Genetic Diversity Studies in 29 Accessions of Okra (Abelmoschus spp L.) Using 13 Quantitative Traits

Aims: Twenty nine (29) local and exotic lines (accessions), of okra (Abelmoschus spp L.) were evaluated for variation in phenotypic traits. Study Design: They were laid out in a Randomised Complete Block Design (RCBD) with four replications and evaluated based on 13 quantitative characters. Place and Duration of Study: Research farm of the Biotechnology and Nuclear Agriculture Original Research Article Amoatey et al.; AJEA, 5(3): 217-225, 2015; Article no.AJEA.2015.025 218 Research Institute (BNARI), Ghana Atomic Energy Commission (GAEC), Department of Nuclear Agriculture and Radiation Processing, Graduate School of Nuclear and Allied Sciences, University of Ghana, between June 2011 and July 2012. Methodology: The accessions were grown in the field, each on a subplot measuring 3.5 m x 2.5 m, with seeds sown at a spacing of 0.70 m x 0.50 m. Data were collected using the International Plant Genetic Resources Institute (IPGRI) Descriptor List for okra. Results: The accessions exhibited significant variation in all quantitative traits studied. Block coefficients of variation were extremely low, implying that results obtained are reliable and repeatable over replications. Cluster analysis based on Canberra, Furthest Neighbour Similarity Matrix grouped the accessions into two major clusters and subsequently into four sub-clusters, with no duplications, based on the characters studied. Seven pairs of quantitative traits were positive and significantly correlated (P ≤ 0.05) while three were highly significantly associated (P ≤ 0.01). The highest correlation (r = 0.95) was between number of days to 50% flowering (NDFl) and number of days to 50% fruiting (NDFr). Conclusion: The pattern of clustering showed some degree of association between quantitative characters and geographic origin of the collections. Five Principal Components (PCs) accounted for 78.51% of the total variance, with PC1 recording 32.44%. Different traits contributed differently to total genetic variance.


INTRODUCTION
Production and consumption of okra (Abelmoschus spp L. Moench) is widespread across West Africa [1,2,3], where all vegetative and reproductive parts as well as the fresh fruits are used variously for food preparation [2,4]. Minor applications are found in folk medicine and industry [3,5].
In Ghana, the vegetable is accepted for consumption in all regions. It is cultivated as a garden or commercial crop [6]. Intense cultivation is found in peri-urban areas to meet an evergrowing urban population, with targeted exports from elite farmers. Selection of varieties for cultivation is, therefore, based on end-user preference and adaptation to local agro-ecology.
Currently, genotypes available include many locally adapted landraces as well as some exotic lines selected to meet specifications of export destinations in Europe and America. On-going breeding work in okra is limited [7,2]. Hence, characterisation of these genotypes is incomplete.
Characterisation based on phenotypic traits is not easily reproducible particularly, since these traits are influenced largely by environmental variations [8]. In addition, it requires a large tract of land and/or greenhouse space in which to grow large populations of plants, making it labour intensive and difficult to manage [8,9].However, the tool has remained useful as a necessary first step prior to more in-depth biochemical or molecular studies in okra germplasm exploitation [10].
By and large, the potential value of germplasm is hugely dependent on the efficiency of techniques designed to facilitate detailed study of individual traits and to differentiate among accessions [11,12,13]. Hence, characters recorded on individual accessions can serve as diagnostic descriptors for those accessions [13]; to help breeders as well as genebank curators keep track of such accessions and check for genetic integrity over a number of years of conservation. The objective of the study was to assess variability in quantitative characteristics of some accessions of okra collected across eight out of ten geographic regions of Ghana.

Twenty-nine
(29) accessions of Okra (Abelmoschus spp L.) were assembled from eight geographic regions of Ghana using [14] passport data as indicated in Table 1 below.
The study was conducted at the Nuclear Agricultural Research Centre (NARC) of the Biotechnology and Nuclear Agriculture Research Institute (BNARI), Ghana Atomic Energy Commission (GAEC). The soil at the site is the Nyigbenya-Haatso series, which is a typically well-drained Savannah Ochrosol (Ferric Acrisol) derived from quartzite Schist [15].

Experimental Design and Field Layout
A total land area of 60 m x 32 m was cleared, ploughed and harrowed to a fine tilth for planting. The Randomised Complete Block Design (RCBD) was used with four replications; each replicate measured 30 m x 12.5 m, separated by a distance of 2 m and consisted of 30 subplots (within the block). Each subplot had a dimension of 3.5 m x 2.5 m and spaced by a distance of 1 m.
Field cultivation was done from July 2011 to February 2012. Seeds were sown at a depth of 2 cm, at a spacing of 0.70 m x 0.50 m within and between rows with three to four seeds per hill and later thinned to two after germination. No fertiliser was applied, but weeds were controlled fortnightly and water was supplied during the dry season using watering can.

