The ANOVA results for the measured traits are displayed in Table 2. The analysis indicated that there is significant variation among the sugar beet populations for gross sugar content, white sugar content, white sugar yield, extraction coefficient, and root dry matter (P < 0.01). Additionally, significant variation was observed for roots number, root weight, the amount of sodium, the amount of potassium, and the average length (P < 0.05). The existence of notable variations among population implies their distinctive genetic diversity.
Table 2
Analysis of variance for the measured traits in sugar beets genotypes
S.O. V | df | RN | RY | LY | SC | Na | K | N | WSC | WSY | ECS | MC | DM | L | D |
Rep | 4 | 1.91 | 1.26 | 0.68 | 2.72 | 0.2 | 0.55 | 0.07 | 5.48 | 0.08 | 59.64 | 0.58 | 3.22 | 4.75 | 1.56 |
Pop | 7 | 3.67* | 1.57* | 0.63 ns | 5.28** | 0.08* | 1.36* | 0.06 ns | 8.27** | 0.2** | 66.9** | 0.42 ns | 4.67** | 12.49* | 1.48 ns |
error | 28 | 1.52 | 0.47 | 0.31 | 0.61 | 0.03 | 0.51 | 0.03 | 0.74 | 0.05 | 7.93 | 0.12 | 0.86 | 4.81 | 1.55 |
C.V% | | 15.82 | 11.81 | 15.12 | 5.10 | 8.61 | 9.99 | 9.45 | 7.96 | 12.20 | 4.00 | 9.05 | 3.79 | 8.75 | 12.72 |
*, ** and ns: Significant at 5%, 1%, and non-significant respectively |
Root number (RN), root weight (RY) (tons/ha), leaf weight (tons/ha) (LY), gross sugar content (SC), sodium (Na), potassium (K), nitrogen (N), white sugar content (WSC), white Sugar yield (WSY), extraction coefficient (ECS), molasses content (MC), root dry matter (DM), the average of length (L) and the average of diameter (D).
Based on the results of Duncan’s Multiple Range Test (Table 3), it has been proven that M224 has produced the highest RN, RY, LY, and WSY compared to other populations with a mean of 81.1, 46 (tons/ha), 16 (tons/ha), and 8.4, respectively. On the other hand, the P17 had the lowest RN with 45.4, 7233.p29 had the lowest RY with 25 ton/h, 181 had the lowest LY with 10.2 ton/h, and P17 had the lowest WSY with 4 ton/h. A statistically significant difference (P < 0.05) was found between P191 and other populations for SC, N, WSC, ESC, and DM traits. The means for these traits were 17.3, 3.7, 13.2, 76.5, and 26.3, in P 191, respectively. Among these traits, P17 had the lowest amounts of SC (13.8%), N (2.7), and WSC (8.8%). For ESC, M249 had the lowest amount at 69.1%, while 7233.p29 had the lowest amount of DM at 24.8%. According to the results reported in Table 3, P.17 had the highest amount of MC and K with 4.4% and 8.11, respectively, while the lowest amount for these traits was observed in P191 (3.4% and 6.4, respectively. On the other hand, showed the highest amounts of average length (L) and diameter (D) of sugar beet (27.4 and 132.3 respectively). Also, P.5 had the lowest amount of average length and diameter of sugar beet.
In total, population 191 displayed better traits in terms of sugar content, nitrogen content, pure sugar content, extraction coefficient, and root dry matter. Similarly, population M224 exhibited better root number, root yield, leaf weight, and sugar yield as compared to other populations. However, in population 191, the levels of sodium, potassium, and molasses were found to be low, indicating a negative relationship between these traits and root performance, as well as sugar content. According to the data, population P.17 had the lowest values for traits such as root number, sugar content, nitrogen content, pure sugar content, sugar yield, extraction coefficient, and root dry matter. In contrast, population P.17 had the highest values for sodium, potassium, and molasses content. These traits are considered unfavorable as they negatively impact the percentage of sugar extraction. Smith and Martin (1989) found that selecting sugar beet roots with lower sodium ion concentrations increases syrup purity and extractable sugar content, which is consistent with our study findings.
