Combining performance and estimated genetic diversity among soybean parents and F1 populations

ABSTRACT Studies about the combining ability and genetic diversity can provide information on the genetic resources contained in the working collection of germplasm. The objective of this study was to estimate the combining ability of ten soybean parents and their hybrids in the F1 generation using diallel analysis and genetic diversity analysis to evaluate the dissimilarity between parents and crosses. The experiment was carried out in a greenhouse, in a completely randomized design with five replications, during the 2019/2020 crop season, in the municipality of Rio Verde, State of Goiás, Brazil. Agronomic and physiological variables were evaluated. Data were subjected to analysis of variance. Diallel analysis was performed using the Griffing method 2. Analysis of genetic diversity was performed with the Mahalanobis generalized distance and the hierarchical average linkage clustering. Additive gene effects were predominant in the control of all variables analyzed in the F1 generation. The parents SYN 1163 RR and CG 7370 RR showed good complementarity for the variable grain yield. The thirty-five genotypes were separated into six groups, and the variable weight of one hundred seeds had the greatest contribution to genetic diversity.


INTRODUCTION
Soybean breeding programs require a period of six to eight years for the development of cultivars, consisting of various stages such as hybridizations, conduction of segregating populations, and productivity tests.Given the time required for a breeding cycle, choosing the parents and the design that will be used in the interbreeding are determinant steps for success (Daronch et al., 2014).Therefore, the selection of parents is one of the main steps in plant breeding programs (Bilyeu et al., 2016).
After the selection of parents, hybridizations are carried out to obtain segregating populations with maximum genet-ic variability and a high potential for extraction of superior lines during the selection process.In this context, the prior assessment of the genetic potential of the parents allows for the expansion of genetic variability and, consequently, greater genetic gains can be achieved (Rocha et al., 2019).
Aiming at increasing the efficiency of breeding programs, biometric techniques such as diallel analysis were developed to help researchers when choosing parents.
Diallel analysis allows the prediction of the potential of populations as well as inference about the combining ability of soybean parents (Umar et al., 2017).

Samuel Leandro Soares et al.
Additionally, it is important in a breeding program to infer about the genetic diversity between parents and segregating populations, aiming at the maintenance of alleles that express significant agronomic characteristics for the crop.Studies on genetic diversity in breeding processes allow the germplasm bank systemization, assistance in a core collection generation, and especially the maintenance of genetic resources (Sousa et al., 2015).
Among the methods to assess genetic diversity, the generalized Mahalanobis distance (D2) allows analyzing data with repetitions and can combine hierarchical and non-hierarchical clustering methods (Cruz et al., 2020).The UP-GMA hierarchical method calculates the distance between two clusters as the average distance between all pairs of items of two clusters (Oliveira et al., 2017).Methods that group the genotypes based on some similarity or dissimilarity measures aim to separate an original group observed into several subgroups, in a way to obtain intra-subgroup homogeneity and inter-subgroup heterogeneity (Rodrigues et al., 2017).
The objective of this study was to evaluate the effects of the combining ability of ten soybean parents and their hybrid combinations in the F 1 generation in partial diallel crosses and to evaluate the genetic diversity between parents and crosses using the method of the generalized distance of Mahalanobis and the hierarchical clustering method (UPGMA).

Plant material and hybridizations
Partial diallel crosses were carried out with no reciprocals involving ten soybean cultivars, in February 2019.The cultivars were divided into two parental groups (Table 1), and the purple-and white-flower parents were allocated to groups I and II, respectively.Hybridizations were conducted in a greenhouse, in which ten parent seeds were sown in 12-dm 3 pots, in the municipality of Rio Verde, Goiás, Brazil.To ensure flowering synchronism, staggered sowing was done every five days in the four seasons.
Hybridizations were performed mainly with white-flower cultivars as pollen receptor parents.This strategy was adopted to eliminate plants from self-fertilization in the F 1 generation, since the purple flower has complete dominance over the white flower, that is, in the F 1 generation, all crossed plants must present the purple flower.

