Prasher2017). This has enabled the development of predictive models and the identification of classifiers that distinguish individuals with contrasting doshic proportions. Evidence of phenotypes to phenotype connectivity is also being uncovered through phenome-wide association studies (PHEWAS). The EHR records in millions of subjects are used to create a picture of composite phenotype through a common genetic lead and associated overlapping phenotypes (Denny et al., Reference Denny, Ritchie, Basford, Pulley, Bastarache, Brown-Gentry, Wang, Masys, Roden and Crawford2010, Reference Denny, Bastarache, Ritchie, Carroll, Zink, Mosley, Field, Pulley, Ramirez, Bowton, Basford, Carrell, Peissig, Kho, Pacheco, Rasmussen, Crosslin, Crane, Pathak, Bielinski, Pendergrass, Xu, Hindorff, Li, Manolio, Chute, Chisholm, Larson, Jarvik, Brilliant, CA, Kullo, Haines, Crawford, Masys and Roden2013; Pendergrass and Ritchie, Reference Pendergrass and Ritchie2015; Liu and Crawford, Reference Liu and Crawford2022). Integration of assessment of composite phenotypes/Prakriti in these cohorts can accelerate the phenome assembly process by providing a phenotype scaffold in a manner analogous to genome sequence assemblies (Abbas et al., Reference Abbas, Kutum, Pandey, Dakle, Narang, Manchanda, Patil, Aggarwal, Bansal, Sharma, Chaturvedi, Girase, Srivastava, Juvekar, Dash, Prasher and Mukerji2020, Reference Abbas, Chaturvedi, Prakrithi, Pathak, Kutum, Dakle, Narang, Manchanda, Patil, Aggarwal, Girase, Srivastava, Kapoor, Gupta, Pandey, Juvekar, Dash, Mukerji and Prasher2022). Just like in the absence of a scaffold, a de novo genome assembly requires millions of reads, PHEWAS studies require millions of health records for phenome assemblies.

Prakriti phenotyping in retrospective cohorts could be an alternative route, for the identification of endophenotypes that could govern differential outcomes (Table 2). The importance of Prakriti assessment in disease cohorts is borne out of two studies on rheumatoid arthritis that demonstrate conditioning disease groups with Prakriti labels allows identification of variants with higher effect sizes and also different endophenotypes (Juyal et al., Reference Juyal, Negi, Wakhode, Bhat, Bhat and Thelma2012, Reference Juyal, Pandey, Garcia, Negi, Gupta, Kumar, Bhat, Juyal and Thelma2022; Table 2). These are also supported by genetic and exome studies that demonstrate risk alleles that associate with different diseases are partitioned differently amongst the healthy individual of different constitutions. In a population, the frequency of these risk alleles is masked due to extensive phenotypic heterogeneity (Aggarwal et al., Reference Aggarwal, Negi, Jha, Singh, Stobdan, Pasha, Ghosh, Agrawal, Prasher and Mukerji2010; Abbas et al., Reference Abbas, Chaturvedi, Prakrithi, Pathak, Kutum, Dakle, Narang, Manchanda, Patil, Aggarwal, Girase, Srivastava, Kapoor, Gupta, Pandey, Juvekar, Dash, Mukerji and Prasher2022).

Although the composite phenotypes in individuals of extreme phenotypes are readily distinguishable and explainable, methods are needed to objectively define the dual types that comprise a major fraction of the population (~90%) and have mixture of different doshic proportions. Since the doshas govern physiological entities, it is likely that some emergent phenotypes in the dual types could help distinguish and stratify these groups. Large-scale phenotyping using digital devices that can capture the compendium of phenotypes described for Prakriti in diverse settings could be of enormous utility. As a large number of combinatorial possibilities exist for dual Prakriti, the development of AI-based predictive models for objectively assigning Prakriti in large datasets is required. These could be attempted in large cohorts that are being developed by the government of India initiatives, which aim to integrate Prakriti information in health cards and involves phenotyping individuals across the country in AYUSH health and wellness centers using uniform protocols.