In both the M224 and P.5 populations, the leaf weight trait has the highest value. Also, M224 population, has the highest values in root yield and sugar yield. This indicates that the leaves have a positive impact on sugar yield. In a study conducted by Hajimobina et al. (2011), it was found that the yield has a positive and significant correlation with the leaves area and biomass. Additionally, the number of leaves has a positive and significant correlation with gross sugar due to its direct role in photosynthesis.
The results of this study showed that there is not much diversity among the studied germplasm in the range of low changes in certain traits, such as sugar content, sodium, potassium, nitrogen, molasses content, root dry matter, and average length. However, the range of high changes in some traits such as root number, root yield, leaf weight, pure sugar content, sugar yield, extraction coefficient, and average diameter indicates that there is considerable potential for improvement by selection in these traits among the studied populations.
Table 3
Mean comparison of sugar beets genotypes based of measured traits
Pop | RN | RY | LY | SC | Na | K | N | WSC | WSY | ECS | MC | DM | L | D |
181 | 59.8a-d | 35.1ab | 10.2c | 16.2b | 3.8bc | 7.0bc | 3.3ab | 11.8b | 7.0ab | 73.1ab | 3.7bcd | 25.0b | 23.6bc | 132.3a |
7233.p29 | 49.0cd | 25.0c | 12.6abc | 15.2bc | 3.8bc | 6.7bc | 3.7a | 10.9bc | 5.5c | 71.6c | 3.6cd | 24.8b | 24.0bc | 88.3b |
BP-Mashhad | 52.6bcd | 34.2bc | 14.1abc | 15.0c | 4.2abc | 7.2bc | 3.0ab | 10.5c | 5.6bc | 70.0bc | 3.9bcd | 24.0bc | 26.4ab | 97.0ab |
M249 | 72.8abc | 39.2ab | 15.7ab | 14.9c | 4.6ab | 6.9bc | 3.4ab | 10.3c | 7.3ab | 69.1c | 3.9abc | 24.2b | 27.4a | 91.3ab |
191 | 66.2a-d | 30.5bc | 10.8bc | 17.3a | 3.5c | 6.4c | 3.7a | 13.2a | 8.7ab | 76.5a | 3.4d | 26.3a | 24.8abc | 92.1ab |
M224 | 81.1a | 46.0a | 16.0a | 15.1c | 4.1abc | 7.6ab | 3.3ab | 10.5c | 8.4a | 69.4c | 4.0abc | 23.9bc | 26.2ab | 92.5ab |
P.5 | 73.4ab | 37.8ab | 16.7a | 15.2bc | 4.9a | 7.3abc | 2.7b | 10.5c | 7.6ab | 68.8c | 4.1ab | 24.2b | 22.7c | 90.9ab |
P.17 | 45.4d | 29.7bc | 10.9bc | 13.8d | 4.7a | 8.1a | 2.7b | 8.8d | 4.0c | 63.9c | 4.4a | 23.0c | 25.1abc | 93.4ab |
Populations with at least one similar letter do not show significant differences according to the LSD test (P ≤ 0.05).
Root number (RN), root weight (RY) (tons/ha), leaf weight (LY) (tons/ha), gross sugar content (SC) (%), sodium (Na), potassium (K), nitrogen (N), white sugar content (WSC), white Sugar yield (WSY) (tons/hectare), extraction coefficient (ECS) (%), molasses content (MC) (%), root dry matter (DM) (%), the average of length (L) and the average of diameter (D).
