Introduction

The genus Vigna is an agro-sociological important plant taxon. It comprises more than 100 leguminous plant species distributed in wide and diverse areas of Africa, America, Australia and Asia (Lewis et al. 2004). Nine Vigna species are considered as domesticated crops, including cowpea (V. unguiculata (L.) Walps.), zombi pea (V. vexillata (L.) A. Rich), Bambara groundnut (V. subterranean (L.) Verdc.), mungbean (V. radiata (L.) Wilczek), azuki bean (V. angularis (Ohwi) Ohwi and Ohashi), rice bean (V. umbellata (Thunb.) Ohwi and Ohashi), black gram (V. mungo (L.) Hepper), moth bean (V. aconitifolia (Jaqc.) Maréchal), and créole bean (V. reflexo-pilosa Hayata) (Tomooka et al. 2002, 2011). These species are grown mainly for dry seeds by small farmers in several cropping systems of tropical and sub-tropical regions. In addition, some other Vigna species are cultivated and/or harvested for seeds and pods as food, and for leaves and stems as feedsby local people in Africa, Asia, Australia and America (Tomooka et al. 2011).

Zombi pea is an underutilized herbaceous legume belonging to the subgenus Plectrotropis of the genus Vigna. Wild forms of zombi pea occur widely in Africa, America and Australia and Asia, while the cultivated form occurs in a few localities of Africa and Asia (Tomooka et al. 2011). Wild zompi pea often develops storage roots which are edible and consumed by people in Africa, India, Australia and Southeast Asia (Duke 1981; Lawn and Cottrell 1988; Sasikumar and Sardana 1988; Karuniawan et al. 2006), although seeds are sometimes consumed. Edible tubers of V. vexillata contain a protein content of about 15% which is roughly three times higher than that of potato and yam and six times than that of cassava (Chandel et al. 1972). Moreover, yield of V. vexillata tuberous roots has been reported to be as high as 7000 kg/hectare which is comparable to sweet potato (Bhattacharyya et al. 1984). There are two types of cultivated zombi pea. One form is grown for both edible seeds and tuberous roots, while the other is grown solely for tuberous roots. The former is reported in Sudan (Ferguson 1954), while the latter is reported in Bali and Timor islands of Indonesia (Karuniawan et al. 2006) and India (Bhattacharyya et al. 1984; Asati and Yadav 2004).

Zombi pea is a very heterogeneous legume species. Eight taxonomic varieties including macrosperma, vexillata, angustifolia, dolichomena, yunnanensis, plurifora, lobatifloria and ovata have been described/recognized (Maréchal et al. 1978; Pasquet 2001; Pienaar and Kok 1991; Maxted et al. 2004). Variety macrosperma is the cultivated type, while the others are wild type. Vanderborght (1989) examined germination habits of zombi pea germplasm and found that accessions from America were epigeal whereas those from Africa (except Nigeria) and Australia were hypogeal. Due to its wide distribution, zombi pea has been reported to adapt well to environmental stress conditions such as infertile soil (Karuniawan et al. 2006), alkaline soil (Lawn and Watkinson 2002), acid soil (Vanderborght 1989), saline soil (Roecklein and Leung 1987), drought (Roecklein and Leung 1987), and waterlogging (Miller and Williams 1980). Zombi pea is also reported resistant to several insect pests including Callosobruchus maculatus, Zabrotes subfasciatus, Maruca testulalis and Clavigralla tomentosicollis (Birch et al. 1986; Jackai and Oghiakhe 1989) which are major pests of legume crops. In addition, this legume is resistant to viral diseases such as cowpea mottle carmovirus (CPMoV), cowpea aphid borne mosaic virus (CAbMV) and cowpea yellow mosaic virus (CYMV) (Gomathinayagam et al. 1998; Thottappillyet al. 1994). Thus, zombi pea is potentially a gene source in breeding for resistance to biotic and abiotic stresses, as well as a new future crop.

So far, there are a limited number of reports on genetic diversity of zombi pea. The previous genetic diversity analyses in zombi pea were conducted using molecular markers, viz. seed storage proteins (Piergiovanni 1998), isozymes (Garba and Pasquet 1998; Jaaska 2001) and RAPD markers (Spinosa et al. 1998). The results from these studies suggested that Africa is the center of diversity of zombi pea. Nonetheless, these studies have provided little information on the extent of genetic diversity and genetic structure of zombi pea because these studies used a small number (<150) of accessions and the molecular markers used were low informative.

