Micronutrient deficiency (MNDs) or “Hidden hunger” is considered a global crisis affecting more than 2 billion people in the developing countries of South Asia, Africa, and Latin America (Bouis et al., 2013; Kumssa et al., 2015; Migliozzi et al., 2015; Wakeel et al., 2018). It was estimated of the major micronutrients that 60% of the world’s population is iron (Fe) deficient, over 30% are zinc (Zn) deficient, and 12.2% are protein deficient (Food and Agriculture Organization of the United Nation [FAO], 2014; Stein, 2014). The deficiencies of other essential micronutrient components of the diet, including calcium (Ca), are also prevalent and ultimately hamper the well-being of humanity (Kumssa et al., 2015; Beal et al., 2017; Narayanam et al., 2021). Therefore, pressing needs toward reducing MNDs through biofortification was considered a key component by the United Nations in their Millennium Development Goals (MDGs) program. Systematic breeding efforts for improving the nutrient status and the development of biofortified varieties offer great scope for mitigating MNDs judiciously. As a prerequisite, substantial variability should be available in the germplasm set in terms of grain micronutrient concentrations as well as knowledge regarding inheritance pattern and genetic relationship of these micronutrients in the targeted crop (Upadhyaya et al., 2012, 2016; Velu et al., 2014; Phuke et al., 2017).

Pulses are a major source of plant-based protein and other nutrients like phosphorus, vitamins, minerals, riboflavin, and essential amino acids. The genus Vigna savi is one of the most important genera among all the pulse crops, containing more than 200 domesticated and wild species (Pratap et al., 2015). Wild Vigna species possess a high potential for utilization as human food and fodder for animals (Popoola et al., 2015). Some of them are already utilized as human food, for example, V. marina (Burm.) Merr. is used in Australia (Tomooka et al., 2011), V. vexillata (L.) A. Rich. in Southern Asia (Karuniawan et al., 2006; Tripathi et al., 2021), V. racemosa (G. Don) Hutch & Dalzeil. in Nigeria (Folashade et al., 2017), and V. stipulacea (Lam.) Kuntz. and V. trilobata (L.) Verdc. in Asia (Siddhuraju et al., 1992; Bravo et al., 1999; Gore et al., 2019). However, the utilization of wild Vigna species as food and fodder is still very restricted because of unawareness around their potential. Increased scientific efforts to explore the hidden potential of wild Vigna species are recommended while planning future food strategies (Harouna et al., 2018). The nutritional composition plays a crucial role in food acceptability and food preferences as it is directly linked to consumers’ health and well-being. V. stipulacea, known locally as Minni payaru, is being utilized in the southern part of India, mainly in Tamil Nadu, for animal feeding, manure production, and in some traditional dishes like “Idli” and “Vada” (Tomooka et al., 2011; Gore et al., 2021). Unfortunately, V. stipulacea is an under-researched legume and no reports are available to date regarding the variability of minerals viz., Fe, Zn, Ca, etc., and the protein concentration in the grains of this species. Mineral and micronutrient concentrations in many major crops reported quantitative inheritance with low heritability, thus confirming the convoluted role of soil composition and other environmental factors toward mineral availability (Stangoulis et al., 2007; Shi et al., 2008). Therefore, a holistic approach for deciphering genotype × environment interaction (henceforth GEI) is necessary to identify stable micronutrient-rich genotypes. Additionally, breeders are also willing to detect target test locations for conducting future screening programs. This species of Vigna is generally cultivated in dry areas with a rainfed ecosystem where substantial variation is observed regarding soil micronutrient status. Therefore, GEI for grain nutrient concentration is expected to be significant and treated as a step toward developing stable genotypic performance across environments.