Data collection
Data were collected on five randomly selected and tagged plants within the central rows, using the International Plant Genetic Resources Institute, [14] Descriptor List for okra. Characters on which data were taken include:

Data analysis
Mean values of data collected were used for Analysis of Variance (ANOVA) and Duncan's Multiple Range Test (DMRT) for mean separation. Correlation analysis was used to determine the degree of association among the traits. Further, the Principal Component Analysis was employed to assess percentage contribution of each trait to total genetic variability among the accessions. Cluster analysis based on Canberra, Furthest Neighbour Similarity Matrix was also employed to obtain a dendrogram depicting the deduced genetic relationships among the accessions based on evaluation of the 13 characters. Genstat Statistical Software Programme [16], Microsoft Excel Software, and Statgraphics Plus XV.I [17] were used for all the data analyses. Table 2 shows phenotypic variability in 13 quantitative traits among the 29 accessions of okra. The accessions exhibited significant variation with respect to all thirteen quantitative characters. DKA recorded the highest number of days to 50% germination (NDG), number of days to 50% flowering (NDF l ) and number of days 50% fruiting (NDF r ). Similarly, Nkran Nkuruma recorded the highest maximum plant height (MPH), maximum number of internodes (MNI) and first fruit-producing node (FFPN).

Variability in Quantitative Traits
In the same vein, Yeji-Local recorded the highest total number of leaves per plant (TNLP) and number of seeds per fruit (NSPF) as did Kortebortor-BAR for stem diameter at the base (STB), and total number of fruits per plant (TNFP). Four other accessions, Asontem NV, Akrave, Amanfrom and Legon fingers recorded the highest values for maximum number of internodes (MNI), first fruit-producing node (FFPN), total number of fruits per plant (TNFP) and 1000-seed weight (TSW), respectively.

Cluster analysis based on 13 quantitative traits
Genetic relationships among the 29 accessions of okra, based on 13 quantitative traits are displayed in the form of dendrogram (Fig. 1), generated using the coefficient of Canberra, Furthest Neighbour Similarity Matrix. Two clusters were formed at (67.90%) similarity, each re-grouping into two sub-clusters, making a total of four sub-clusters at 76.30% genetic similarity. The four sub-clusters comprised 10, 5, 10 and 4 accessions, respectively (Table 4). Clustering pattern revealed in the dendrogram indicates some degree of convergence with geographical origin of accessions. Summary statistics of the 13 quantitative traits (Table 3) also shows great diversity among the accessions.
The first and last sub-clusters exhibited the highest inter-cluster distance and may be useful as sources of variable genes in future okra improvement programmes through hybridisation. The accessions Cs-Legon and Nkran Nkuruma, were the most divergent, and accordingly could be utilised for obtaining heterobeltiosis [7,18]. Cs-Legon, Legon fingers, Atomic, Indiana, Clemson spineless; and Yeji-Local, Kortebortor-BAR and Nkran Nkuruma were placed in subclusters 1 and 4, respectively, coinciding with their geographical origins of collection, a reflection of adaptation to similar environmental conditions or related ancestry. This is in consonance with reports of [19,20].

Correlations among 13 Quantitative
Traits of Abelmoschus spp L. Table 5 shows the associations among thirteen quantitative traits of the various okra accessions. NDG was negatively correlated to all other traits except NFPH to which it was positive, but poorly correlated. Similarly, NDF l and NDF r showed negative correlation with 50.00% and 41.67% respectively, of the other traits. NDF r was positive and significantly associated with TNLP and STB as did MPH with FFN. MNI was positive and significantly correlated with FFN as did also TNLP withNSPF as well as NSPF with NFPH and NFPH with TNFP. FFN was also positive and highly significantly correlated with both NFPH and TNFP. The highest positive and significant correlation (r = 0.95) was between NDF l and NDF r . This corroborates findings of several researchers [2,21,22,23]and suggests that component breeding would be very effective when there is positive association of major yield characters [7] as found in this study. This is in consonance with findings by [24,25], where factor scores of nine and twelve characters for rice accounted for variance among accessions and were mostly correlated with PC 1 , PC 2 , PC 3 and PC 4 . The total contribution of the five principal component axes (78.51%), in this study, was higher than observations made by [21,22,25,26]where the principal component axes contributed 64.32%, 66.37%, 76.62% and 64.5% to variation, respectively. In the current study, all the eigen values were lower than those observed by [22]. First fruit-producing node and total number of leaves per plant were found to have contributed positively and significantly to total genetic variance in this study, confirming a similar observation by [22].

CONCLUSION
The 29 accessions of okra (Abelmoschus spp L. Moench) exhibited great diversity in the 13 quantitative traits studied. Cluster analysis      grouped the accessions into four sub-groups with a bearing on geographical origin. No duplicates were detected while the accessions Cs-Legon and Nkran Nkuruma were the most divergent, and may provide variable genes useful in future okra improvement programmes, through hybridisation. The highest character association (r = 0.95) was found between number of days to 50 % flowering (NDF l) and number of days to 50 % fruiting (NDF r ), implying that selection for one trait will lead to a high positive response in the other. Five Principal Components (PCs) accounted for 78.51% of total variance. The first principal component (PC 1 ) which contributed 32.44% to the total genetic variation was mostly correlated with number of fresh fruits per plant per harvest, first flowering node, total number of fruits per plant, maximum plant height, total number of seeds per fruit, maximum number of internode, stem diameter at the base, number of leaves per plant and 1000-seed weight.