Genetic relationship of mungbean and blackgram genotypes based on agronomic and photosynthetic performance and SRAP markers

Genetic identification is at the base of parental selection in varietal development programs. Agronomic and photosynthetic characters and sequence-related amplified polymorphism (SRAP) markers were analyzed for two legume species which included 23 mungbean (Vigna radiata (L.) Wilczek) and four blackgram (Vigna mungo (L.) Hepper) genotypes. The results revealed that the seeds/pod, plant height, pods/plant, pod length, days to flowering, and all photosynthetic characters studied had a significant correlation with the yield/plant. Using UPGMA analysis with phenotypic data, five clusters and two individuals were identified. Twenty-five SRAP primer combinations generated 562 amplified bands, of which 507 were polymorphic (90.2%). The average numbers of scorable and polymorphic bands/primer pair were 22 and 20, respectively. Two major clusters coincided with two species having a 100% bootstrap value. Within the mungbean cluster, there were two subclusters containing 12 and 11 mungbean genotypes. Mantel’s test demonstrated that the polymorphisms given by SRAPs were associated with agronomic and photosynthetic variability (r = 0.734, p < 0.01). These results allow promising mungbean genotypes to be identified through genetic diversity and field performance which can be utilized as potential parents towards future breeding programs. Moreover, the factors which contribute most to yield/plant can be simultaneously used as selection criteria for yield improvement.


Introduction
In Thailand, the mungbean (Vigna radiata (L.) Wilczek) and blackgram (Vigna mungo (L.) Hepper) are considered to be strategic crops for local and national agribusiness. They contain significant amounts of enzymes, phytonutrients, and antioxidants within their dried seeds, forage or green/black pods, and fresh seeds, which are essential for maintaining better health (Somta and Srinives, 2007). The growing awareness of their benefits has led to an increased demand in recent years. However, production is not sufficient to keep up with demand. This is mainly because the existing varieties have a genetically low yield potential and are susceptible to biotic stresses (insect pests and diseases) and abiotic stresses (salinity, heat stress, waterlogging, and drought). Therefore, new varieties with an improved high yield potential and resistance to stresses are required as soon as possible. Finding suitable parents is the initial step in any breeding approach. Phenotyping is the usual way to allow breeders to select parents with the best characteristics. For these two Vigna species, only few studies have been conducted on photosynthetic performance diversity and relationships in combination with the agronomic characters (Hossain et al., 2009;Islam and Razzaque, 2010;Gao et al., 2015). With advances in molecular biology, DNA markers provide a feasible choice for assessing genetic information because they can overcome several limitations found by using phenotyping alone i.e., the magnitude of environmental factors on trait expression, the limited number of characters, and discrimination of genotypes with high phenotypic similarity (Smith and Smith, 1989). Recently, genome/transcriptome sequencing projects have evolved which influence a trend away from structural markers towards functional markers. Functional markers located near any genes are very useful for assessing functional diversity due to the polymorphisms generated from the exon and intron regions (Poczai et al., 2013). SRAP developed by Li and Quiros (2001) is a simple functional marker and has gained in popularity over other multilocus markers. Because of its simplicity, high reproducibility and because no prior genome sequence is needed, this method has become attractive for breeding programs. Aneja et al. (2013) and Alghamdi et al. (2017) found that all SRAP primer combinations exhibited 100% polymorphism, thereby being useful in genetic assessment amongst mungbean accessions, as well as other Vigna spp. The objectives of this study were: i) to evaluate agronomic and photosynthetic performance of 23 mungbean and four blackgram genotypes and ii) to estimate their genetic diversity and relationships using SRAP markers.