Many statistical tools are available to measure the confounding role of environment followed by characterizing and grouping genotypes and environments. Among them, the GGE biplot is gaining popularity over other biplot analyses. This methodology was proposed by Yan et al. (2000), which highlighted that both genotype (G) and genotype × environment (GE) are the two sources of variation that should be considered concurrently for evaluation of genotypes and test environments precisely (Yan and Tinker, 2005). The beauty of this method in comparison to other biplot analyses is that it excludes the major influence of the environment (E) and considers the main genotypic effect coupled with the GEI effect in a lucid way for the evaluation of genotypes and testing locations (Yan et al., 2007; Yan and Holland, 2010). In earlier studies for comprehending GEI in food legumes concerning micronutrient availability, several statistical tools were used such as Eberhart and Russell (1966)’s method in peanut (Upadhyaya et al., 2012); mungbean (Singh et al., 2013); lentil (Kumar et al., 2013, 2014; Darai et al., 2020); or by using AMMI biplot in cowpea (da Silva and Santos, 2017) and lentil (Singh et al., 2017). Recently, Gupta et al. (2020) have deployed GGE biplot analysis for deciphering GEI toward identifying stable genotype in urdbean. To the best of our knowledge, this is the first report of expanding this methodology for precise appraisal of V. stipulacea genotypes concerning grain micronutrients and antinutritional factors. Therefore, keeping these in the backdrop, the present study was deliberated toward assessment of genetic variability of a diverse set of V. stipulacea genotypes concerning nutrients (Fe, Zn, Ca, and protein) and antinutritional factors (phytate) followed by determination of the GEI interaction for appraisal of nutrient-rich stable genotypes through GGE biplot approach. The findings of this study will make it easier for global researchers to select nutrient-rich V. stipulacea germplasm for future breeding and genomic studies.

A total of 99 accessions of V. stipulacea were grown in the Kharif (monsoon season) of 2018–19 at two locations. The details of the accessions and their geographical origin are listed in Table 1. Hereafter, the two testing locations are referred to as Loc1 for ICAR-Indian Institute of Pulses Research (IIPR), Kanpur located at 26°27′N latitude, 80°14′E longitude, 152.4 m above mean sea level (AMSL), and Loc2 for ICAR-National Bureau of Plant Genetic Resources (NBPGR), New Delhi, located at a latitude of 28°40′N and longitude of 77°12′E and an altitude of 228 m AMSL.

At both locations, accessions were sown under natural field conditions. The recommended package of practices for growing Kharif Vigna species was followed. For meeting the balanced nutrient demand of the crop, 100 kg ha^–1 Di-ammonium Phosphate (DAP) was applied as a basal dose before sowing. One pre-sowing irrigation was given to ensure proper germination in the field, while light irrigation was also given at the flowering stage. Sufficient moisture was available during the entire cropping season at both locations. The mechanical scarification of seeds was done to ensure maximum germination (Pratap et al., 2015; Singh N. et al., 2020). At ICAR-IIPR, accessions were grown in customized pots in a Vigna wide hybridization garden. Each pot measured 1 m in diameter and was about 1 m high. Further, each pot had provision of drainage at the bottom and sides so that water stagnation did not occur during excess precipitation. In each accession, 20 seeds were sown in individual pots around the periphery maintaining an equal distance of 5–7 cm between two plants. At ICAR-NBPGR, every accession was grown in two rows, each with 4-m length, with a row-to-row spacing of 60 cm. No insecticide was sprayed and manual weeding was done at both locations 25–30 days after sowing.

Micronutrient deficiency (MND) or “hidden hunger” is considered a global crisis (Bouis et al., 2013; Wakeel et al., 2018). The basic reason for MNDs in developing countries is the consumption of diets based only on cereals and one or two staples, with less diversification in their food platter due to acute poverty. Pulses are an integral component of the daily diet of vegans as well as the peoples of developing countries. Besides being a chief source of protein, pulses are also a good source of dietary fiber, low molecular weight carbohydrates, essential amino acids, polyunsaturated fatty acids, minerals, and vitamins (Thompson, 2019; Venkidasamy et al., 2019). Hence, pulses are a good candidate for biofortification, and progress of pulse biofortification through conventional breeding programs is also at a good pace (Nair et al., 2013; Petry et al., 2015; Upadhyaya et al., 2016; Kumar et al., 2018; Tan et al., 2018; Jha and Warkentin, 2020; Wu et al., 2020). However, most of the earlier biofortification program targeted major pulses, and the potential genetic diversity of underutilized pulses and crop wild relatives (CWR), which are a rich source of essential elements (Difo et al., 2015), were not harnessed assiduously to unravel the novel alleles governing micronutrient accumulation and their further transfer into the cultivated background. Although V. stipulacea, like other Vigna species, is nutrient-dense, it received less attention concerning micronutrient composition in the prevention of MNDs, mainly owing to its minor crop status.

