3.4. Multivariate analysis

3.3.2. Nutritional traits

86 Rediscovery of Landraces as a Resource for the Future

Mineral elements and protein

K 1.00

Ca 0.24\*\* 1.00

P 0.75\*\* 0.34\*\* 1.00

\*\* significant at the 0.01 probability level.

okra genotypes.

Mg 0.75\*\* 0.63\*\* 0.74\*\* 1.00

Na 0.32\*\* 0.35\*\* 0.36\*\* 0.06 1.00

Fe 0.01 0.33\*\* 0.00 0.23\*\* 0.34\*\* 1.00

Al 0.05 0.30\*\* 0.09 0.16 0.29\*\* 0.91\*\* 1.00

B 0.14 0.50\*\* 0.07 0.14 0.36\*\* 0.42\*\* 0.27\*\* 1.00

Zn 0.55\*\* 0.26\*\* 0.73\*\* 0.58\*\* 0.37\*\* 0.14 0.18\*\* 0.001 1.00

Mn 0.46\*\* 0.35\*\* 0.42\*\* 0.45\*\* 0.32\*\* 0.35\*\* 0.34\*\* 0.05 0.33 1.00

Cu 0.37\*\* 0.23\*\* 0.48\*\* 0.38\*\* 0.25\*\* 0.111 0.039 0.110 0.54\*\* 0.27\*\* 1.00 Protein 0.03 0.06 0.05 0.02 0.04 0.26\*\* 0.16 0.064 0.11 0.27\*\* 0.11

Table 4. The correlation coefficients between mineral elements and total protein contents evaluated in immature fruits of

Moderate to highly significant association was observed for the mineral and protein content determined in the okra genotypes (Table 4). Highly to moderate significant positive associations were observed between K and Ca, P, Mg, Zn, Mn, and Cu; while negative and moderate correlation was observed between K and Na. Similarly, there were significantly positive association between Ca and P, Mg, Na, B, Zn, Mn and Cu. It was also observed that there were highly significant associations between P and Mg, Zn, Mn, and Cu. These results suggested that high P content might be accompanied with increased concentration of Mg, Zn, Mn, and Cu contents of okra fruits and vice-versa. Na was negatively and significantly associated with all micronutrients except B indicating that high Na content associated with the low contents of the Fe, Al, Zn, Mn, and Cu. An extremely strong association was observed between Al and Fe (0.91) which might be the indication of the existence of genetic control compared to the rest of the traits evaluated. Strong association was also found between K and P as well as K and Mg. Significantly, negative association was observed between protein and Fe, and between protein and Mn. In general, in the current study most of the traits evaluated showed highly and significantly positive and moderate associations among them indicating that there were some functional interaction existed among the mineral elements and protein content. Ref. [55] reported the positive correlation among the mineral elements in rice was due to the interaction between ions whose chemical properties were sufficiently similar, and they compete for site of absorption, transport, and function in plant tissues. Hence in the present study, positive association between and among the mineral elements and protein contents showed that

Concentration of mineral elements (mg kg<sup>1</sup>

) and total protein content (%) in dry mass basis

K Ca P Mg Na Fe Al B Zn Mn Cu

#### 3.4.1. Morphological phenotypic traits

The principal component analysis (PCA) was used for the reduction of data set and transforming the available raw data set into principal components or component factors, which are equal to the number of evaluated morphological phenotypic traits (Table 5). From the current experiment, the PCA transformed 18 raw set of data into 18 factors loadings or principal components with the pattern that the first principal component (PC1) contributed the most variability and the last principal component (PCn) contributed the lowest variability, which accounted for the entire (100%) variability. However, the PC1, PC2, PC3, and PC4 showed high significant variability compared to the rest of the PCs (Table 5) with the eigenvalues greater than one and cumulatively accounted for 68.49% of the total variation among the okra genotype. These PCs had eigenvalue more than 1 [56, 57]; while the rest of the PCS had eigenvalue less than 1 [58], and would not be considered in the interpretation of the results obtained and removed, as they were not significantly influencing and contributing to the variability among the genotypes. Morphological phenotypic traits showed different pattern of contribution to the variability in the principal components loading suggesting the existence of genetic variability that would be used in the okra improvement programme. The current cumulative variation explained by the first four PCs was comparable with what [59, 60] reported in contributing to the variations among different okra genotypes. In the first principal component, grain yield, number of fruits, fruit diameter, shelled pod weight, thousand seed weight, number of seeds per fruit and fruit length, respectively, contributed high variability with positive loading compared to the rest of the traits. This principal component alone explained 33.10% of the total variability among the okra genotypes with the eigenvalue of 5.96. The PC2 accounted for 16.10% of the total variation and was mainly influenced by vegetative growth traits such as number of branches, stem diameter, leaf width, leaf length, number of internodes and number of leaves with positive loading. The PC3 with 12.72% variance distinguished the okra genotypes based on fruit harvest index, plant height, fruit yield, and internode length with all positive loading except fruit yield. Similarly, the PC4 was associated and dominantly influenced by the number of leaves, leaf length and number of branches with all positive loadings except leaf length and this PC accounted 6.58% variances. The remaining phenotypic traits had no any significant contribution to the variation in the four PCs and hence were of minor importance in the characterization of okra genotypes.

