Bambara nuts qualitative traits variation
The findings revealed that there is significant morphological variability in qualitative characteristics such as terminal leaflet shape, growth habit, seed color, eye pattern, testa pattern and seed shape. Similar findings by Beket et al., 2019 reported that the majority of the accessions with 74.26 % produced elliptic leaflets, followed by accessions with oval leaflets having 17.82 %. Only 6.93% of the accessions had lanceolate leaflets, and round leaflets were uncommon with 0.99 %. In a similar study, the terminal leaflet shape reported that 64 % were oval, 24% were elliptic and 12% were lanceolate (Valombola et.al., 2022). The study also reported that 52.5% of the Bambara groundnut germplasm had a semi bunch growth habit while 42.5% had a growth habit which is spreading and 5% had a growth habit which are bunch type. This is in agreement with the findings by Gbaguidi et al., 2018, in Benin who reported that there were more bunch landraces among the three types of growth habits. In a similar, Beket et al., 2019 reported that the spreading type was the most common with 59.41% among the accessions studied, followed by the bunch type with 31.68%. The semi-bunch style was less common with 8.91 %. Another study reported that 44 % were bunch type, 24 % were semi-bunch, and 32 percent were slightly spread (Valombola et.al., 2022). That could be explained by the fact that the bunch landraces were more advantageous to the farmer particularly during harvest when the roots and stems were unearthed easily and with less seed left in the soil (Gbaguidi et al., 2018).
A study by Judicaëlle et al., 2020 reported that seed color showed a high degree of variability, with cream seed predominating having 31.11 %. Colors that had the lowest proportions include dark brown seeds, black, dotted dark brown on a cream background, light brown marbled spots on a cream background and black rhombic spots on a cream background on the micro hair but not on the hairy ends with 1.11 %. Similarly, eye patterns were reported to be irregular brown with 3.33 %, circular grey with 2.22 %, thick grey on both sides with 1.11 %, butterfly grey with 52.22 %, triangular grey with 7.77 %, irregular black with 1.11 %, and dark red butterfly with 6.67 % (Judicaëlle et al., 2020). Another study reported that 64 % of the accessions had no eye color, 16 % had amber color, and 4 % had brownish, brown, grey, black, or purple color (Valombola et.al., 2022). It was also reported that Seed testa pattern was 20% for the tan color, 12% for the red and cream, 8% for the dark tan and purple, and a minimum of 4% for speckled brown, cream, brown, and black (Valombola et.al., 2022).
Principal components analysis for Bambara groundnut quantitative traits
Principal Component Analysis (PCA) reported a total variation of 53%. The first principal component (PC1) has an Eigen value of 6.791 and represents a cumulative proportion of 27%, representing the equivalent of four individual variables. The results are in agreement with the findings by Olanrewaju et al., 2021, who reported that PC1 and PC2 accounted for 42.3 % of all observed variances. PC1 accounted for 24.67% of total variations, while PC2 accounted for 17.63%. Khan et al., 2020 reported a 45.8% variations in Bambara groundnut and some of the traits contributing to the variation were yield (Kgha-1). In another study by Jonah et al., 2014 reported that a 71% of the observed variation in the first three principal components with 100 seed weight as one of the traits contributing positively to the variations while a research by Mohammed et al., (2020) reported a 79% of the total variations with seed weight and biomass contributed to the total variations among the Bambara groundnut accessions. A similar study reported that the first three axes expressed 84.01 % of the total variability with yield as one of the traits that contributed positively to the total variation (Touré et al., 2012).