Plant materials
A total of 23 mungbean and four blackgram genotypes were included in this study (Table 1). Among mungbean tested genotypes, five were Thai certified varieties that are popularly grown by Thai farmers; 'CN36', 'CN72', 'CN84-1', 'KPS1', and 'SUT1', and one was a Suranaree University of Technology (SUT) developed variety; 'SUT4' ( Table 1). The first three varieties were developed by Chai Nat Field Crops Research Center, Thailand, while 'KPS1' and 'SUT1' were developed by Kasetsart University, Kamphaeng Saen Campus and Suranaree University of Technology, Thailand, respectively. Seventeen other mungbean genotypes were obtained from the World Vegetable Center in Taiwan. Four blackgram genotypes were also included, two of which were Thai certified varieties; 'CN2' and 'CN80' developed by the Chai Nat Field Crops Research Center, and the others were 'BR-1' and 'PAK40592' from the World Vegetable Center collection.
Experimental site, design and crop establishment A field experiment for agronomic and photosynthetic characteristics was carried out at Suranaree University of Technology Farm, Nakhon Ratchasima, Thailand (latitude: 14°52'39"N, longitude: 102°00'15"E, altitude: 227 m) during March to June, 2020. The field trial was established in a randomized incomplete block design with four plots per genotype. Four replications were used when the plants of each genotype in each plot grew well all the way up to field characterization, while three replications were used if uneven growth was observed in one of the plots. In each replication, seeds were planted in two-meter-long rows with a spacing of 0.5 and 0.2 m between rows and between plants within rows, respectively, which was kept after thinning. Two seedlings were maintained per hill for each genotype. Agronomic practices were performed as described by Khajudparn (2009 Agronomic and photosynthetic characterization Eleven quantitative agronomic traits were measured including the number of days to flowering, days to maturity, clusters/plant, branches/plant, plant height, pods/plant, pod length, pod width, 100 seed-weight, seeds/pod, and yield/plant. Measurement techniques followed those established by IBPGR Secretariat (1980) and Chai Nat Field Crops Research Center (2018). Ten randomly selected plants in the middle of each plot were used for measurement (Table 2). Four photosynthetic traits were measured with three plants in each plot, which were selected so as to be the same plants used for agronomic measurements on four consecutive cloudfree days. Disease responses were evaluated according to Chankaew et al. (2011) and Khajudparn et al. (2007) for Cercospora leaf spot (CLS) and powdery mildew (PM) in 2017 and 2018, respectively, when diseases were evenly spread across the fields. SPSS 16.0 software (SPSS Inc., Chicago, IL, USA) was used to perform an analysis of variance (ANOVA) followed by Duncan's post hoc statistical tools of agronomic and photosynthetic traits and disease responses. This software was also conducted to calculate correlation of agronomic and photosynthetic traits. Maximum length of four to ten pods (in case of curved pods, the longest straight line from the base to the tip of pod was measured.). Average of ten plants/plot. 8 Pod width (mm) Maximum width of four to ten pods. Average of ten plants/plot. 9 100-seed weight (g) Weight of 100 randomly selected seeds. Average of ten plants/plot (50 seeds from two plants were combined if necessary).