Experimental design and procedures
An experiment was conducted with the F 1 generation in a completely randomized design, with thirty-five treatments, in which twenty-five were hybrid combinations and ten were parents, with five replications.Each experimental unit consisted of a pot with two plants, and the value of each treatment was represented by the mean of the two plants.
The treatments were sown in 12-dm 3 plastic pots, in greenhouses, on December 12, 2019, in the municipality of Rio Verde, Goiás (17°44'24.9"S and 50°57' 31.4"W, 853 m altitude).To carry out the sowing, two holes three cm deep were made on opposite sides of the pot, and two seeds were planted per hole.To ensure equal number of plants per treatment, thinning was performed at five days after the emergence of the seedlings, leaving one plant per hole.
During the experiment, daily irrigation was applied for providing uninterrupted supply of the water demanded by the soybean crop and preventing the occurrence of water restrictions.Pests and diseases management was carried out following the technical recommendations for the soybean crop.

Analyzed variables
The variables analyzed were: days to flowering -stage R1 (DTF): plant height at flowering (PHF); days to maturity -stage R8 (DTM); and plant height at maturity (PHM).Evaluation of plant development was based on the phenological scales proposed by Fehr & Caviness (1977).
Other variables evaluated included first pod insertion height (FPIH); number of nodes per plant (NNP), which was used only for analysis of genetic diversity,; the number of pods per plant (NPP); "a" and "b" chlorophyll content (TCab) measured with a Clorofilog Falker 1030 chlorophyll meter, at thirty days after seedling emergence, in the third fully expanded trifoliolate leaf from the apex to the base of the plant, always in the central leaflet as described by Yokoyama et al. (2018); weight of one hundred seeds (WHS); and grain yield expressed as grams plant -1 (GY).
Combining performance and estimated genetic diversity among soybean parents and F1 populations

Statistical-genetic analyses
The analysis of variance was used to evaluate the effect of the treatments.Then, the Griffing's (1956) diallel analysis, model 2, was applied to the variables with significant effect, with adaptation for partial diallel performed by Geraldi & Miranda Filho (1988), according to the model: Where, Y ij : mean of the cross between the i th parent of group I and j th parent of group II; µ: general mean of the diallel; d1 + d2: differences between means of groups I and II and the general mean; g i : effect of the general combining ability of the i th parent from group I; g j : general combining ability of the j th parent from group II; S ij : effect of specific combining ability; Ɛ: experimental mean error.
The variables with significant effect (p < 0.05) were subjected to genetic diversity analysis.The distances between Samuel Leandro Soares et al.
pairs of genotypes were calculated by the Mahalanobis generalized distance matrix (D2) defined as , and the clustering of genotypes was performed using the hierarchical method of between-group average linkage (UPGMA).
Where: D 2 ii' : generalized Mahalanobis distance between i and i' genotypes; ψ: matrix of residual variances and covariances; δ': [d1 d2 ...dv] where dj Yij -Yi´j and Yij: mean of the i th genotype in relation to the j th variable.
The dendrogram shows the hierarchical clustering of the genotypes.The efficiency of the dendrogram in reproducing the dissimilarity matrix (D2) was calculated by the cophenetic correlation coefficient (CCC).The dendrogram cutoff point was applied according to Mojena (1977).To verify the contribution of each variable to genetic diversity, the principal component analysis was applied as described by Cruz et al. (2012).The variance and diallel analyses were performed using the Genes statistical software (Cruz, 2016), and the Rbio statistical software was used for the other analyses (Bhering, 2017).

