Selection of okra parents based on performance and genetic divergence

A total of 200 okra accessions with wide variability and a potential for genetic improvement were stored in the Vegetables Germplasm Bank of the Federal University of Viçosa (UFV-BGH) in Viçosa, Minas Gerais State, Brazil. The objective of this work was to select parents by genetic divergence and behavior per se in 70 okra accessions from the BGH-UFV by quantitative and qualitative descriptors of economic interest. Analysis of individual and combined variance, by clustering of means by Scott-Knott test, of the accessions by Tocher’s method and selection based on qualitative descriptors and behavior per se using the methodology of sum of inverted positions was made. The variability of the characteristics of the accessions as verified by the Scott-Knott test formed different groups and subgroups by Tocher’s method. Fifteen accessions were selected with the qualitative descriptors, and based on the sum of inverted positions for quantitative descriptors the BGH-132, BGH-547, BGH-693, BGH-740, BGH-961, BGH-7863, BGH7865, BGH-3196 and BGH-4890 okra accessions were selected as potential parents.


INTRODUCTION
India is the diversity center of cultivated okra [Abelmoschus esculentus (L.) Moench] (Haq et al., 2013).Also, this plant has great economic importance, especially in tropical and subtropical regions (Yuan et al., 2014), due to the fact that it does not need high technological input to be successfully grown (Aguiar et al., 2013).The climatic conditions of Brazil, especially in the northeastern, midwestern, and southeastern regions favor the intensive cultivation of okra, mainly by small producers (Aguiar et al., 2013).Okra fruit is a good dietary supplement which is free of fat and low in calories (Patil et al., 2013).Genetic improvement mainly depends on the amount of genetic variability present in the population (Koundinya et al., 2013).In any crop, the germplasm serves as a valuable source of base population and offer major source of variability (Ramya and Senthil Kumar, 2009).The Vegetable Germplasm Bank of the Federal University of Viçosa (UFV-BGH) has more than 200 okra accesses.The genetic characterization of these accessions provides useful information for understanding and harnessing genetic diversity (Cerqueira-Silva et al., 2014).Furthermore, these assessment works can be used to infer the variability of the collection and to identify its potential uses in genetic improvement programs (Gonçalves et al., 2009).
The genetic diversity assessed by biometric techniques or predictive processes, provides opportunity for plant breeders to develop new and improved cultivars with desirable characteristics (Govindaraj et al., 2015) and also provides information of potential parents for breeding programs (Costa et al., 2011).Moreover, multivariate methods based on predictive processes are being used for okra crop (Garg et al., 2011;Singh et al., 2012;Das et al., 2013).The choice of parents based on behavior per se is crucial for breeding programs and can generate superior populations (Santos et al., 2011).The objective of this study was to characterize and evaluate the genetic divergence of 70 okra accessions from the BGH-UFV to select potential parents based on the predictive processes and behaviors per se.

Plant genetic resources and experiments
Seventy okra accessions from BGH-UFV were evaluated in two experiments in a randomized block design with three repetitions, in the Vegetable Crops Sector of the Department of Plant Science, Universidade Federal de Viçosa (20º45'14 "S,42º52'53" W and 648,74 m) in Viçosa, Minas Gerais, Brazil.Two experiments were conducted with 37 and 33 okra accessions evaluated in the first and second experiments between September 2010 and February 2011 and September 2011 and February 2012, respectively.The following controls were used: Santa Cruz 47 (cultivar pure line), TPX903 and TPX4460 (hybrids) (Table 1).