10 Seeds/pod Number of seeds/pod of four to ten pods. Average of ten plants/plot. 11 Yield/plant (g) Total seed yield from two harvests. Average of ten plants/plot. 12 Photosynthetic rate (Pn) (µmol m -2 s -1 ) Three uppers most fully-expanded terminal leaves of 40 days-old plant (full bloom stage, R2 showing the greatest photosynthetic potential in mungbean (Gao et al., 2015)) were measured by a

SRAP analysis
Fresh young leaves from three individual plants of each genotype, reaching homozygosity whose uniformity had been previously checked, were taken and bulked. Genomic DNA was isolated according to the modified CTAB protocol of Lodhi et al. (1994). DNA concentration and purity were assessed by a ND-1000 spectrophotometer at A260 and A280 (NanoDrop Technologies, Inc., Wilmington, DE, USA). Five different forward and reverse primers were combined randomly to generate 25 SRAP primer combinations (Table 3). A polymerase chain reaction (PCR) was carried out in 20 μL reaction volumes containing 150 ng of DNA, 1x buffer (150 mM Tris-HCl, pH 8.75 at 25 °C, 500 mM KCl, 20 mM MgCl2, 1% Triton X-100), 200 μM of each deoxyribonucleotide triphosphate (dNTPs), 0.6 μM of each forward and reverse primer, and 1U of Taq DNA polymerase (Vivantis, Selangor Darul Ehsan, Malaysia). The PCR conditions according to Aneja et al. (2013) were followed. T100 TM Thermal Cycler (Bio-rad Laboratory, Inc., California, USA) was used for DNA amplification. The PCR products were separated on 6% denaturing polyacrylamide gel at 200 V using vertical electrophoresis for 70 min. The gel was stained using silver nitrate (Sambrook and Russell, 2001). The molecular weights of the DNA bands were estimated by comparison with a 100 bp DNA Ladder as a marker (Invitrogen, California, USA).  The SRAP amplified bands were coded as "0" and "1" for their absence or presence, respectively, similar to other dominant markers. PIC was calculated according to PIC = 1-ΣPi 2 , where i is the total number of alleles detected for the SRAP marker, and Pi is the frequency of the i th allele in the genotypes studied. The unweighted pair group method average (UPGMA) dendrogram of agronomic and photosynthetic characters and polymorphic SRAP loci was constructed based on Euclidean distance to measure the genetic dissimilarity coefficients in a pair-wise comparison across all genotypes using PAST software (version 4.03) with a bootstrap frequency of 1000. The goodness of fit for the genotypes to a specific cluster in the UPGMA algorithm was executed by the normalized Mantel statistic Z test (Mantel, 1967). Correlation between two cophenetic metrics from agronomic and photosynthetic characters and SRAP markers was analyzed based on the Mantel matrix correspondence test through XLSTAT software (version 2015) (Addinsoft, Inc., Paris, France) (Mantel, 1967). A correlation value (r) appearing greater than 0.5 indicates that there is a statistical significance at 0.01 probability level, if over 15 taxonomic units are observed (Lapointe and Legendre, 1992). PAST software (version 4.03) was used again to construct principal coordinate analysis (PCoA) to represent the clustering pattern of genotypes in the first two principal coordinates, which accounted for the highest variation.

Results and Discussion
Agronomic and photosynthetic character analysis The mungbean and blackgram genotypes displayed significant differences for all of their agronomic characters and for the two disease responses (p < 0.01) but not for their photosynthetic characters (p > 0.05) ( Tables 4 and 5 Crop yield implicates a variety of other contributing component characters. An analysis of the correlation between yield and its independent variables can potentially identify their relative significance to improve higher yield. In this analysis, yield/plant corresponded to several agronomic characters: seeds/pod (r = 0.612 ** ), plant height (r = 0.569 ** ), pods/plant (r = 0.435 ** ), pod length (r = 0.430 ** ), and days to flowering (r = 0.200 * ), as well as all photosynthetic characters studied including Pn (r = 0.463 ** ), Gs (r = 0.406 ** ), WUE (r = 0.235 * ), and Tr (r = 0.210 * ) ( Table 6), indicating that increasing these characters can possibly improve yield. However, days to maturity, clusters/plant, branches/plant, pod width, and 100-seed weight showed no correlation with yield. Most previous reports of yield and agronomic character relationships in mungbean also found that seeds/pod and pods/plant exhibited strong positive correlation with yield, while the associations