This experiment was conducted with the aim of evaluating a large number of genotypes of V. stipulacea to understand the presence of variability regarding grain nutrients (Fe, Zn, Ca, and protein) and the antinutritional factor (phytate), followed by contemplating the complex role of environments on the inheritance of these quantitative traits for identifying stable genotypes. In the present study, the mean performance of the accessions revealed the presence of ample genetic variability in most of the studied traits. As a result, the potential to obtain desirable recombinants was emphasized by using promising genotypes as parents in the biofortification efforts. The variability observed in grain nutrients and phytate concentration in the present study mostly corroborated with the variability observed in other Vigna species especially for well-studied traits like Fe and Zn (Dwivedi et al., 2012; Nair et al., 2013; Wu et al., 2020). In the present study, the seed Fe concentration of V. stipulacea genotypes was 34.3–128.8 mg kg^–1 over the locations with an average value of 66.7. In earlier studies, with other Vigna species like mungbean, an average value of 30.0 and 40.0 mg kg^–1 of seed Fe have been reported (Wu et al., 2020), which confirmed that V. stipulacea is the potential source for mitigating Fe deficiency of rural folks especially adolescent children and female with acute problems of anemia. Similarly, the present finding regarding grain Fe concentration was corroborated with the wider variation reported in the case of other small-seeded legumes like lentil (Thavarajah et al., 2017; Kumar et al., 2018; Shrestha et al., 2018); common bean (de Araújo et al., 2003; Blair et al., 2009) and cowpea (Boukar et al., 2011; Santos and Boiteux, 2015). On the contrary, a narrow range of variation (20.6–71.0 mg kg^–1) was reported regarding grain Zn concentration compared to grain Fe in our study. In earlier reports with mungbean, a range of 25.0–30.0 mg kg^–1 of grain Zn was reported (Wu et al., 2020), which further proved that considerably higher Zn concentration was present in V. stipulacea genotypes. However, a relatively higher concentration of Zn concentration of 5–134 mg kg^–1 was reported in urdbean (Gupta et al., 2020). Earlier works in common bean reported a smaller range of seed Zn concentration (Gelin et al., 2007; Blair et al., 2009). The present study revealed that V. stipulacea accessions have a greater variation concerning grain Fe concentration than the grain Zn concentration. The same trend was observed in both small and large-seeded grain legumes, perhaps due to changes in embryo size and the proportion of seed coat to cotyledonary tissues (Ariza-Nieto et al., 2007; Cvitanich et al., 2010). In the present study, the differences in Ca concentration also showed very promising results, with an average value ranging between 159.1 and 161.6 mg kg^–1 across the locations. Differences in the concentration of minerals like Ca are less studied in Vigna species, which is highly influenced by the soil pH (Wu et al., 2020). The variation regarding grain protein concentration in the studied V. stipulacea accessions was narrow and it was well corroborated with earlier studies in mungbean (Akaerue and Onwuka, 2010; Nair et al., 2013); cowpea (Gupta et al., 2010; Ravelombola et al., 2016) and urdbean (Shaheen et al., 2012). Phytate or phytic acid is considered as one of the antinutritional factors and inhibits the bioavailability of minerals and proteins (Kumar et al., 2010; Nissar et al., 2017). Paradoxically, phytate also has nutraceutical properties and is considered a useful suppressor of cardiovascular diseases and cancers in human beings (Selle et al., 2012; Thavarajah et al., 2014). Besides, phytate is an essential component for plant metabolism (Oomah et al., 2011). Therefore, biofortification efforts for reducing phytate concentration should be considered meticulously so that an optimum balance between seed phytate concentration and bioavailability of other minerals can be judiciously accomplished. An earlier study by Chitra et al. (1995) reported that phytate concentration was the highest in urdbean followed by mungbean with a concentration of 13.7 mg g^–1 and 12.0 mg g^–1, respectively. The phytate concentration in V. stipulacea accessions was higher (15.4 mg g^–1) than both urdbean and mungbean, which requires urgency to initiate a research program for the reduction of the phytate concentration of this Vigna species to an optimum level so that the bioavailability of the minerals can be improved. Although, the phytate concentration of V. stipulacea was comparatively lesser than legumes such as soybean (mg g^–1), as mentioned by Chitra et al. (1995). Furthermore, if the heritability for the characteristic of interest is high, the variability existing in the population is helpful and can be harnessed. In the present study, high heritability coupled with high GA for most of the studied traits justified that any selection will be rewarding for improving these traits in V. stipulacea.