Characterization and morphoagronomic selection
The quantitative and qualitative descriptors proposed by the International Plant Genetic Resources Institute (IPGRI) (1998) and others deemed necessary for their agronomic importance were used to characterize and evaluate the okra accessions (Table 2).The agronomic descriptors of economic interests were: (PI): precocity index (%) -the ratio between the sum of the mass of the fruits of the first six harvests and the total fruit multiplied by 100, (NFr): number of fruits per plant, (AFW): average fruit weight (g), and (TWF): total weight of fruits per plant (g).The accessions were selected as potential parents in three stages: (1) selection based on qualitative descriptors of economic interest (that is, green colour; predominant format of round fruit, angular pods, or smooth fruit surface), (2) per se behavior using the methodology of sum of inverted positions based on the weight posts index (Mulamba and Mock, 1978) for quantitative descriptors of economic interest, and (3) genetic divergence among accessions.Data from each experiment were submitted to the analysis of variance (F test) and the group means were compared by the Scott-Knott test at 5% probability.The best accessions were identified based on the sum of inverted positions methodology classifying the accessions for each of the characters of economic interest, aiming to improve favorable ones, and getting the sum of ranks.The means were compared by Scott-Knott test and used in the sum of the ranks with weights reversed based on access grouping.The accessions of groups of greatest interest received the highest score, and so on.The number of grades allocated was the same as that of accesses.For example, for the characteristic average fruit weight of the second experiment (Table 3), accessions were clustered into three groups: the first and more productive, received the grade three; the intermediate received grade two, and the lower production, received grade one.
The study of genotype × environmental interaction (Freeman, 1973) was made in the evaluation of genetic divergence among accessions to the characteristics of economic interest using the common control in the experiments as well as in the variance analysis and multivariate analysis, the Mahalanobis distance was used as dissimilarity measure (Mahalanobis, 1936) and the Tocher's method to form groups (Rao, 1952).Analyses were performed using the computer application GENES (Cruz, 2013).

RESULTS
The 70 okra accessions showed variability when characterized by qualitative descriptors (IPGRI,1998).In the first stage based on the qualitative descriptors of economic interest (that is, green colour; predominant format of round fruit, angular pods, or smooth fruit surface), 15 of the 70 accessions were selected: BGH-132, BGH-311, BGH-7838, BGH-359 , BGH-547, BGH-693, BGH-740, BGH-961, BGH-7863, BGH-7865, BGH-1248, BGH-3196, BGH-4162, BGH-4890, BGH-4913, 11 from the first experiment and 4 from the second, representing 22% of the accessions.The genotypes showed differences in both experiments (p ≤ 0.05).For quantitative descriptors of economic interest, earliness index (%), number of fruits per plant, and average and total weight of fruits per plant (Table 3).Quantitative descriptors of economic interest allowed observing different groups of means by the Scott-Knott test at 5% probability.However, the average of the total fruit production in the first experiment and the number of fruits per plant in the second was similar to those of the control, while the average fruit weight in the second formed three groups (Table 3).The sum of inverted positions for quantitative descriptors of economic interest (Table 3) selected the BGH-132, BGH-547, BGH-693, BGH-740, BGH-961, BGH-7863, BGH-7865, BGH-3196 and BGH-4890 accessions with the highest scores in the first stage and therefore as potential parents.The scores of these accessions were higher than the control, that is, Santa Cruz 47 and similar to the hybrid in different experiments, with higher scores for the BGH-4890.A total of 13% of genotypes was selected based on their per se to the economic performance descriptors.
The BGH-132 BGH-547, BGH-961 and BGH-7865 accessions were allocated to the group of the earliest production, highlighting the BGH-961, the most precocity index, superior to the hybrid controls (TPX903 and TPX4460) with high precocity.Moreover, productivity of these accessions was similar to that of the control group (Table 3).Interaction at the 1% probability level by F test for genotype × environment for characteristics of economic interest was not observed (Table 4).Cluster analysis by the Tocher's optimization method (Table 5) initially separated the genotypes into two groups; the first group having the largest number of genotypes (98.5%) and the second with BGH-6988 access.However, the total and average weight of fruits and especially the early precocity of the latter were not desired.Eleven subgroups were formed for group I due to its large number of genotypes ( The BGH-132 and BGH-3196 accessions allocated in the first subgroup and the BGH-740 BGH-693, BGH-7865 and BGH-547 in the second subgroup with three controls were selected as parents, based on their behavior per se.  1 Means followed by the same letter, for column belongs to the same group, according to the grouping of Scott-Knott test, at 5% probability. 2Notas received for the group allocated according sum of inverted positions. 3PI Precocity index (%); (NFr): number of fruits per plant; AFW: average fruit weight (g); TWF: total weight of fruits per plant (g). 4 Treatment mean square, significant at the 5% probability.; NFr: number of fruits per plant; TWF: total weight of fruits per plant (g).AFW: average fruit weight (g).
Table 5. Groups (Gr.) and Subgroup (subgr.)formed with 70 okra accessions from the BGH-UFV and three cultivars and distance intra and inter groups by Tocher's method based on Mahalanobis' generalized distance (D²).