Molecular correlates in Trisutra framework: Platform for evidence-based usage of Ayurveda formulation and drug repurposing

Most often a herb that forms a constituent of a medicinal formulation in Ayurveda is used in drug discovery programs. However, the therapeutic approach in Ayurveda contrasts with the conventional pharmacological reductionist approach that is inspired from herbal formulations. In the latter, the focus is mostly to identify the active principle that could be further synthesized chemically, and then taken forth for further mechanistic and preclinical/clinical studies. This approach though has been useful for the discovery of many small molecules, has not been able to provide an understanding of their mechanisms from a clinical and systems perspective. A drug in such a form (with only the active principle) cannot be prescribed in an Ayurveda setting. Drugs typically are discovered with a particular disease focus, but now with thousands of genome-wide association studies, it is becoming evident that there is an extensive disease gene network. Different diseases could have shared mutations and genes. In most complex diseases ultimately a patient has to have a polypill which could be either multiple medicines (modern) or a formulation with multiple active principles (Ayurveda). There is a large focus on research on drug repurposing to reduce the time and cost for drug discovery as has been evident throughout the COVID-19 pandemic.

In Ayurveda, a systemic understanding seems to be implicitly applied in practice as a single herb is most often used in multiple and diverse conditions, and a single clinician can treat multiple diseases (Patwardhan and Mashelkar, Reference Patwardhan and Mashelkar2009). Also, the treatment strategies seem to consider emergent properties in terms of the disease, as well as diseased states within a spatio-temporal context. In today’s precision medicine, this is the ultimate aim. The “trisutra” framework resonates with the network medicine concept where the features (phenotypes), causes (diseases), and therapy (drugs) are unified through a common organizing principle of tridosha. If tridosha has a molecular correlate then it is possible to query this network from any of the axes. The two following examples would demonstrate how this is testable and applicable for translation, where the query is initiated from the phenotype and the other from the drug perspective.

In the first example, a gene “EGLN1” that is an oxygen sensor and regulates the hypoxia responsiveness axis through HIF1 differed between two constitution types “Pitta and Kapha” amongst healthy individuals at the expression and genetic level (Aggarwal et al., Reference Aggarwal, Negi, Jha, Singh, Stobdan, Pasha, Ghosh, Agrawal, Prasher and Mukerji2010). This genetic difference contributes to either adaptation to high altitude or susceptibility to high altitude pulmonary edema. A small molecular inhibitor of EGLN1 in a mouse model of asthma revealed that modulation of hypoxia axis could lead to airway hyper-responsiveness, exacerbate inflammatory consequences and mitochondrial dysfunction and elicit steroid-resistant asthma features (Ahmad et al., Reference Ahmad, Kumar, Mabalirajan, Pattnaik, Aggarwal, Singh, Singh, Mukerji, Ghosh and Agrawal2012). A herbal formulation, Adathoda vasica, widely used to treat asthma and is described in Ayurveda to modulate perturbations of the Pitta-Kapha axis was further tested in the above model (Gheware et al., Reference Gheware, Panda, Khanna, Bhatraju, Jain, Sagar, Kumar, Singh, Kannan, Subramanian, Mukerji, Agrawal and Prasher2021b). The hypothesis was if in Ayurveda, medicines are selected on the basis of doshic imbalances, then the molecular correlates to the doshic imbalances would be actionable points. When used in the above model, Adathoda vasica alleviated all the above conditions including inflammation, restored mitochondrial dysfunction and was validated as a HIF1a inhibitor. The effect of the formulation in the hypoxia axis was also evident in the lung transcriptome, thus explaining why this medicine could work in other diseases where hypoxia is a cause or consequence, and could be repurposed in preclinical models of sepsis and COVID-19 (Gheware et al., Reference Gheware, Dholakia, Kannan, Panda, Rani, Pattnaik, Jain, Parekh, Enayathullah, Bokara, Subramanian, Mukerji, Agrawal and Prasher2021a). This entire framework provided EGLN1 as a biomarker, which is now been widely accepted to be a target a large number of diseases and associated genetic variations have been replicated in multiple studies and altitudes of the world (Simonson et al., Reference Simonson, Yang, Huff, Yun, Qin, Witherspoon, Bai, Lorenzo, Xing, Jorde, Prchal and Ge2010; Brutsaert et al., Reference Brutsaert, Kiyamu, Revollendo, Isherwood, Lee, Rivera-Ch, Leon-Velarde, Ghosh and Bigham2019). Differences in the wound healing axis between Pitta and Kapha Prakriti have been corroborated in genome-wide expression and exome studies (Prasher et al., Reference Prasher, Aggarwal, Mandal, Sethi, Deshmukh, Purohit, Sengupta, Khanna, Mohammad, Garg, Brahmachari and Mukerji2008; Abbas et al., Reference Abbas, Chaturvedi, Prakrithi, Pathak, Kutum, Dakle, Narang, Manchanda, Patil, Aggarwal, Girase, Srivastava, Kapoor, Gupta, Pandey, Juvekar, Dash, Mukerji and Prasher2022). Adhathoda vasica has been extensively studied from the active principle perspective, but this is the first time that a complete formulation was demonstrated to work on the hypoxia axis. This provides a unifying link to its usage in multiple indications and in all instances hypoxia could be a primary determinant. Thus the trisutra framework for translation of concepts of ayurveda that is linkage between feature-cause-therapy could be established through a molecular correlate (EGLN1) of tridosha (the common organizing principle) (Figure 2).