Simple sequence repeat (SSR) or microsatellite is the marker of choice for molecular genetics study in crops because of its advantages as being co-dominant, multi-allelic, reliable, PCR-based, and easy to score. SSR markers from one species can also be used in other related species. In Vigna species, SSR markers developed from azuki bean, cowpea, and mungbean were shown to have moderate to high rates of amplifcation in other Vigna species (Chaitieng et al. 2006; Gupta and Gopalakrishna 2010; Somta et al. 2009; Tangphatsornruang et al. 2009).

In this study, we report a genetic diversity analysis of 422 zombi pea accessions from various origins using SSR markers from azuki bean, cowpea and mungbean. The objectives of this study were to assess genetic diversity and genetic structure of zombi pea. The results will be useful for plant geneticists/breeders to better understand the diversity and domestication of this legume species.

Materials and methods

V. vexillata germplasm and DNA extraction

A total of 422 (408 wild and 14 cultivated) accessions of V. vexillata originating from various origins, including Africa (201 accessions), America (126 accessions), Australia (85 accessions), Asia (5 accessions) and unknown (5 accessions) were used in this study (Table 1 and supplementary Table S1). They were sown in an experimental field of Kasetsart University, Kamphaeng Saen campus, Nakhon Pathom, Thailand. In addition, two seeds of each accession were germinated on a germination tray using peat moss as the culture media, and their germination habit was recorded as hypogeal or epigeal.

Table 1 Source and number of V. vexillata accessions used in this study with number of alleles, gene diversity (H E) and observed heterozygosity (H O) and average allelic richness (Ar) calculated from allelic data of 20 SSR markers
Full size table

Young leaves from one plant of each accession were collected and extracted for total genomic DNA. The DNA extraction was carried out using a CTAB method described by Lodhi et al. (1994). Quality and quantity of the DNA were measured on 0.8% agarose gel electrophoresis by comparing with a known concentration of lambda DNA.

SSR marker analysis

One thousand and twenty-four SSR markers from azuki bean (Wang et al. 2004; Chankaew et al. 2014), mungbean (Tangphasornruang et al. 2009; Somta et al. 2009), and cowpea (Li et al. 2001; Kongjaimun et al. 2012) were screened for amplification and polymorphism in six accessions (CIAT4295, CIAT4525, CIAT4338, AusTRCF320518, JP235863, and TVNu1583) of V. vexillata. These accessions were randomly selected from different geographical regions. A PCR mixture was prepared following Somta et al. (2008) with minor modification. In brief, the PCR mixture of 10 µl contained 10 ng genomic DNA, 5 pmol of each forward and reverse primers, 1 × Taq buffer, 2 mM dNTPs, 1.5 mM MgCl2, and 1 U Taq DNA polymerase (Thermo Scientific). The PCR amplification was carried out in a GeneAmp PCR System 9700 Thermal Cycler (Applied Biosystems, Foster City, Calif., USA) or PTC-200 Thermal Cycler (MJ Research, Waltham, MA, USA) using cycling profile as followed: 94 °C for 2 min, followed by 35 cycles of 94 °C for 30 s, 55 °C for 30 s, 72 °C for 1 min and the final extension at 72 °C for 10 min. Amplified products were separated on 5% denaturing polyacrylamide gel (w/v; 19:1 acrylamide-bisacrylamide) with 7 mol/L urea and 1× Tris-borate-EDTA buffer. Electrophoresis was run at 70 W constant power for 1.5–2 h (depending on allele size) using a pBR322 DNA/BsuRI (HaeIII) (Thermo Scientific) as the size standard. The amplified products were visualized by silver staining. Twenty markers (supplementary Table S2) showing polymorphic DNA bands were used to analyze DNA of all the 422 V. vexillata accessions.

Genetic data analysis

Allele data were used to calculate number of alleles per locus, observed heterozygosity, gene diversity (expected heterozygosity), and average allelic richness using FSTAT 2.9.3.2 software. The data were also used to calculate the polymorphism information content (PIC) (Anderson et al. 1993).