DISCUSSION
The variability of the qualitative descriptors between the accessions provides information of potential parents for hybridization and development of new cultivars (Sanwal et al., 2012).These characteristics are linked to the economic value of the okra fruits, as the green colour of the fruit is preferred by consumers (Solankey et al., 2013).The selection of BGH-132, BGH-311, BGH-7838, BGH-359, BGH-547, BGH-693, BGH-740, BGH-961, BGH-7863, BGH-7865, BGH-1248, BGH-3196, BGH-4162, BGH-4890 and BGH-4913 accessions was based on qualitative descriptors, namely, colour of fruit; and predominant format of round fruit, angular pods, or smooth surface.However, some accessions with angular pods format were selected because Santa Cruz cultivar, one of the most planted in Brazil, has fruits with round shape, which does not meet the preference of the foreign market.This shows there is a necessity in the production system, including cultivars with angular pods format adapted to Brazilian climate and to increase export of preferred okra in the foreign market (Purqueiro et al., 2010).Differences between genotypes for quantitative descriptors of economic interest in precocity index, number of fruits per plant, and average and total weight of fruits per plant confirmed the high variability of okra accessions in the germplasm bank, which is important for the genetic improvement of population of this plant (Koundinya et al., 2013).Moreover, these features contribute to increased crop yield, which is the main objective of okra breeding programs (Dhankhar and Mishra, 2005).The BGH-132, BGH-547, BGH-693, BGH-740, BGH-961, BGH-7863, BGH-7865, BGH-3196 and BGH-4890 accessions were selected by the sum of inverted positions for quantitative descriptors of economic interest.This index is an efficient technique to select genotypes, as for other crops (Vivas et al., 2012;Almeida et al., 2014).The highlight of the BGH-132, BGH-547, BGH-961, and BGH-7865 accessions by early precocity and productivity is important because genotypes that bloom first provide earliest crops with longer fruiting period and higher number of fruits per plant (Medagam et al., 2012).The selection and introduction of these accessions enable breeding programs to explore heterosis obtaining cultivars or hybrids with higher productivity and income (Jindal et al., 2010).The non-significance of genotype × environment interaction for traits of economic interest demonstrated no differences in responses compared to the control to environmental variations, which may complicate studying morphological and agronomic traits of economic interest (Kyriakopoulou et al., 2014).To determine the divergence, the genetic distances between the genotypes were initially estimated by the Generalized Distances of Mahalanobis.Group with single genotype (II) indicates their independent identity and importance due to unique characters possessed by them.The intra group values were less than inter-groups values indicating the homogenous and heterogeneous nature of the genotypes within and between the groups (Singh et al., 2013).Both groups initially formed by Tocher's method showed genetic distance between accessions, which is important to qualify the degree of divergence among populations and to evaluate the relative contribution of different components of the total divergence (Bhutia et al., 2015).The BGH-6988 access, the only one in the second group, presented inadequate mass and total fruit production and precocity index.In addition, this access has purplish fruits, a standard not acceptable to the national and international market that prefers green fruit (Solankey et al., 2013).The estimated distances with high magnitude between pairs of individuals within group I justify the establishment of a subgroup, and crosses between accessions of different subgroups could enable restoration of transgressive segregant in advanced generations (Bhutia et al., 2015).
The BGH-740, BGH-693, BGH-7865 and BGH-547 accessions selected as parents made up the same subgroup as the three controls, based on their behavior per se.Accessions similar to the cultivars can be interesting because they could be used as parents in breeding programs, because is faster for the recovery of the agronomic characteristics (Adalid et al., 2012).

Table 1 .
Code name, origin, and year of collection of 70 okra accessions from the Vegetable Germplasm Bank of the Federal University of Viçosa (UFV-BGH) and commercial controls, Viçosa, Minas Gerais State, Brazil.
*Local collection of reference or introduction of the accessions.

Table 2 .
Quantitative and qualitative descriptors to characterize and evaluate morphological and agronomic traits of 70 okra accessions in the vegetative (VS) and fruiting (FS) stages of the Vegetable Germplasm Bank of the Universidade Federal de Viçosa (BGH-UFV), Viçosa, Minas Gerais State, Brazil.

Table 3 .
Means 1 and grades 2 of the sum of inverted positions of four descriptors of economic interest for the accessions and control in the first and second experiments with okra genotypes of the BGH-UFV and three commercial cultivars.

Table 4 .
Summary of variance analysis of the environments (Env.) in relation to the commercial control for the descriptors of economic interest (Desc.) of okra genotypes of the BGH-UFV and three control cultivars.