Biplot analysis was carried out based on the first two PCs. The genotypes and morphological phenotypic traits were shown on a biplot to clearly visualize their associations and differences (Figure 1). This PCA biplot more explained the 49.20% of total variability among the genotypes, displaying that number of branches, number of seeds, grain yield, number of fruit, fruit harvest index, and days to 50% flowering were considered as the most discriminating parameters (Figure 1). The genotypes that were positioned on the right top quadrant were closely associated and characterized by longest fruit, largest seed size, heavy fruit shell, highest fruit yield, highest number of seeds per fruit, tallest plant, longest internode, and highest number of


Table 5.

Principal component

 factors,

eigenvalues,

 individual,

 and cumulative

internodes. The genotypes demarcated on the top left quadrant were associated with highest number of branches and leaves, widest stem and leaves, longest leaves as well as late maturing genotypes. Furthermore, the biplot demarcated the genotypes on the left bottom quadrant based on derived traits called fruit harvest index. This trait is the most important trait to select the genotypes for drought tolerance and these traits were suggested to have drought tolerance traits. Similarly, the right bottom quadrant consists of genotypes with highest grain yield, fruits per plant and widest fruits. The genotypes concentrated around the origin had similar genetic characteristics, while the genotypes that were found far from the origin are discriminated from the rest of the group due to their peculiar genes/alleles and considered as unrelated genotypes. Therefore, selection of these genotypes as potential parents would result in successful hybridization to develop heterotic groups in the okra-breeding programme (Figure 3).

Agronomic Performance, Nutritional Phenotyping and Trait Associations of Okra…

http://dx.doi.org/10.5772/intechopen.70813

89

The data set of all the mineral elements and crude protein contents were subjected to principal component analysis (PCA), which removed the highly inter-correlated and redundancy nature of the prevalent variations among the okra genotypes (Table 6). The PCA grouped the mineral

3.4.2. Nutritional traits

Figure 1. Plant characteristics.

 variability for the

morphological

 phenotypic

 traits.

Phenotypic

 traits Factor loadings Agronomic Performance, Nutritional Phenotyping and Trait Associations of Okra… http://dx.doi.org/10.5772/intechopen.70813 89

Figure 1. Plant characteristics.

internodes. The genotypes demarcated on the top left quadrant were associated with highest number of branches and leaves, widest stem and leaves, longest leaves as well as late maturing genotypes. Furthermore, the biplot demarcated the genotypes on the left bottom quadrant based on derived traits called fruit harvest index. This trait is the most important trait to select the genotypes for drought tolerance and these traits were suggested to have drought tolerance traits. Similarly, the right bottom quadrant consists of genotypes with highest grain yield, fruits per plant and widest fruits. The genotypes concentrated around the origin had similar genetic characteristics, while the genotypes that were found far from the origin are discriminated from the rest of the group due to their peculiar genes/alleles and considered as unrelated genotypes. Therefore, selection of these genotypes as potential parents would result in successful hybridization to develop heterotic groups in the okra-breeding programme (Figure 3).

#### 3.4.2. Nutritional traits

Phenotypic

 traits Factor loadings

F1

PH NFP

NB

NL

NI

IL SD

LL

LW

D50%F

FL FD FYLD Grain yield

NSF TSwt

Swt FHI

Eigenvalue

Variability (%)

Cumulative

Table 5.

Principal component

 factors,

eigenvalues,

 individual,

 and cumulative

 variability for the

 %

 33.10

 49.20

 61.91

 68.49

 73.91

 78.42 82.15

 85.63

 88.41

 90.91

 93.09 morphological

 phenotypic

 traits.