Genetic distance (D A) (Nei et al.1983) for all pairs of accessions and for all pairs of geographical groups was calculated using POPULATIONS 1.2.32 software. The D A for all pairs of accessions was then subjected to a neighbor-joining analysis using MEGA 6 software (Tamura et al. 2013) and to a principle coordinate analysis using package Factoextra in R-program 3.0.1 software (R Development Core Team 2013) to reveal genetic relationships among accessions. Also, D A for all pairs of geographical groups comprising more than 7 accessions was calculated using POPULATIONS 1.2.32 software.

The genetic structure of the 422 V. vexillata accessions was determined by STRUCTURE 2.3.4 software that employs a model-based Bayesian Markov Chain Monte Carlo (MCMC) algorithm. Initially, 20 simulation runs were carried out using a number of assumed populations (K) of 1 to 10 with a burn-in period of 10,000 and 50,000 replicates of MCMC. The true K was determined by the ad-hoc ΔK method (Evanno et al. 2005). Subsequently, a run with optimum K, burn-in period of 100,000 and 500,000 replicates of the MCMC algorithm was performed to assign each individual to a cluster.

Results

Variation in germinations of V. vexillata germplasm

Germination was epigeal in 186 and hypogeal in 236 V. vexillata accessions (supplementary Table S1). Wild accessions showed both germination types, while cultivated accessions showed only hypogeal germination. Wild accessions from Africa were very heterogeneous for the germination habit. In contrast, the accessions from America and Australia were each highly homogeneous for the germination habit. All except one of the American accessions were of epigeal type, while all except two of the Australian accessions were of hypogeal type. Only two Asian accessions were available in this study and showed one germination type each.

SSR variation in V. vexillata

Of the 1,024 SSR markers used to screen for polymorphism in 6 accessions of zombi pea, 595 were able to amplify these DNA, 217 showed polymorphism. Twenty SSR markers (supplementary Table S2) were chosen to analyze the 422 accessions of V. vexillata and detected 273 alleles in total (Tables 1, 2). All the alleles, except one, were found in the wild germplasm. The number of alleles detected per marker was between 6 (CEDG091, CEDG174) and 20 (CEDG107), with an average of 13.65 (Table 2). Among the wild accessions with known origins, those from Africa possessed the highest number of unique alleles (alleles found in one origin but not in the others; 73), followed by the germplasm from Australia (7) and America (6).

Table 2 Linkage groups (LG), number of detected alleles per locus, allele size range and polymorphic information content (PIC) of the 20 polymorphic SSR markers used in diversity analysis
Full size table

The polymorphism information content (PIC) values of the markers varied from 0.132 (CEDG174) to 0.915 (CEDG214) with an average of 0.704 (Table 2). Among the wild accessions, average PIC value was highest in those from Africa but not different from those from Australia (Table 2), while in the cultivated germplasm, average PIC value was 0.601 (Table 2).

Observed heterozygosity (H O) over all accessions was 0.148. H O values between wild and cultivated accessions were similar, being 0.169 and 0.143, respectively. In the wild germplasm, H O was highest in African accessions, followed by Australian and American accessions, respectively (Table 1).

Gene diversity (H E) among all accessions was 0.72. H E values in the wild and cultivated accessions were comparable. Among the wild germplasm, H E was highest in the accession from Africa (0.75) and lowest from America (0.44) (Table 1). Wild germplasm from Africa showed highest average allelic richness (Ar), followed by Australia and America (Table 1). Of the African wild germplasm, East African germplasm had highest H E and Ar, although not much different from the other African regions.

Neighbor-joining analysis

A phylogenetic tree reconstructed from neighbor-joining (NJ) analysis using D A showed that, in general, accessions from the same geographical region were clustered together showing three major clusters (I, II and III) (Fig. 1). Cluster I was the largest cluster comprising mainly African accessions, all Asian accessions, 15 Australian accessions and 8 American accessions. All cultivated accessions were included in this cluster. Cluster II comprised principally accessions from America together with 5, 1, 1, and 2 accessions from East Africa, Central Africa, Australia, and unknown origin, respectively. Cluster III comprised most accessions from Australia and six accessions from East Africa.

Fig. 1
figure 1

A Neighbor-joining tree constructed from Nei’s genetic distance (D A) showing genetic relationship among the 422 accessions of Vigna vexillata. D A was computed from allelic data of 20 simple sequence repeat markers

Full size image

The 14 cultivated accessions were separated into three sub-clusters within the cluster I (Fig. 1). Sub-cluster I comprised only accessions from Bali island of Indonesia, sub-cluster II included only accessions from Sudan, and sub-cluster III contained all accessions from America and one accession each from Sudan, Botswana and Central African Republic. Interestingly, the accessions from Bali were clustered with wild accessions from various origins.