with other characters, i.e. days to flowering, plant height, or pod length were not consistent among different works (Khattak et al., 1995;Khajudparn and Tantasawat, 2011;Mondal et al., 2011;Das and Barua, 2015;Hemavathy et al., 2015;Garg et al., 2017;Kumar et al., 2018;Manivelan et al., 2019;Tahir et al., 2020). This may be largely due to the differences in the plant materials, range of the characters, and evaluation time or environmental conditions studied in each work. For yield and photosynthetic character relationships, Srinivasan et al. (1985) demonstrated that photosynthetic rate was significantly associated with seed yield of mungbean at early pod development but not at the vegetative stage. However, Islam and Razzaque (2010) found that Pn, Gs, Tr, and WUE were not correlated with seed yield of mungbeans. It should be noticed that the photosynthetic rate and its related factors are not always correlated with crop yield (Curtis et al., 1969;Rhodes, 1972;Long et al., 2006). These inconsistent attributes may result from the high vulnerability of the plant's photosynthetic process to changes in environmental factors over a short period of time under evaluation or because of changes in photosynthesis and yield relatedness with growth stages (Islam et al., 1994). In addition to the relationships between yield/plant, the interrelationships between other characters were also evaluated. For example, days to flowering was significantly positively related to days to maturity, plant height, pod length, seeds/pod, and Gs, suggesting that selection for early genotypes may reduce plant height, pod length, seeds/pod and Gs. Notes: CLS and PM were evaluated in 2017 and 2018, respectively. CLS infection was scored at 65 days after sowing (DAS) on a scale of 1-5, where 1 = no disease symptom, 2 = 1-25% of total leaf area infected, 3 = 26-50% of total leaf area infected, 4 = 51-75% of total leaf area infected, and 5 = 76-100% of total leaf area infected. PM infection was scored at 65 DAS on a scale of 1-9, where 1 = no disease symptom, 2 = 2-3 lesions on the lower part of leaves, 3 = 2-3 lesions on the lower part of leaves, where spore formation can be observed, 4 = full spore formation on the lower part of leaves, and a few lesions can be observed on the middle part of leaves, 5 = like number 4, but chlorosis leaves and much of spore formation can be observed, 6 = like number 5, but full spore formation can be observed, 7 = spore formation on all parts of leaves, and 25% dry leaves can be observed, 8 = like number 7, but 25-50% dry leaves can be observed, and 9 = like number 7, but over 50% dry leaves can be observed. Means within the same column not showing the same letters are significantly different (p ≤ 0.05) based on DMRT. We also found strong positive correlations between clusters/plant with branches/plant and pods/plant. This finding is not surprising because the pods increased, together with the clusters, which are in turn formed by the branches. Note that clusters/plant, branches/plant, and pods/plant were strongly negatively related to pod length, pod width, and 100-seed weight. In this study, the mungbean genotypes, i.e. 'ML-131', 'VAR A-G', and 'BARI MUNG2', possessed higher clusters/plant, branches/plant, pods/plant and lower pod length, pod width, and 100-seed weight. Moreover, there were significant positive correlations between plant height with pod length, pod width, seeds/pod, Pn, Gs, and WUE and significant negative correlation with branches/plant. The genotypes, i.e. 'CN36' and 'EG-MD-6D', appeared to have taller plant, higher pod length, pod width, seeds/pod, Pn and lower branches/plant.
The average pair-wise Euclidean distance of the phenogram among all genotypes based on all the agronomic and photosynthetic variables and disease responses was 5.483, ranging from 1.015 ('NM92' vs. 'WALET') to 9.386 ('SUT1' vs. 'PAK40592'). The cophenetic correlation coefficient value of 0.7836 (p < 0.01) was observed, demonstrating a high association for these genotypes to a specific cluster represented in the dendrogram ( Figure 1A). However, bootstrap analysis revealed high values only for cluster I, IV and V, and lower values in the remaining nodes, indicating a lack of robustness of the clustering in some clusters. A low support value can be obtained if genotypes cover an intermediate position between major groups (García-Martínez et al., 2006). All blackgram genotypes were grouped in cluster I. Several mungbean genotypes formed four clusters including cluster II, III, IV and V consisting of five, nine, five, and two genotypes, respectively. In cluster II, 'V4718', 'V4758', and 'V4785', originated in India whose genetics are related to disease resistance in Thailand (Poolsawat et al., 2017;Tantasawat et al., 2020;Tantasawat unpublished data) showed high resistance to CLS and PM (Table 4). However, the others failed in their resistance to both diseases, although some of them, i.e. 