 95.08

 96.63

 97.89

 98.62

 99.20

 99.67 100.00

 33.10

 16.10 12.72

 6.58

 5.41

 4.52

 3.73

 3.48

 2.77

 2.50

 2.18

 1.99

 1.55

 1.26

 0.73

 0.58

 0.47

 0.33

 5.96

 2.90

 2.29

 1.18

 0.97

 0.81

 0.67

 0.63

 0.50

 0.45

 0.39

 0.36

 0.28

 0.23

 0.13

 0.10

 0.08

 0.06

0.06

0.12

 0.49

 0.08

0.30

0.43

 0.09

0.19

0.26

0.27

0.14

 0.05

0.11

0.16

0.28

0.28

 0.02

 0.23

0.34

 0.09

0.26

0.02

0.06

 0.03

0.06

0.19

 0.13

0.30

 0.01

0.07

 0.07

 0.03

0.45

0.31

0.37

0.46

0.33

 0.08

 0.12

0.04

0.19

0.28

0.04

0.22

 0.01

0.32

0.13

0.31

 0.22

 0.21

 0.51

 0.30

 0.09

0.18

 0.37 0.32

 0.17

0.07

0.01

 0.12

0.31

0.20

 0.01

0.11

 0.45

 0.04

0.16

 0.43

0.12

0.29

0.10

 0.42

 0.06

0.02

 0.08

 0.11

 0.03

0.06

 0.15

 0.10

 0.09

 0.05

0.18

 0.27

 0.03

 0.06

0.35

 0.64

0.29

 0.25

0.27

 0.11

0.38

0.01

 0.05

 0.20

0.17

0.15

 0.05

0.41

0.04

0.03

0.21

0.08

 0.01

0.09

 0.28

 0.61

0.32

0.35

0.02

0.15

 0.04

 0.08

0.26

 0.04

0.03

0.23

 0.29

 0.16

0.03

0.29

0.38

 0.39

0.15

0.46

 0.09

 0.01

 0.04

0.09

0.24

 0.10

 0.23

 0.51

0.04

 0.14

0.28

0.03

0.06

 0.46

 0.09

0.40

0.01

 0.11

0.19

 0.18

0.29

0.09

 0.41

0.37

0.08

0.02

0.38

0.13

0.31

 0.38

0.05

 0.32

 0.05

0.02

0.01

0.09

0.10

 0.36

0.15

0.25

0.35

0.10

 0.31

0.47

 0.11

 0.34

 0.16

 0.06

0.28

 0.27

0.09

 0.07

 0.07

 0.05

0.07

 0.34

0.07

0.56

0.26

0.01

 0.13

 0.34

0.07

0.21

 0.03

 0.20

 0.25

0.44

 0.05

 0.08

0.03

0.02

0.07

 0.39

 0.17

 0.11

 0.00

0.34

0.42

 0.31

 0.54

0.01

0.01

 0.00

0.35

0.02

 0.01

0.02

0.02

0.03

0.05

0.11

 0.21

 0.35

0.19

0.04

 0.44

0.47

0.28

0.19

 0.23

0.40

 0.14

0.06

0.07

 0.06

0.02

0.11

0.05

 0.32

 0.25

 0.07

 0.50

 0.08

 0.54

0.17

 0.30

0.04

0.26

0.06

 0.10

0.19

 0.09

0.19

 0.03

 0.02

0.11

 0.31

0.10

 0.59

0.30

 0.08

0.03

0.07

 0.01

 0.01

 0.09

 0.35

 0.44

 0.03

 0.19

0.15

0.16

 0.14

0.04

 0.44

0.02

 0.36

0.05

 0.17

 0.13

 0.21

0.48

0.07

0.01

0.38

0.28

0.09

0.16

 0.21

 0.10

0.18

88 Rediscovery of Landraces as a Resource for the Future

0.35

0.10

 0.08

 0.16

0.05

 0.06

 0.10

 0.01

 0.00

0.04

 0.14

 0.56

0.25

0.14

 0.10

 0.00

 0.49

0.40

0.16

 0.22

 0.40

0.15

 0.29

 0.09

0.06

 0.04

0.18

0.16

 0.67

 0.08

 0.06

 0.32

0.02

0.05

0.10

 0.13

 F2

 F3

 F4

 F5

 F6

 F7

 F8

 F9

 F10

 F11

 F12

 F13

 F14

 F15

 F16

 F17

 F18

> The data set of all the mineral elements and crude protein contents were subjected to principal component analysis (PCA), which removed the highly inter-correlated and redundancy nature of the prevalent variations among the okra genotypes (Table 6). The PCA grouped the mineral