The perplexing role of the environment and the networking between GEI is a serious noise toward the phenotypic expression of any quantitative traits due to a reduction in heritability (Phuke et al., 2017). Therefore, GEI should be judged properly by growing the genotypes in different locations to confirm the stable phenotypic expression regarding complex traits of interest. GEI is the product of different kinds of genetic association across the environments (Ye et al., 2006). In the present study, the ANOVA for all the studied traits represented highly significant genotypic variance across the environments based on pooled analysis. Similarly, the interplay between genotype and location interactions (σ²gl) was also highly significant for individual context, suggesting the predominant role of genotype, environment, and their interaction (GEI) toward the inheritance of these traits. This finding justified the partitioning of GEI interaction. Grain micronutrients concentration is highly variable on environmental factors, especially soil properties of the location (Bashir et al., 2014; Kumar et al., 2018), which hinders the progress of genetic analysis of these traits. It was observed that the genotype × location (σ²gl) was the highest for phytate concentration due to higher genotypic variance for this trait in V. stipulacea genotypes. Moreover, high σ²gl was also observed in grain Fe concentration which was corroborated with the findings of Gupta et al. (2020) in urdbean; mungbean (Wu et al., 2020), and cowpea (Boukar et al., 2011; Santos and Boiteux, 2015).

In a crop biofortification program with the objective of improving the nutritional quality of food crops, determining the environmental stability concerning grain micronutrients is imperative (Welch and Graham, 2004; Upadhyaya et al., 2012; Bouis and Saltzman, 2017). GEI impedes genotypic selection based on phenotypic expression and ultimately reduces the genetic gain under selection (Comstock and Moll, 1963). Despite these challenges of GEI in crop biofortification program, breeders were enabled to develop nutrient dense biofortified varieties in many food legumes through exploiting various methodologies for stable genotype delineation (Boukar et al., 2011; Upadhyaya et al., 2012; Santos and Boiteux, 2015; Kumar et al., 2018). Recently among the various statistical tools for determining GEI, GGE biplot is gaining attention to analyze multi-locational data for deciphering complex GEI in a graphical mode (Yan, 2001; Yan and Tinker, 2006; Yan et al., 2007). In GGE biplot analysis, the complex GEI are presented in the form of various PCs for graphical presentation of the data against each PC (Yan and Tinker, 2006). Earlier findings stated that the first two PCs should explain more than 60.0% of the variability present within the data set for determining the competence of the methodology (Yan et al., 2010). In the present study, both the PCs were able to explain the total variations for all the studied traits. For detection of the ideal genotype, both the mean performance and consistency over the locations should be considered (Yan et al., 2007). The presence of high GEI for all the studied traits influenced the rank of the genotypes across the locations, suggesting the presence of COI as reported earlier (Das et al., 2019; Singh B. et al., 2020). The presence of COI recommended breeding for specific adaptation.