Correlation coefficient
The correlation coefficient among the measured traits in this study is shown in Table 4. The results indicated that the number of roots has a significant effect on RY, LY, Na, and WSY, but does not have a significant correlation with other traits. RY correlated with LY, Na, and WSY, while LY is inversely correlated with WSC, ECS, and DM, and directly correlated with Na. Liu et al. (2022) also find an extremely significantly positive relationship between the root yield and sugar yield, potassium content and α-N content.
this experiment showed that SC affected significantly by N, WSC, WSY, ESC, and DM, and is also inversely correlated with Na, K, and MC. Na concentrations decreased with higher amounts of WSC, ECS, and DM, and increased with higher amounts of MC and L. Significant correlations were observed between K and MC, and negative correlation with WSC, ESC, and DM. The amount of N was only significantly affected by DM.
The relative decrease of WSC was strongly correlated with MC and directly affected by WSY, ESC, and DM. No significant correlations were observed between WSY and MC, L, and D in this study. However, WSY was directly affected by DM and ESC. It seems that although further enhancement of ESC in DM is possible, there is a significant decrease in MC.
Correlation between traits is crucial in plant breeding. One-sided selection for traits without considering their correlation may lead to unfavorable results in a breeding program. Based on the findings presented in Table 4, it can be inferred that leaf weight had a negative and significant impact on both the pure sugar content and extraction coefficient. This implies that as the leaf weight increases, the pure sugar content and extraction coefficient decrease. It was also observed that pure sugar content exhibits a negative and highly significant correlation with sodium content, potassium content, and molasses content. This suggests that an increase in the aforementioned attributes leads to a decrease in pure sugar content. Moreover, when the impurities of sodium, potassium, and nitrogen increase, more extractable sugar enters the molasses sugar, resulting in a decline in the extraction coefficient. On the other hand, the extraction coefficient displays a positive and significant correlation with sugar content and pure sugar content, and a negative and significant correlation with sodium, potassium, and molasses content.
The results of this study revealed that there is a positive and significant correlation between leaves yield, sodium content, sugar yield, and molasses content. This means that these traits will be increased as the weight of the leaf increased. However, an increase in leaf weight negatively affects the increase in sugar content in the root. Despite this, sugar yield still increases due to the high effect of leaf weight on root performance. A previous study by Watson et al. (1972) reported that the production of large leaves in sugar beet plants leads to a decrease in root storage materials due to the increase in respiration compared to photosynthesis. This result is consistent with our findings. Ober et al. (2005) concluded there was a high positive correlation between transpiration rate and sugar yield. The results suggest that green foliage cover and water use patterns could help identify superior genotypes in elite germplasm. These data should enable tools to be developed for the indirect selection of genotypes. Chołuj et al. (2014), suggest that traits that are strongly correlated with root and sugar yield could be used as potential selection criteria in physiological-associated breeding strategies to improve drought tolerance in sugar beet.