Biplot of principal component analysis for Bambara groundnut quantitative traits
PCA biplot loading for both variables and germplasm revealed how strongly each trait influences a PC and how they are related. Based on Principal Component (PC) 1 and 2, Yield (Kgha-1), were positively correlated with 100 seed weight in grams, number of leaves per plant, and petiole length. Therefore these traits are the standard features that can be used to determine similarities and diversity of species phenotype. In other words, they show large variations across the germplasm studied, implying that they are the best discriminators of the morphological traits under consideration. In a similar study, it was reported that yield and 100 seed weight was the major contributing traits and these can be used as discriminant in seed production for Bambara groundnut. Internode length and flowers per peduncle on the other hand, had the shortest vectors, indicating that they are the least discriminator of the morphological traits under study (Olanrewaju et al., 2021; Valombola et al., 2019). A study by Mohammed et al., 2020 also reported that the PCA biplot revealed a significant variation among Bambara groundnut genotypes.
Analysis of Molecular Variance (AMOVA) for the 188 Bambara groundnut germplasm samples genotyped
Analysis of molecular variance revealed a variation of 13 % within the Bambara groundnut germplasm genotyped and 84 % variation among the Bambara groundnut germplasm genotyped while 3 % variations were observed among the populations. This is in contrary with the phenotypic data, which revealed a total variation of 53 % among the Bambara groundnut germplasm using Principal component analysis. But the results are in agreement with the study by Uba et al., 2021 where the AMOVA reported that 89% of genetic variation occurred among the populations, 8% between regions, and 3% between populations. The greater the genetic diversity of the germplasm, the greater the likelihood of success in breeding desirable traits. The high percentage value of genetic diversity within-population obtained from AMOVA could be due to natural adaptation or extensive seed exchange among farmers between collection sites, or it could be due to the population's common origin, which could have resulted in Bambara groundnut growers using the same seed continuously, without new introductions. In Malawi, seed sources for planting Bambara groundnut by farmers includes; farmer-saved seeds, exchange and market purchase (Pungulani et al., 2012). Due to the crop's autogamous breeding system, this is likely to result in a heterogeneous population of landraces and thus higher intra-landrace diversity, as opposed to the homogeneous population that would be expected.
Pairwise Nei’s standard genetic distance between the ten populations of Bambara groundnut evaluated using 1048 SNP markers.
The average Nei genetic distance between germplasm within each population revealed Pop 5 vs Pop3, Pop 6 vs Pop 3 and Pop 10 vs Pop 2 had the smallest genetic distance of 0.000. The study also revealed that the closest distance was observed in Pop 7 vs Pop 6, Pop 8 vs Pop 3 and Pop 8 Vs Pop 5 with a genetic distance of 0.001. The largest distance were observed in Pop 7 vs Pop 2 with a genetic distance of 0.111. Uba et al., 2021 reported similar findings where the closest distance based on geographical regions were reported to be 0.002.
Cluster analysis of the 188 Bambara groundnut germplasm samples genotyped
Cluster analysis grouped the genotypic data for 188 germplasm samples into 10 clusters with the largest cluster having 38 germplasm while the smallest cluster having 7 germplasm. The dendrogram based on geographical distribution of the germplasm revealed that most of the Bambara groundnut germplasm from the same region did not cluster accurately based on their collection site. This could be due to low genetic differentiation between populations, implying that the genetic background of Bambara groundnut population does not always correlate with their geographical origin or region. Uba et al., 2021 reported that the grouping of some accessions from different regions into the same cluster may indicate the degree of relatedness between accessions from different regions, which can be attributed in part to the transfer and exchange of seeds between regions via gene banks and human migration.
Geographical distribution of the 188 Bambara groundnut germplasm based on cluster analysis
The current study has identified the number of germplasm per geographical area as revealed by genetic analysis and it has been discovered that out of the 188 germplasm samples genotyped, NPGRC had a total of 62 germplasm samples followed by Chipala EPA with 32 germplasm samples and the least was Kasungu with 4 germplasm samples. Similar study by Olukolu et al., 2012 reported that Bambara groundnut germplasm were grouped into several clusters based on geographical region. Ontong et al., 2021 also reported similar observation on geographical distribution of Bambara groundnut accessions using simple sequence repeats (SSR) markers.