'PUSA-105', which also originated in India was found to have high resistance to CLS in India (Marappa, 2008), indicating the influence of race specific resistance. Cluster III and IV consisting of several mungbean representatives from different regions including Thailand, the Philippines, Australia, Pakistan, Taiwan, and Indonesia are interesting, as members in both clusters shared similar characters, particularly pod width, pod length, and 100-seed weight (Table 7). However, the average plant height of cluster IV was higher than that of cluster III which may separate them into two distinct clusters. Note that several members, i.e. 'NM92', 'CES55', 'WALET', and 'TAINAN SEL#5' in cluster III and 'CN36', 'CN84-1', and 'SUT1' in cluster IV also produced high yield/plant. Some of these varieties which originated in other countries, i.e. 'WALET', the local high yielding variety of Indonesia (Hakim, 2008) also had high yielding potential when grown in Thailand. Moreover, Thai certified varieties including 'CN36', 'KPS1' and 'SUT1' in cluster IV, as well as 'SUT4' in cluster III, which were reported to have resistance to CLS and PM were susceptible to both diseases in this study. This may be attributed to their resistance breakdown by virulent pathogens. Two mungbean genotypes including 'ML-131' and 'BARI-MUNG 2' showing high pods/plant and high yield/plant were grouped into cluster V. In previous studies, cluster analysis based on the characters related to yield, i.e. seed yield/plant, biological yield, and harvest index was not allocated in different clusters/subclusters because they characterized a large number of mungbean accessions, i.e. 340 and 533 accessions in Yimram et al. (2009) andTahir et al. (2020), respectively. While NM94 and VAR A-G were individuals, not grouped into any clusters. The PCoA revealed that the utmost total variation of 49% from the sum of 89% was largely explained by the first principal coordinate, and 40% was from coordinate two. The clustering pattern of the PCoA was similar to the dendrogram generated by UPGMA analysis. However, most mungbean genotypes on the PCoA axis were not well distinguished as seen in the dendrogram ( Figure 1B).

SRAP analysis and genetic relationships
A set of 25 SRAP primer combinations revealed a high polymorphism percentage of 90.2% (representing the profiles of 507 polymorphic bands from 562 scorable bands with a size range of approximately 200 to 1,400 bp as shown in Table 3). However, Aneja et al. (2013) and Alghamdi et al. (2017) found 100% polymorphism among Indian mungbean varieties and Saudi mungbean germplasms, respectively. Although we obtained a comparatively lower polymorphism percentage than those of previous studies, our study used higher numbers of primer combinations and scorable DNA bands, both of which may affect polymorphism percentage. We used silver staining on polyacrylamide gel, which detected higher number of scorable bands than ethidium bromide staining on agarose gel as used in previous study (Aneja et al., 2013). These results indicate that polyacrylamide gel with smaller pore sizes and higher resolution than agarose gel can effectively visualize and differentiate small DNA bands, mostly obtained from multilocus marker systems (Bassam and Gresshoff, 2007;Tantasawat et al., 2010). Each primer combination generated an average of 22 bands, of which 20 exhibited polymorphism. The lowest number of polymorphic bands was achieved with the primer combination me2em3 (7) with polymorphism percentage of 50. Whereas, the primer combinations me1em2 generated the highest numbers of polymorphic bands (38). This primer combination together with me3em1 (21), me3em2 (20), me3em4 (22), me4em3 (19), me4em4 (21), and me4em5 (16) had the highest polymorphism percentage of 100. From all primer combinations, an average polymorphism percentage of 89% was revealed across all genotypes. This was comparable to that found in mungbean and blackgram (89.51%) using inter-simple sequence repeat (ISSR) markers (Tantasawat et al., 2010). However, the polymorphism levels may be restricted by the self-pollinated nature of these two species.