Table 4
Correlation (simple matching similarity coefficient) analysis of sugar beet genotypes
| RN | RY | LY | SC | Na | K | N | WSC | WSY | ECS | MC | DM | L | D |
RN | 1.00 | | | | | | | | | | | | | |
RY | 0.82** | 1.00 | | | | | | | | | | | | |
LY | 0.78** | 0.81** | 1.00 | | | | | | | | | | | |
SC | 0.11 | -0.11 | -0.26 | 1.00 | | | | | | | | | | |
Na | 0.37** | 0.45** | 0.57** | -0.58** | 1.00 | | | | | | | | | |
K | -0.03 | 0.06 | 0.06 | -0.33* | 0.13 | 1.00 | | | | | | | | |
N | 0.10 | 0.06 | 0.10 | 0.30* | 0.00 | 0.00 | 1.00 | | | | | | | |
WSC | 0.02 | -0.20 | -0.35* | 0.97** | -0.70** | -0.47** | 0.20 | 1.00 | | | | | | |
WSY | 0.91** | 0.66** | 0.55** | 0.50** | 0.05 | -0.22 | 0.19 | 0.42** | 1.00 | | | | | |
ECS | -0.09 | -0.26 | -0.39** | 0.85** | -0.78** | -0.61** | 0.07 | 0.95** | 0.30* | 1.00 | | | | |
MC | 0.26 | 0.373 | 0.45** | -0.57** | 0.80** | 0.69** | 0.14 | -0.75** | -0.07 | -0.91** | 1.00 | | | |
DM | 0.01 | -0.23 | -0.30** | 0.91** | -0.50** | -0.23 | 0.46** | 0.86** | 0.35* | 0.73** | -0.43** | 1.00 | | |
L | 0.07 | 0.09 | 0.12 | -0.28 | 0.28* | -0.16 | -0.03 | -0.25 | -0.03 | -0.20 | 0.10 | -0.28* | 1.00 | |
D | -0.03 | 0.09 | -0.05 | 0.04 | -0.06 | 0.00 | -0.12 | 0.05 | -0.01 | 0.07 | -0.06 | -0.01 | -0.06 | 1.00 |
*, ** and ns: Significant at 5%, 1%, and non-significant respectively |
Root number (RN), root weight (RY), leaf weight (LY), gross sugar content (SC), sodium (Na), potassium (K), nitrogen (N), white sugar content (WSC), white Sugar yield (WSY), extraction coefficient (ECS), molasses content (MC), root dry matter (DM), the average of length (L) and the average of diameter (D).
Factor analysis
Tables 5, 6 and 7 exhibited the results of the factor decomposition of sugar beet populations through the principal component method and varimax rotation. Varimax rotation is the most appropriate approach to identify the most effective components. It helps to interpret the rotated factors by setting many coefficients close to zero. Tables 5, 6 and 7 present the eigenvalue of each factor, the amount of variance of each factor (in percentage), which indicates its importance in the overall interpretation of the data, and the common variance that represents a part of the attribute variance shared with the factors, along with the specific variances of each trait. In total, the three factors explained 87.6% of the total variance.
The first factor, named WSC, explained 54.5% of the total variation. The largest positive factor coefficients in this factor belonged to ECS, WSC, SC, and DM, respectively. The second factor, called RY, accounted for 22.8% of the total variance. In this factor, the largest positive factor coefficients were RN, WSY, and RY. The third factor contributed to 10.2% of the total variance, and the two attributes D and RY had the largest positive factor coefficients. The remaining 12.6% of the variance was divided among four other factors. Liu et al. (2022) conducted that principal component analysis of ten quantitative traits showed that the cumulative contribution of the first four principal components was 71.325%.
Table 5
The total variance is explained using principal component analysis.
Component | Initial Eigenvalues | % of Variance | Cumulative % |
1 | 7.6 | 54.5 | 54.5 |
2 | 3.2 | 22.8 | 77.3 |
3 | 1.4 | 10.2 | 87.6 |
4 | 1 | 7.0 | 94.6 |
5 | 0.3 | 2.4 | 97.1 |
6 | 0.3 | 1.9 | 99 |
7 | 0.1 | 1.0 | 100 |
Table 6
The coefficient matrix for component scores underwent varimax rotation with Kaiser normalization.