In the wild African germplasm which is highly heterogeneous for germination habits, germplasms with the same germination type generally clustered together (supplementary Fig. S1).

Principal coordinate analysis

PCoA analysis based on D A revealed that the first three PCs together accounted for 37.3% of the total variation, with PC1, PC2 and PC3 explained 18.2, 11.2 and 7.9, respectively. A scatter plot of the zombi pea germplasm based on PC1 and PC2 showed that accessions from Africa, America and Australia were largely differentiated, except for some accessions from America and Australia which were associated with those from Africa. Accessions from America and Australia were clearly distinct, while accessions from Asia mingled with those from Africa.

Wild accessions from East, West, Central and South Africa were not clearly differentiated. Among them, wild accessions from East Africa showed the widest distribution, whereas wild accessions from Central and South Africa showed a narrow distribution and were not differentiated. Wild accessions from Australia were clustered into two subgroups; one comprised accessions from East, North and West Australia, and the other included only accessions from East and West Australia. The latter group showed close genetic relationships with accessions from Africa. Wild accessions from Asia showed clear genetic differentiation, although there were only two accessions from this region, they showed close genetic relationships with accessions from Africa. All 14 cultivated accessions in this study showed a wide distribution and a clear distinction from each other. Even the wild African accessions showing the same germination type by NJ analysis were not grouped together by PCoA analysis.

Population structure analysis

Bayesian clustering of the zombi pea germplasm was performed using STRUCTURE software. Based on Evanno’s ad hoc ΔK method (Evanno et al. 2005), there were three subpopulations among the 422 zombi pea accessions (Fig. 3). Subpopulation I was the largest subpopulation with 231 accessions. This subpopulation comprised almost all accessions from wild Africa (193 accessions), some wild accessions from America and Australia, all accessions from Asia and all cultivated accessions. Subpopulation II comprised 122 accessions; all except four (three from Africa and one from Australia) accessions are from America. Subpopulation III comprised 70 accessions, all from Australia. In most cases, the results of germplasm clustering using NJ, PCoA and STRUCTURE analyses were congruent.

Genetic differentiation among geographical groups

D A analysis among groups of the wild zombie pea germplasm (Table 3) revealed that their genetic distance was low (0.151 for South Africa vs. East Africa) to relatively high (0.415 for wild West Africa vs. wild Australia). Among the African germplasm, East African germplasm showed the closest genetic relationship with South African germplasm, while South African germplasm showed the widest genetic relationship with Central African germplasm. Wild American germplasm showed the closest genetic relationship with East and West African germplasm, suggesting that the American germplasm was genetically originated from East or West Africa or both regions. Wild Australian germplasm showed the closest genetic relationship with East African germplasm, and is likely to be genetically derived from East Africa.The cultivated germplasm from Africa showed closest genetic relationship with wild germplasm from East Africa (D A = 0.351) and highest differentiation with the wild germplasm from Australia.

Table 3 Pairwise Nei’s genetic distance (D A) among the six geographical groups of V. vexillata based on 20 SSR markers
Full size table

Discussion

Association between germination habits and geographical origins in zombi pea

Both epigeal and hypogeal germinations have been reported in zombi pea (Vanderborght 1989) and these germination habits are associated with geographical origins (Spinosa et al. (1998). Similar results were found in this study where zombi pea germplasm from Africa showed both germination types, while most Australian accessions had epigeal germination and American accessions had hypogeal germination (supplementary Table S1). This implies that the zombi pea germplasm from Africa possesses higher genetic diversity than the germplasm from America and Australia, and the genetic structure of V. vexillata from America and Australia is associated with germination habits.

Center of origin, diversity and domestication of V. vexillata

Maréchal et al. (1978) studied the morphological diversity of V. vexillata and proposed that this legume originated in Africa. In this study, SSR allelic data clearly revealed that wild V. vexillata germplasm from Africa had the greatest gene diversity (Table 1). Our results support the proposal of Maréchal et al. (1978). V. vexillata accessions from East Africa showed higher gene diversity and broader distribution on the PCoA scatter plot than the other germplasm of African origins (Fig. 2), suggesting that East Africa is more likely the center of origin and center of diversity of V. vexillata.