Figure 2. Iterative approach of Ayurgenomcs for the development of an operational framework of network medicine based on concepts of Ayurveda. In Ayurveda, the three axes of causes, features, and therapy (trisutra) are linked through a common organizing principle (tridosha) that govern three physiological entities described broadly by kinetic (Vata), metabolic (Pitta), and structural (Kapha) attributes respectively. Tridosha feature descriptions resonate with modern GO (gene ontology) descriptors. Relative proportion of tridosha governs human constitutions (Prakriti) with extreme types having predominance of one of the doshas. The extreme individual Prakriti have contrasting phenotypes, responses and health trajectories and are more vulnerable to specific groups of diseases. Multiomic studies on extreme constitution reveal molecular correlate of Tridosha that have different GO and pathway enrichments and are associated with contrasting susceptibilities, environmental response, and drugs. Perturbation of the dosha associated genes in model systems provide molecular evidence for dosha involvement in disease states (Vikriti) through specific axis. This provides a test bed for exploring mechanism of action of herbal formulation used for modulation of specific disease and dosha perturbations. The above framework has been validated with respect to hypoxia axis through this iterative approach and is described in the text. It has also provided mechanism of action of Cessamplous pareira from its usage perspective and evidence for recommendation in COVID-19 both from its effect on host and pathogen.

A second way to probe the networked concept of trisutra could be from the therapy perspective. In Ayurveda, a drug can be used for multiple and seemingly diverse diseases. This could be because multiple molecules in a formulation are targeting different diseases, or that a shared pathway connects diverse diseases. Our group took an example drug, Cessamplous pareira (CIPA) that is widely used to treat fever-like conditions and hormonal disturbance in women to test this hypothesis. We tested this through a connectivity map (CMAP) framework. Connectivity maps provide a compendium of transcriptional signature profiles from millions of perturbations (small molecules, drugs, gene knock outs, over-expression) from different cell lines (Lamb et al., Reference Lamb, Crawford, Peck, Modell, Blat, Wrobel, Lerner, Brunet, Subramanian, Ross, Reich, Hieronymus, Wei, Armstrong, Haggarty, Clemons, Wei, Carr, Lander and Golub2006; Subramanian et al., Reference Subramanian, Narayan, Corsello, Peck, Natoli, Lu, Gould, Davis, Tubelli, Asiedu, Lahr, Hirschman, Liu, Donahue, Julian, Khan, Wadden, Smith, Lam, Liberzon, Toder, Bagul, Orzechowski, Enache, Piccioni, Johnson, Lyons, Berger, Shamji, Brooks, Vrcic, Flynn, Rosains, Takeda, Hu, Davison, Lamb, Ardlie, Hogstrom, Greenside, Gray, Clemons, Silver, Wu, Zhao, Read-Button, Wu, Haggarty, Ronco, Boehm, Schreiber, Doench, Bittker, Root, Wong and Golub2017). One can use a test drug and compare the signatures to infer possible modes of action of a novel compound or a mixed formulation. CMAP analysis of CIPA revealed that the molecules in this drug have matching signatures with protein synthesis inhibitors, many of which are antiviral, and this could work through ESR1 axis (Haider et al., Reference Haider, Dholakia, Panwar, Garg, Gheware, Singh, Singhal, Burse, Kumari, Sharma, Ray, Medigeshi, Sharma, Prasher and Mukerji2021). These were validated in infection models of dengue in cell lines and also showed a repurposing possibility in SARS-CoV2 infection (Haider et al., Reference Haider, Anand, Enayathullah, Parekh, Ram, Kumari, Anmol, Shukla, Dholakia, Ray, Bhattacharyya, Sharma, Bokara, Prasher and Mukerji2022). Moreover, the complete formulation was more effective compared to single molecules in SARS-CoV2 inhibition. Transcriptome analysis and conjoint analysis with GSEA also revealed its involvement in the estrogen axis and possible usage in diseases where these need to be alleviated. Also docking studies revealed potential binding of a large number of molecules of CIPA to bind to ESR1 as also SARSCoV2. There is accumulating literature on the evidence of a link between the estrogen axis and viral inhibition. This study not only provides an evidence-based usage of Ayurveda formulation, but also provides new targets and molecules for therapeutic intervention. Noteworthy, both studies described above have integrated the principles and usage of the formulation from Ayurveda perspective in the study. Both drugs have been widely used and studied, but the above studies highlight the difference if traditional knowledge is integrated in these studies.