The PIC values for each primer are most often analyzed to estimate discriminatory power by accounting both the number of alleles at a locus and their relative frequencies (Nagl et al., 2011), and it can reach a maximum of 0.5 (in case of dominant markers). The PIC values of ≥ 0.28 in eight primer combinations; me1em2, me2em2, me3em3, me4em1, me4em3, me4em4, me5em3, and me5em5, and 100% polymorphism in three of these primer combinations; me1em2, me4em3, and me4em4 suggested that they are preferable for genetic characterization. These results demonstrate that using only eight primer combinations, including me1em2, me2em2, me3em3, me4em1, me4em3, me4em4, me5em3, and me5em5 allows one to distinguish the genotypes evaluated within and between both species. Therefore, the use of this marker system can reduce cost, time, and labor in the identification of varieties of mungbean and blackgram. Tantasawat et al. (2010) obtained similar findings when using only six most informative ISSR primers. However, genetic identification based on SRAP loci is more likely to reflect phenotypic trait expressions than other multilocus markers including ISSR markers, which are largely located in non-coding genomic regions (Shao et al., 2010).
Pair-wise Euclidean distance coefficient values varied from 4.107 ('CN36' vs. 'CN72') to 57.846 ('V4758' vs. 'BR-1') with an average of 26.447. Average distance coefficient between V. radiata genotypes was 15.776. 'CN36' and 'CN72' which are widely used commercially in Thailand and possess some relationship in their pedigree (Table 1) showed the lowest Euclidean distance (4.107), while 'V4758' and 'CES55' without any relationship showed the highest Euclidean distance (21.566) in the mungbean group. 'CN36', 'CN72', and 'CES55' had high pod length, pod width, and 100-seed weight, whereas 'V4758' did not. In the blackgram group, the genetic distances ranged from 24.749 to 28.734 found between 'CN2' with 'CN80' and between 'CN80' with 'PAK40592', respectively, with an average of 27.120. Phylogenetically genotypic data between both species revealed their genetic distance of 55.749, indicating a high interspecific diversity that reflects the differences observed in their several agronomic and photosynthetic characters, i.e. plant height, pod length, seeds/pod, yield/plant, Pn, Tr, and Gs. The SRAP-based UPGMA dendrogram showed two distinct clusters with 100% bootstrap value where all four blackgram genotypes were representatives of cluster I, while cluster II comprised all 23 mungbean genotypes (Figure 2A). The coefficient cophenetic value with Mantel's test reached 0.9959 (p < 0.01), indicating high reliability of distance matrix data represented in the dendrogram.  In cluster II, there were two subclusters largely consistent with their special features reported between high yielding potential (subcluster IIA) and resistance to diseases (subcluster IIB) (Table 1). These results support our previous report, which used 11 expressed sequence tags-simple sequence repeat (EST-SSR) primer pairs to determine the genetic diversity and relationships of these 27 genotypes (Chueakhunthod et al., 2018).
SRAP-based clustering was apparently clearer than that of the EST-SSR system because 11 of those 18 genotypes reported to have resistance to diseases including 'PUSA-105', 'VAR A-G', 'NM92', 'TAINAN SEL#5', 'GELATIK', 'BARI MUNG2', 'NM94', 'ML-131', 'V4718', 'V4758', 'V4785' were separated from the others reported to have high yielding potential, while only five of these were grouped together by EST-SSR markers. This may be due to the higher number of polymorphic bands derived from SRAP (507 bands) compared to EST-SSR (56 bands). All the genotypes which originated from India with resistance to diseases were grouped in subcluster IIB, separated from subcluster IIA where the genotypes from Thailand and the Philippines as well as Indonesia and Australia with high yielding potential were grouped together. There were 12.909 and 10.886 genetic distances among genotypes from the Philippines and Thailand, respectively, which were lower than those from India (17.318 genetic distance in subcluster IIB). In the case of the Philippines individual similarity, this can be explained by the fact that the Philippines improved varieties or lines were initially developed from only six purified local varieties as parents (Ballon et al., 1978;Catipon et al., 1988;Somta et al., 2009). Several outstanding Philippines varieties/lines developed were intensively used as parents both nationally and internationally (Catipon et al., 1988), and this may account for the sharing of common parents and a narrow genetic base. For example, MG50-10A obtained from crossing between 'GLOSSY GREEN S1' and 'GLABROUS GREEN' was utilized to develop 'CES55' and 'BPI GLABROUS #3'. Somta et al. (2009) also revealed this close relationship among the Philippines mungbeans when using SSR markers.