Fig. 2
figure 2

A scatter-plot of PC1 and PC2 from principal coordinate analysis based on Nei’s genetic distance (D A) showing genetic relationship among the 422 accessions of Vigna vexillata. D A was calculated from allelic data of 20 simple sequence repeat markers

Full size image

There are two types of cultivated V. vexillata; African and Asian (Bali and East Timor in Indonesia) types. The African type shows erect and determinate growth habits, photoperiod-insensitive, early flowering, and hairy pods, while the Bali type shows viny and indeterminate growth habits, photoperiod-sensitive, late flowering, and glabrous pods. As compared to the African type, the Asian type has larger seed size and longer seed development duration. Ferguson (1954) noted that V. vexillata is a unique crop of Sudan where it is called “Babun” and grown only on a small scale in one locality for edible root and seeds. Karuniawan et al. (2006) reported that V. vexillata is grown in some localities of Bali and Timor Islands of Indonesia for edible root. Damayanti et al. (2010) studied variation of agronomic and morphological traits in some V. vexillata germplasm, including wild form and African and Bali cultivated forms, and found that the Bali form was distinct from the other forms, although small number of accessions were studied. Recently, Takahashi et al. (2016) determined phylogenetic relationship of many Vigna accessions including 9 accessions of V. vexillata using DNA sequences of nuclear DNA-ITS and chloroplast atpB-rbcL spacer regions and found that the Bali cultivated accessionsare genetically different from the other accessions. These suggested that Bali cultivated V. vexillata has been domesticated independently. In this study, cultivated V. vexillata accessions from Africa, America and Asia (Bali) were used and similar results were obtained. Neighbor-joining tree (Fig. 2) and PCoA scatter plot (Fig. 3) separated cultivated accessions into three different groups, although they were all clustered with wild African accessions. The Bali accessions were in a clear separated group. Several accessions from Sudan were grouped together, while one accession from this country was grouped with all accessions from America and accessions from Botswana and Central African Republic. Based on the information gathered by Damayanti et al. (2010) the group of the Sudanese cultivated accessions was traced back to the accession named ‘Babun’. Nonetheless, these results suggested that cultivated V. vexillata were domesticated twice and independently; once in Africa and once in Asia. The results supported the proposal of Garba and Pasquet (1998) based on herbarium specimens, morphological and isozyme variation that domestication of V. vexillata occurred independently in Africa and Asia. The clear grouping of cultivated accessions from Sudan (Fig. 1) supported the note of Ferguson (1954) that V. vexillata was first cultivated in Sudan. These together indicated that Sudan is one of the places of V. vexillata domestication in Africa and that the American cultivated accessions originated from the African cultivated accession, possibly from Sudan. This supported the report of Damayanti (2010) who traced the origins of American cultivated V. vexillata (also used in our study) and found that they were originally from Sudan.

Fig. 3
figure 3

Genetic structure of the 422 accessions of Vigna vexillata determined by STRUCTURE analysis based on simple sequence repeat allelic data at 20 loci. Each bar represents one accession

Full size image

In the case of the cultivated Bali accessions, they were genetically close to wild accessions from India (AusTRCF66514), Papua New Guinea (CIAT4944) and the Philippines (TVNu1632) (Fig. 1) which all have hypogeal germination. We found that hybridizing the accession AusTRCF66514 as the female parent by Bali cultivated accession (JP235863) as the male parent, resulted in partial self-fertile F1 plants, despite a low number of hand-pollinated flowers (Dachapak and Somta, unpublished data). This suggested that the Bali cultivated accessions constitute the primary or secondary gene pool of V. vexillata and they are domesticated from wild V. vexillata of India. This is possibly due to the fact that (1) V. vexillata is cultivated in India (Bhattacharyya et al. 1984; Asati and Yadav 2004), (2) plant material similar to Bali cultivated V. vexillata exists in Sri Lanka (Damayanti 2000; personal communication with Dr. Rémy Pasquet by Dr. Robert Lawn), and (3) there was a past cultural and social (Hinduism) relationship between South Asia and Indonesia, especially Bali. Additional diversity study using more wild and cultivated germplasm of V. vexillata from Asia in particular from India and Indonesia is necessary to elucidate the domestication site of Bali cultivated V. vexillata.