RESULTS AND DISCUSSION
The analysis of variance detected a significant effect (p < 0.01) of the source of variation and treatments for all variables analyzed (Table 2), indicating the presence of genetic variability between parents and F 1 hybrids.
Table 2: Mean square of the partial diallel variance analysis for days to flowering (DTF), plant height at flowering (PHF), days to maturity (DTM), plant height at maturity (PHM), number of nodes per plant (NNP), number of pods per plant (NPP), "a" and "b" chlorophyll content (TCab), weight of one hundred seeds (WHS), and grain yield as g plant -1 (GY), evaluated in the F 1 progenies derived from a set of diallel crosses involving ten soybean cultivars.Rio Verde -GO, 2020 The general combining ability (GCA I) in Group-I was significant (p < 0.05) for all variables, except for the weight of 100 seeds (Table 2).In Group-II, the general combining ability (GCA II) was significant (p < 0.05) for all variables, except for tdays to flowering and "a" and "b" chlorophyll content (Table 2), suggesting that among the parents in Group-II there was no significant difference in the allelic concentration for the two traits.In cases where GCA is significant, it can be inferred that at least one of the parents differs from the others regarding the concentration of favorable alleles (Pimentel et al., 2013).The source variation SCA showed significant effect for all variables analyzed (Table 2).
The GCA/SCA ratio with the highest GCA estimates indicates a predominance of additive gene effects in controlling this characteristic.When additive effects are pronounced, gains of greater magnitude will be predicted (Bhering et al., 2017).On the other hand, the highest SCA values indicate a predominance of non-additive gene effects in controlling these traits (Bi et al., 2015).
The additive gene effects were predominant for all the characteristics assessed, with the lowest GCA/SCA ratio found for grain yield (1.3) and the highest GCA/SCA ratio for plant height at flowering (18.4).Rocha et al. (2019) demonstrated the importance of additive gene effects in controlling the characteristic grain yield when evaluating Furthermore, the complexity of inheritance of quantitative traits and the effect of the genotype x environment interaction may explain the divergence of the results obtained (Goksoy et al., 2019).
Because precocity is one of the primary objectives of breeding programs, parents that contribute to the reduction of the cycle must be inserted in new crossing blocks.Thus, the parents NS 4823RR (DTF: -1.7 and DTM: -4.0) and FMT Anta 82 RR (DTF: -0.5 and DTM: -0.4) presented negative GCA estimates or both days to flowering and days to maturity (Table 3).3).This is relevant information for soybean improvement, because the number of pods per plant is the variable that contributes the most to the increase in grain yield (Mahbud et al., 2015;Bisinotto et al., 2017).Considering the variables APF, APM, NN and NVP simultaneously, as well as the heritability of a character with an additive gene effect (Ramalho et al., 2012), the parent CG 7370 RR showed to be promising to be included in new crossing blocks.
The variable "a" and "b" chlorophyll content, among the parents of group-I had positive and high GCA (1.3) effects for the cultivar NS 4823 RR (1.3).Group II showed no significant effect for GCA, and among the parents, CG Samuel Leandro Soares et al.
The variable weight of one hundred seeds had no significant effect for GCA in group-I.For the parents in group-II, the cultivars BMX Desafio RR and TMG 7262 RR contributed the most to increasing the weight of one hundred seeds in their offspring, with values of 0.5 and 0.8, respectively (Table 3).
Among the parents that contributed the most to grain yield, the cultivars SYN 1163 RR and CG 7370 RR stood out with positive GCA estimates of 2.9 and 4.8, respectively (Table 3).Grain yield is a characteristic strongly influenced by the environmental effect as it is a quantitative trait, thus, parents that contribute with allelic frequency to simultaneously increase several traits should be used in breeding programs aiming at higher yields.In this aspect, the parent CG 7370 RR stood out for presenting estimates that contribute to all variables, even with the lowest-magnitude estimates for days to flowering and maturity and weight of one hundred seeds.
Estimates of specific combining ability are shown in  4).
Later in this section, we discuss that the variable number of pods per plant makes a considerable contribution to the genetic divergence found in this study, which infers the existence of genetic variability for this attribute, providing greater selection gains.According to Bagateli et al. (2020), positive estimates of SCA reveal the existence of a large amount of heterozygous loci with greater potential for genetic variability, in addition to being associated with high means.
Except for PRE 6336 RR x BRSGO 6959 RR, all the other crosses have at least one parent with a good GCA estimate in their genealogy.In cases where the parents do not present a good GCA, the population potential can be predicted based only on the SCA, indicating superiority of parents based only on their GCA, which infers the existence of gene complementation between the parents.x FMT Anta 82 RR, indicating it is superior to the parents.
In addition, crosses with the highest SCA estimates can release greater variability after self-fertilization generations, essential for plant breeding (Abreu et al., 2004).Combining performance and estimated genetic diversity among soybean parents and F1 populations In relation to the importance of the variables for genetic diversity, this study found that the weight of one hundred seeds had the greatest contribution (13.3%), followed by number of nodes per plant (11.2%), and number of pods per plant (11.0%) (Figure 2).were plant height at maturity, number of pods with one seed, and grain yield, with a relative importance of 31.0,19.5, and 13.2%, respectively.Torres et al. (2015), in a similar study, found that number of pods per plant was the trait that contributed the most to genetic diversity.
Moreover, the variables that contributed the least to genetic diversity were grain yield, days to maturity, and "a" and "b" chlorophyll content, with values of 8.9%, 8.8%, and 7.4%, respectively.The finding that the variable grain yield