Thai mungbean genotypes in this study also shared common parents in their pedigree. For example, 'CN84-1' was derived from the γ irradiated mutants of 'CN36', or 'CN72' was selected from mutated 'KPS2', which contains genetic background of 'CN36' (Table 1) subcluster IIB; 'V4718', 'V4758', and 'V4785', whose CLS and PM resistance genes were identified by Chankaew et al. (2011);Poolsawat et al. (2017) and Tantasawat (unpublished data), and found to be highly resistant to both diseases (Table 4) had genetic distance of 14.216. In subcluster IIA, 'SUT1', 'CN36', 'CN84-1', and 'EG-MD-6D' which had high yielding potential with regard to high yield/plant and large seed size and were grouped in the same subcluster IV of the dendrogram based on field performance ( Figure 1A) exhibited high genetic relationships with only 10.729 genetic distance. To understand the genetic relationships among the genotypes better, the PCA was also performed. Two principal coordinates explained 77% of the total variation, of which the first two axes accounted for 68 and 9%, respectively. Similar to UPGMA analysis, this method showed two distinct groups according to species similar to the cluster analysis ( Figure 2B). SRAPs implicate a greater influence of the characters due to their exclusive strategies, which are proved with an association of 0.734 (p < 0.01) between this functional marker and the characters related to field performance, indicating a concordance between both marker systems. This finding was consistent with Aneja et al. (2013), who stated that the information given by SRAP enabled one to classify several mungbean genotypes on the basis of micronutrient content. Ferriol et al. (2003) also reported that SRAP corresponded more closely to morphological characters and to the evolutionary history by means of the morphotypes than amplified fragment length polymorphism (AFLP). This coherence between SRAP and morphological pattern has also been reported in other legume crops, i.e. pea and chickpea (Espósito et al., 2009;Khan et al., 2016).
The mungbean and blackgram genotypes were largely grouped into separate clusters according to species by each marker system. However, some variations in the clustering from different marker systems should also be noted ( Figures 1A and 2A). SRAPs proved to be more effective than agronomic and photosynthetic characters in unravelling genetic differentiation between these V. radiata and V. mungo genotypes, since the bootstrap value of the V. radiata and V. mungo clusters was 100%. In addition, cophenetic correlations for testing the goodness of fit from SRAPs were higher than those of the agronomic and photosynthetic characters (0.9959 vs. 0.7836). Moreover, the SRAP marker loci were capable of classifying most mungbean genotypes according to their special features between disease resistance and high yielding potential.

Conclusions
These results demonstrate that the multilocus SRAP marker system with apparent strategies in functional analysis mediated by the amplification of DNA using a single forward primer with numerous interchangeable reverse primers has proved to be a valuable tool in determining variability and the relationships among these two Vigna species and in trait mapping or marker-assisted selection. The promising mungbean genotypes, including Thai certified varieties, i.e. 'SUT1', 'CN84-1', and 'CN36' and those from other regions, i.e. 'EG-MD-6D' from the Philippines possessed characters, i.e. pod length, pod width, 100-seed weight, yield/plant, and photosynthetic characters that are very useful for improving yield. All of these candidates can contribute to future breeding programs, while the others, i.e. 'V4718', 'V4758', and 'V4785', which contain resistance genes enabling them to have high resistance to CLS and PM can also be used as parents in a gene pyramiding program for durable disease resistance. In addition, seeds/pod, plant height, pods/plant, pod length, days to flowering, and all photosynthetic characters studied including Pn, Gs, WUE, and Tr, which contributed the most to yield/plant can be used as selection criteria for yield improvement in both mungbean and blackgram.