Genetic structure of zombi pea

As many as 8 taxonomic varieties have been described for V. vexillata due to the high morphological variation of this species (Maxted et al. 2004). Accessions characterized as variety angustifloria, macrosperma, vexillata, and ovata were included in this study. NJ tree and PCoA plot showed that accessions of the same varieties were not always clustered together (Figs. 2, 3; see also supplementary Table S1). In stead, the genetic structure of this species is geographical dependent. Similar finding was reported earlier by Spinosa et al. (1998) who used RAPD markers to analyze V. vexillata accessions of var. angustifloria, macrosperma and vexillata and found no relationship between botanical varieties and geographical origins.

Damayanti et al. (2010) showed that the Bali cultivated V. vexillata is not cross-compatible with the wild V. vexillata or African cultivated V. vexillata. Based on this result the authors concluded that the Bali cultivated accessions are not in the primary gene pool of V. vexillata. As mentioned above, hybridization between a Bali cultivated accession (JP235863) and an Indian wild accession (AusTRCF66514) resulted in partially self-fertile F1 plants. However, it is worth noting that when we conducted this hybridization (November 2014 to February 2015) the Bali accession showed very low (<3%) pod setting, despite setting a lot of flowers. However when the same accession was grown in the following year (November 2015 to February 2016) it set more of pods (~15%) (Sujinna and Somta, unpublished data). Environmental conditions appear to influence fertility of the Bali cultivated V. vexillata and possibly its offspring. Additional hybridization study is necessary to determine genepool of the Bali cultivated V. vexillata.

The previous diversity studies in V. vexillata using isozyme markers (Garba and Pasquet 1998; Jaaska 2001) and RAPD markers (Spinosa et al. 1998) showed that the African and American accessions are genetically different. The same result is obtained in this study using SSR markers (Figs. 1, 2, 3). By incorporating morphological variation with isozyme marker data, Pasquet (2000) proposed that V. vexillata from America and Australia were introduced from West and East Africa, respectively. Similarly, Jaaska (2001) proposed that American germplasm was derived from the Western Gondwanaland V. vexillata population by vicariance through a continental drift, and after that evolutionary change in the American germplasm was mostly at intraspecific level. These proposals are supported by our results which demonstrated that V. vexillata from America is genetically similar to that from East Africa (Figs. 1, 2; Table 3), thus suggesting that the American germplasm was derived from East African V. vexillata. Nonetheless, our results (Figs. 1, 2; Table 3) supported the proposal that the Australian accessions originated from East Africa V. vexillata. It is worth noting that the number of West African V. vexillata used in our study was only 21 accessions and different result may be obtained when the sample size from this region is increased.

By using isozyme marker analysis, Garba and Pasquet (1997, 1998) illustrated that V. vexillata accessions from Asia are genetically different from those from Africa and suggested that Asian V. vexillata can be considered a distinct subspecies. These are in contrast to our findings where accessions from Asia were clustered with those from Africa and showed closer genetic relationship than those from Australia and America (Figs. 2, 3). Thus our findings suggested that Asian V. vexillata should not be raised as a new subspecies, although not many accessions of Asian germplasm was used in this study. Again, further study using a wider range of germplasm from Asia is required to determine its genepool and taxonomic rank.

Comparison of genetic diversity in zombi pea with other Vigna species

Previous genetic studies in Vigna species using SSR markers on a large number of germplasm accessions included azuki bean (Xu et al. 2008), black gram (Kaewwongwal et al. 2015), mungbean (Sangiri et al. 2007), and rice bean (Tian et al. 2013). We compared our results in zombi pea with these Vigna species (Table 4). Average number of alleles per locus in zombi pea (13.65) was lower than azuki bean (23.92) and mungbean (16.26), but higher than black gram (9.05) and rice bean (12.92). Gene diversity of zombi pea (0.72) was close to azuki bean (0.76), and greater than those of mungbean, blackgram and rice bean. Average allelic richness of zombi pea (13.65) was higher than mungbean (4.82) and blackgram (9.01), but almost the same as rice bean (13.10) and lower than azuki bean (16.09). The high diversity found in zombi pea is likely due to wide geographical distribution of the germplasm used in this study.

Table 4 Comparison of number of alleles per locus, gene diversity (H E), outcrossing rate (t), PIC value and average allelic richness (Ar) between Vigna vexillata and other Vigna species
Full size table