Table 6. Principal component analysis of mineral elements and protein traits in 46 okra genotypes showing eigenvectors, eigenvalues, individual, and cumulative percentage of variation explained by the first three PC axes.

elements and protein traits into 12 components, which accounted for the entire (100%) genetic variability among the evaluated okra genotypes. According to Chatfied and Collins [58, 61], components with an eigenvalue of less than one should be removed so that fewer components with significant meanings are considered. Furthermore, Ref. [56] suggested that eigenvalues greater than one are considered significant and component loadings greater than 0.3 were considered meaningful. Hence, from this study, as it can be seen clearly that only the first three eigenvectors which had eigenvalues greater than one and cumulatively explained about 69.04% of the total variation by the first, second and third principal components in the whole data set for the genotypes and provide discriminatory information in respective to the mineral elements and protein. The first principal component, that is the PC1 alone describes and explains 33.69% of the total variability among the okra genotypes, which was mainly contributed by the variances due to K, P, Mg, Zn, Mn and Cu (Table 6) with positive loading. The second principal component

Figure 3. Scatter plot of the first and second principal component analysis for morphological phenotypic traits and okra

Agronomic Performance, Nutritional Phenotyping and Trait Associations of Okra…

http://dx.doi.org/10.5772/intechopen.70813

91

genotypes.

Figure 2. Fruit characteristics.

Agronomic Performance, Nutritional Phenotyping and Trait Associations of Okra… http://dx.doi.org/10.5772/intechopen.70813 91

Mineral elements and protein

Variability (%)

Cumulative %

Figure 2. Fruit characteristics.

Factor loadings

90 Rediscovery of Landraces as a Resource for the Future

PC1 PC2 PC3 PC4 PC5 PC6 PC7 PC8 PC9 PC10 PC11 PC12

33.693 23.893 11.454 7.455 6.449 5.359 4.508 2.560 1.669 1.540 0.993 0.426

33.693 57.586 69.040 76.495 82.944 88.303 92.811 95.371 97.040 98.581 99.574 100.000

Table 6. Principal component analysis of mineral elements and protein traits in 46 okra genotypes showing eigenvectors,

eigenvalues, individual, and cumulative percentage of variation explained by the first three PC axes.

K 0.410 0.037 0.199 0.128 0.320 0.102 0.181 0.597 0.056 0.201 0.479 0.017 Ca 0.215 0.392 0.399 0.089 0.039 0.144 0.196 0.392 0.394 0.245 0.424 0.155 P 0.441 0.048 0.176 0.080 0.028 0.002 0.214 0.133 0.380 0.725 0.017 0.176 Mg 0.407 0.250 0.003 0.115 0.176 0.142 0.130 0.017 0.082 0.389 0.726 0.073 Na 0.208 0.322 0.321 0.106 0.095 0.669 0.268 0.130 0.339 0.268 0.009 0.081 Fe 0.070 0.522 0.257 0.186 0.168 0.166 0.010 0.023 0.131 0.218 0.087 0.703 Al 0.089 0.475 0.257 0.388 0.229 0.080 0.179 0.133 0.151 0.032 0.126 0.633 B 0.035 0.376 0.349 0.230 0.364 0.579 0.048 0.419 0.023 0.050 0.061 0.176 Zn 0.402 0.021 0.057 0.113 0.394 0.171 0.389 0.375 0.546 0.143 0.160 0.038 Mn 0.325 0.124 0.368 0.172 0.376 0.123 0.506 0.063 0.467 0.277 0.055 0.001 Cu 0.313 0.020 0.056 0.453 0.520 0.270 0.481 0.324 0.027 0.027 0.088 0.062 Protein 0.058 0.146 0.521 0.677 0.276 0.121 0.348 0.106 0.132 0.043 0.031 0.041 Eigenvalue 4.043 2.867 1.374 0.895 0.774 0.643 0.541 0.307 0.200 0.185 0.119 0.051

Figure 3. Scatter plot of the first and second principal component analysis for morphological phenotypic traits and okra genotypes.