Therefore, the incorporation of principle and clinical usage in drug discovery programs that are sourced from herbs described in Ayurveda could provide interesting scaffold for discovering novel links, targets, and molecules. However, there is a need to develop, (1) experimental assays and models that can probe mechanisms in poly-pharmacological framework, (2) Patient-specific models that recapitulate the evolution of different stages of the diseases coupled with multi-omic (genomic, transcriptomic, proteomic, metabolomic) approaches for high-throughput and integrative screening of herbal formulations and discovery of targets, and (3) Poly-pharmacological based computational frameworks for docking, in silico identification of plausible therapeutic targets of the plant secondary metabolites as well as structure-guided discovery of molecules and targets. This should allow combinatorial synthesis of active constituents for the most effective outcomes. The route to such discoveries could be reduced if we are able to build an integrative platform that allows crosstalk between the two systems. Ayurgenomics-based approaches could have wide applications in defining the requirement for a platform for crosstalk and an interoperable framework in integrative medicine.

Ayurgenomics applications in integrative medicine settings

In the present times, patients are faced with multiple options for health and disease management, especially in India and Asian countries where traditional medicine is widely practiced. There is also an increased recognition for this in many populations of West. Besides, changing demographics that include a burgeoning younger population, an increase in mobility to nonnative environments, proportionately high aging population, and other cultural transitions, pose a major economic burden. Integrative medicine that combines principles and practices of traditional medicine, such as Ayurveda, could increase health and disease management options for individuals. This would also complement the current move toward digital, systemic, wholistic, and precise P4 (Predictive, Preventive, Personalized, Participatory) medicine of the contemporary times with an additional promotive component (Topol, Reference Topol2014, Reference Topol2019; Banerjee et al., Reference Banerjee, Debnath and Debnath2015; Mukerji and Prasher, Reference Mukerji and Prasher2015; Lemonnier et al., Reference Lemonnier, Zhou, Prasher, Mukerji, Chen, Brahmacharid, Nobleg, Auffraya and Sagnerh2017).

Ayurgenomics thus provides an operational framework and platform for integration of principles and practices of Ayurveda with modern medicine (Figure 2). Through an iterative approach, the utility of this framework in translation has been demonstrated. This platform can propel new knowledge and developments. It allows the discovery of molecular and multi-omic correlates of the common organizing principle tridosha that govern constitution types “Prakriti.” These molecular correlates (1) differentiate healthy individuals, (2) exhibit different baselines of intermediate patho-phenotypes that are associated with disease predisposition, (3) provide genetic links to disease predisposition and environmental responsiveness, and (4) can provide new leads for early actionable interventions and therapeutic targets. We observed recurrent themes across studies wherein intermediate patho-phenotypic states differ amongst Prakriti at baselines (Table 1). This is relevant in predicting individualized health trajectories and calibrating interventions. The contemporary relevance of this framework is evident from its potential in situations like the COVID-19 pandemic. For example, the pathways discovered to differ between Prakriti govern differential outcomes in COVID-19, more specifically, the ability to cope with hypoxia or the susceptibility to inflammatory consequences. This highlights a scope for risk stratification and prioritization of informative markers using Prakriti methods. Secondly, the intervention points are actionable with Ayurveda medicines as has been demonstrated by both Adathoda vasica and Cessamplous pareira. This evidence-based approach can increase the acceptability of Ayurveda recommendations and also highlights herbal recommendations from Ayurveda are testable in an Ayurgenomics framework. Such studies can provide new biomarker targets, drug repurposing possibilities, and enables bioprospecting for new molecules.