CONCLUSIONS
The additive gene effects were the most important for the expression of the variables analyzed.
Overall, the parent CG 7370 RR is the most promising to be used in new crossing blocks, as it simultaneously contributed to enhancement of the variables analyzed.
Six groups of genotypes were formed based on the UP-GMA cluster analysis, and the variables that presented the greatest relative contribution to genetic diversity were the weight of one hundred seeds, number of nodes per plant, and number of pods per plant.
The generalized Mahalanobis distance matrix was used to confirm the occurrence of genetic diversity of great magnitude among the treatments analyzed.

F
2 populations from diallel crosses.Similar results were also obtained byZorzetto et al. (2009), who verified the Combining performance and estimated genetic diversity among soybean parents and F1 populations importance of additive gene effects in the expression of the characteristics number of seeds per plant, number of pods per plant, and weight of 100-seeds.In contrast,Colombo et al. (2018) obtained a greater contribution of non-additive gene effects in the F 1 generation for the expression of number of days to flowering and maturity, and grain yield.The diverged results may be related to the population of interest, since estimates of genetic parameters are intrinsic to the target population.

Figure 1
Figure 1 shows the dendrogram of the dissimilarity between the genotypes by the hierarchical average linkage clustering (UPGMA).The cophenetic correlation coefficient (CCC) estimated was 0.74.The cophenetic correlation coefficient indicates the efficiency of the dendrogram in
This result has great relevance to the genetic improvement of the crop, because the genetic diversity estimates can be used to guide new crossings and obtain new gains with selection within populations using a greater genetic variability.Sousa et al. (2015) evaluated the genetic diversity among soybean crosses and found that the most contributing variables to genetic diversity among the crosses studied did not contribute much to genetic diversity of soybean in this study corroborates the results obtained by Oliveira et al. (2017) in a study of genetic divergence between 24 soybean genotypes using multivariate clustering methods.

Figure 2 :
Figure 2: Relative contribution of 10 characteristics to quantify genetic diversity among 35 soybean genotypes: weight of one hundred seeds (WHS), number of nodes per plant (NNP), number of pods per plant (NPP), first pod insertion height (FPIH)), plant height at maturity (PHM), days to flowering (DTF), plant height at flowering (PHF), grain yield in g plant -1 (GY), days to maturity (DTM) and " a and b" chlorophyll content (TCab).

Table 1 :
Description of soybean parents and hybrids used in a partial diallel crossing design (5x5) *ID: Identification code

Table 3 :
Estimates of general combining ability for: days to flowering (DTF), plant height at flowering (PHF), days to maturity (DTM), plant height at maturity (PHM), number of nodes per plant (NNP), number of pods per plant (NPP), "a" and "b" chlorophyll content (TCab), weight of one hundred seeds (WHS), and grain yield as g plant -1 (GY), relative to ten soybean parents in the municipality of Rio Verde, Goiás

Table 4 .
The crosses that showed the lowest SCA estimates

Table 4 :
Estimates of specific combining ability for: days to flowering (DTF), plant height at flowering (PHF), days to maturity (DTM), plant height at maturity (PHM), number of nodes per plant (NNP), number of pods per plant (NPP), "a" and "b" chlorophyll content (TCab), weight of one hundred seeds (WHS), and grain yield in g plant -1 (GY), relative to twenty-five F 1 soybean populations in the municipality of Rio Verde, Goiás