elements and protein traits into 12 components, which accounted for the entire (100%) genetic variability among the evaluated okra genotypes. According to Chatfied and Collins [58, 61], components with an eigenvalue of less than one should be removed so that fewer components with significant meanings are considered. Furthermore, Ref. [56] suggested that eigenvalues greater than one are considered significant and component loadings greater than 0.3 were considered meaningful. Hence, from this study, as it can be seen clearly that only the first three eigenvectors which had eigenvalues greater than one and cumulatively explained about 69.04% of the total variation by the first, second and third principal components in the whole data set for the genotypes and provide discriminatory information in respective to the mineral elements and protein. The first principal component, that is the PC1 alone describes and explains 33.69% of the total variability among the okra genotypes, which was mainly contributed by the variances due to K, P, Mg, Zn, Mn and Cu (Table 6) with positive loading. The second principal component

distinct [54]. The okra genotypes in the top right quadrant were closely associated with the mineral elements such as Al, Fe, Cu, Zn, and Mn (Figure 2). The right bottom quadrant consists of the okra genotypes that are closely related with the mineral elements such as K, Mg, P, and Ca. Those genotypes that found on the left bottom quadrant were mostly associated with the low concentration of mineral elements such as Na and B and protein content. In the present study, the genotypes VI056457, VI033796, VI060824, VI060802, VI055423, and VI049632 stand out clearly as the most genetically divergent okra genotypes for the nutritional traits evaluated. This indicated that they might have a peculiar gene/allele that separated them from the group of the genotypes assessed for the nutritional composition and could be used as parental genotypes for hybridization to develop new cultivar for

Agronomic Performance, Nutritional Phenotyping and Trait Associations of Okra…

http://dx.doi.org/10.5772/intechopen.70813

93

In the present study, the existence of genetic variability in the morphological, phenotypic and nutritional traits would help the breeder in selection of the okra genotypes for the improvement for these traits, which would help to increase the frequency of favorable genes in the pre-breeding programme. This is a first pre-breeding programme of okra in South Africa established recently as a prerequisite for the development of new cultivar in the country and beyond for yield and nutritional quality. The okra genotypes in this study showed enormous phenotypic and nutritional variations that would help in the okra improvement programme. The significant positive association between grain yield and yield traits as well as nutritional quality traits could be used as selection criteria for potential and good parental lines in okra breeding programme in South Africa. Understanding and the knowledge of variability and trait association in this study is important in the okra-breeding programme as an initial step to develop new cultivar for the traits of interest. To my best knowledge, this is the first study on this under-utilized fruit vegetable crop

species in South Africa that would contribute to food, nutritional and health security.

Agricultural Research Council (ARC) and National Research Foundation (NRF) for research and funding opportunity, South Africa. The author would also like to acknowledge the Genetic Resources and Seed Unit, World Vegetable Center (AVRDC), Taiwan for providing the okra

Agricultural Research Council, Vegetable and Ornamental Plants, Pretoria, South Africa

the traits of interest in our breeding programme (Figure 4).

4. Conclusion

Acknowledgements

germplasm for the study.

Author details

Abe Shegro Gerrano

Address all correspondence to: agerrano@arc.agric.za

Figure 4. Biplot generated using the concentration of mineral elements and protein content data set of okra genotypes.

(PC2) represents and accounted for 23.89% with eigenvalue of 2.87 had dominantly influenced by the mineral elements such as Ca, Na, Fe, Al, and B with the highest loading vector of Fe followed by Al with positive loading. The nutritional trait that contributed great variability among the genotypes showing 11.45% of variation were protein with the highest positive loading. The mineral elements Ca, Na, B, and Mn also contributed differences in this PC.

The existence of wider nutritional variability among okra genotypes studied was further described by the PCA biplot (Figure 2) using multivariate technique. The PCA biplot provided important information regarding the similarities as well as the pattern of differences among the nutritional traits of the different okra genotypes and of the interrelationships between the quantified nutritional traits. The PCA clustered the okra genotypes into different groups over the four quadrants based on the nutritional traits determined (Figure 2). The okra genotypes scattered in all four quadrants on the axes, indicating that there were a wide genetic variability for the traits studied. Accessions that overlapped and closer to each other in the principal component axes had similar genetic relationships in the nutritional traits. However, genotypes which are far from each other could be considered as genetically distinct [54]. The okra genotypes in the top right quadrant were closely associated with the mineral elements such as Al, Fe, Cu, Zn, and Mn (Figure 2). The right bottom quadrant consists of the okra genotypes that are closely related with the mineral elements such as K, Mg, P, and Ca. Those genotypes that found on the left bottom quadrant were mostly associated with the low concentration of mineral elements such as Na and B and protein content. In the present study, the genotypes VI056457, VI033796, VI060824, VI060802, VI055423, and VI049632 stand out clearly as the most genetically divergent okra genotypes for the nutritional traits evaluated. This indicated that they might have a peculiar gene/allele that separated them from the group of the genotypes assessed for the nutritional composition and could be used as parental genotypes for hybridization to develop new cultivar for the traits of interest in our breeding programme (Figure 4).