In the present dataset, it was detected that the highest grain Fe (G-2), Zn (G-76), Ca (G-60), and protein (G-64) containing genotypes were not stable except in the case of grain Ca concentration where G-60 was the most stable as well as a high performer thus considered as the ‘ideal’ genotype. For grain Fe, Zn, and protein concentration, the ‘ideal’ genotype were G-75, G-76, and G-27, respectively. For phytate concentration, the low performing genotypes (G-99, G-95) exhibited stable performance and were acknowledged as ‘ideal’ (G-95) and ‘desirable’ (G-99). Genotypes which positioned close to the ‘ideal’ genotype are considered ‘desirable’ as the distance between two genotypes resembled the Euclidian distance between them (Yan and Tinker, 2006). Among all the tested traits except for grain Ca, COI was observed for all the traits. Therefore, it can be said that within the same data set, both COI and non-COI were detected which was in affirmation with earlier findings (Dehghani et al., 2006; Sabaghnia et al., 2008; Das et al., 2019, 2020; Singh B. et al., 2020). The identified genotypes (G-75, G-76, G-60, G-27, G-95) for all the five studied traits would be precious genetic stocks for utilization as parents in a V. stipulacea biofortification program.

Besides identification of stable genotypes, delineation of the best testing location is also the choice of breeders. In GGE biplot, an ‘ideal’ test environment should be selected based on “discriminating power” to delineate genotypes, being ‘representative’ as well as having a high “desirability index” (Xu et al., 2014). In the present study, Loc1 was considered as the ‘ideal’ testing location for most of the traits except grain protein, where Loc2 was detected as the most ‘ideal.’ Ideal testing location delimitation would facilitate plant breeders for conducting their trials meticulously for precise genotype selection. Previous studies applied the same principle of GGE for genotype and testing location appraisal in different legumes concerning grain micronutrient concentration (Janila et al., 2015; Phuke et al., 2017; Darai et al., 2020; Gupta et al., 2020; Misra et al., 2020).

The association between different grain micronutrients stated that grain Fe and Zn exhibited a significant positive association that followed the same trend in individual locations. Similar relationships between these micronutrients have been reported in earlier studies in urdbean (Gupta et al., 2020); peanut (Janila et al., 2015); and chickpea (Misra et al., 2020). It can be speculated that the positive association of grain Fe and Zn might be due to overlapping QTLs as reported earlier in common bean (Blair et al., 2009, 2010), suggesting simultaneous selection for both the traits would be effective. Similarly, grain protein exhibited a significant positive association with Fe, which was corroborated with the earlier finding in wheat (Cakmak et al., 2010) and pearl millet (Pujar et al., 2020). On the contrary, a significant negative correlation between grain Fe and protein was detected in cowpea (Gerrano et al., 2019). The present study detected a significant negative association between grain protein and phytate. However, earlier studies reported a positive association between these two traits (Chitra et al., 1995; Dai et al., 2007), though the amplitude of correlation was relatively modest in the case of pulses. Therefore, selection against phytate concentration would be rewarding toward the selection of high protein genotypes in V. stipulacea.

The adequate variability was observed regarding grain nutrients and phytate concentration among the V. stipulacea germplasm. The presence of significant GEI for most of the traits indicated the convoluted role of environments on the phenotypic expression of these traits. A high magnitude of GEI was observed for grain phytate and Fe concentration. GGE biplot analysis revealed the incongruous performance of the genotypes for most of the traits except grain Ca concentration, suggesting precise phenotyping of the traits in a testing location with specific adaptation. GGE biplot methodology deployed in the present study decisively delineated stable genotypes concerning grain nutrients and phytate as well as testing locations for culling out ideal genotypes. The positive and significant association between grain Fe with Zn and protein admitted the possibilities of simultaneous selection of all these three characters. Commonly, most of the Vigna species are native to Asia and the horn of Africa and are prevalent in the food platter of Asian and African communities where MNDs are very acute. Like other Vigna species, V. stipulacea is also easy to cook quickly; therefore, it requires meager energy demands during food preparation. Wider variation of essential nutrients in V. stipulacea followed by utilization of stable genotypes in a biofortification program holds immense promise to combat the malnutrition of Sub-Saharan Africa and other underdeveloped countries of Asia and Africa in a judicious way.

The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding authors.

PG and KT conceptualized the experiments. PG performed the experiments and prepared the manuscript. AD, KT, and SB analyzed the data and prepared the manuscript. AP contributed to the facilitation of the field trials and data recording. RB provided lab facility for biochemical analysis. VG and HD supervised the research trial. RN contributed to the acquisition of fund, revision and finalization of the manuscript. All authors have read and agreed to the final version of the manuscript.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.