Thus, Ayurgenomics provides molecular subtitles to Ayurveda concepts and enables ontological links with modern medicine through a shared genomics dictionary. In addition, the development of technology-enabled platforms that would makes this information interoperable with modern medicine could provide exciting opportunities for integrative medicine especially in precision medicine settings (Singh et al., Reference Singh, Bhargava, Ganeshan, Kaur, Sethi, Sharma, Chauhan, Chauhan, Chauhan, Chauhan and Brahmachari2018; Mukerji and Sagner, Reference Mukerji and Sagner2019). For example, predominant “Vata” individuals have higher cell proliferation rates in baseline and are also more prone to DNA-damaging agents. Inherent differences in cell proliferation have been demonstrated to govern variability in lithium response in cell lines derived from bipolar disorder patients (Paul et al., Reference Paul, Iyer, Nadella, Nayak, Chellappa, Ambardar, Sud, Sukumaran, Purushottam, Jain and Viswanath2020). Also, analysis of predominant Prakriti in a cohort of diabetes has shown that Vata Prakriti individuals are not only more prone but also exhibit heightened response to DNA-damaging agents (Banerjee et al., Reference Banerjee, Biswas, Chattopadhyay, Arzoo and Chattopadhyay2021). Could Prakriti phenotyping provide an additional assistance in pharmacogenomic settings? Prakriti and cell proliferation rate assessment from healthy tissues might be additionally useful for in cancer management. Higher baseline of cell proliferation rates that have evolved to modulate response to DNA-damaging agents could confound calibration of dosage of DNA-damaging agents and promote recurrence. Genetic as well as exome analysis reveals differences in pharmacogenes as well as response to drugs amongst extreme constitution types (Joshi et al., Reference Joshi, Ghodke and Patwardhan2011; Bhalerao et al., Reference Bhalerao, Deshpande and Thatte2012; Prasher et al., Reference Prasher, Varma, Kumar, Khuntia, Pandey, Narang, Tiwari, Kutum, Guin, Kukreti, Dash and Mukerji2017). Thus, the integration of Ayurveda-based phenotyping could be of use in pharmacological settings. A third aspect is the scope of Ayurgenomics in translation. Differences in probing the molecular pathways involved in poly-herbal formulation through this framework could not only provide molecular links to its usage, but also allow development of novel drug discovery paradigms with a poly-pharmacological framework and inspire combinatorial synthesis of multiple molecules for network medicine.

Concluding remarks

“Integrative Medicine” formalized as another option for health management is urgently required. However, Ayurveda faces a disconnect with a modern audience. There is limited scientific evidence in contemporary language about the principles of systems’ medicine and there are also challenges in understanding the heterogeneity in treatment modalities. There is a need for developing and evolving a common language and evidence-based technology solutions for its integration into mainstream and interoperability with other medicinal systems. Moving forward, these gap areas need to be addressed. A technology-enabled ecosystem for evolving harmonized protocol for integrative medicine is required. These would involve the integration of digital devices and IOT-based technologies that not only allow automated capture of large-scale data but also enable the development of AI-based recommendation engines that could assist the Ayurveda clinicians in objective decision-making. Big data analytics and natural language processing-based methods could enable integration of ontologies from modern and Ayurveda systems in an interoperable framework. Integration of multi-omic technologies, including electronics and sensors with devices, chemical, genomics, molecular as well as digital markers for calibrating therapy, evolving standardized protocols, as well as monitoring treatment and outcomes and uniform data sharing platforms and architecture could all be enablers for realizing “Evidence-based Ayurveda” solutions and a participatory framework for managing health. Besides integrative medicine, this could also increase outreach, access, and surveillance, as well as affordable solutions.