Traits/Component | 1 | 2 | 3 | 4 | 5 | 6 | 7 |
RN | -0.086 | 0.408 | -0.027 | -0.14 | 0.512 | -0.585 | 0.437 |
RY | -0.057 | 0.313 | 0.228 | 0.119 | -0.197 | 0.084 | -0.462 |
LY | 0.096 | 0.079 | -0.089 | -0.106 | -1.381 | 0.315 | -0.547 |
SC | 0.166 | 0.029 | -0.048 | 0.05 | 0.281 | 0.61 | -0.009 |
Na | -0.055 | 0.033 | 0.075 | 0.04 | 0.292 | -0.031 | 1.58 |
K | -0.223 | 0.122 | -0.129 | -0.131 | 0.48 | 0.101 | -0.947 |
N | -0.07 | 0.107 | 0.012 | -0.079 | 0.265 | -1.589 | 0.289 |
WSC | 0.181 | 0 | -0.047 | 0.048 | 0.06 | 0.58 | -0.164 |
WSY | 0.018 | 0.317 | -0.112 | -0.093 | 0.516 | -0.101 | 0.147 |
ECS | 0.204 | -0.046 | 0.027 | 0.048 | -0.41 | 0.504 | -0.405 |
MC | -0.171 | 0.081 | -0.054 | -0.031 | 0.513 | 0.04 | 0.165 |
DM | 0.144 | -0.002 | -0.086 | -0.013 | 0.338 | 0.116 | 0.442 |
L | 0.11 | -0.083 | 0.055 | 1.052 | 0.163 | 0.541 | 0.261 |
D | -0.04 | 0.035 | 0.981 | 0.054 | -0.377 | -0.252 | 0.392 |
Root number (RN), root weight (RY), leaf weight (LY), gross sugar content (SC), sodium (Na), potassium (K), nitrogen (N), white sugar content (WSC), white Sugar yield (WSY), extraction coefficient (ECS), molasses content (MC), root dry matter (DM), the average of length (L) and the average of diameter (D).
Table 7
Three components extracted of component matrix using principal component analysis.
| 1 | 2 | 3 |
ECS | 0.99 | | |
DM | 0.98 | | |
WSC | 0.98 | | |
MC | -0.97 | | |
SC | 0.95 | | |
K | -0.89 | | |
Na | -0.86 | | |
N | 0.81 | | -0.43 |
RN | | 0.96 | |
RY | | 0.88 | |
LY | -0.37 | 0.79 | |
WSY | 0.59 | 0.77 | |
D | | | 0.74 |
L | | 0.32 | -0.62 |
Root number (RN), root weight (RY), leaf weight (LY), gross sugar content (SC), sodium (Na), potassium (K), nitrogen (N), white sugar content (WSC), white Sugar yield (WSY), extraction coefficient (ECS), molasses content (MC), root dry matter (DM), the average of length (L) and the average of diameter (D).
The biplot obtained from the first two components (Fig. 1) indicates the grouping of populations. Based on the results extracted from this biplot, the studied populations can be divided into three groups. The populations of M249, M224, and PPM were grouped together. The P17 population was in the second group, while the populations of P181, P191, and 732 were grouped together in the third group. However, the analysis of the principal components for this grouping was unable to include all of the variance. The first two components, on which the biplot was based, only accounted for 77.3% of the total variance. Nevertheless, the biplot was able to draw an acceptable grouping that represents the genetic flow of the populations and their differences to some extent. Cluster analysis utilizes all variance to group populations, providing more accurate results compared to principal component analysis. Hu et al. (2019) conducted a study where they used principal component analysis and cluster analysis to evaluate the variability in quality characteristics of sugar beet roots and to determine the contributions of the variables. In their research, the quality indicators of the fourteen genetic varieties were classified into four principal components based on principal component analysis results, accounting for a cumulative variance contribution rate of 91.8%. Furthermore, the genetic varieties were classified into four groups using the cluster analysis method, which aligned with the results obtained from the PCA method. We also obtained almost the same result in our own study.
Cluster analysis
The populations were classified into three different groups based on cluster analysis, as shown in Fig. 2. The first group comprised populations M249, M224, P.5, and BP-Mashhad. The second group included only population P.17, while the third group comprised populations 7233.P29, 181, and 191. The populations in the first group performed better than the other groups in terms of root number, root yield, and leaf weight. The second group had superior traits such as sodium level, potassium level, and molasses content compared to the other groups. The populations in the third group had superior traits such as pure sugar content and extraction coefficient compared to the other groups.
After analyzing the genetic information and pedigree of the populations, it has been determined that BP-Mashhad, P.5, M249, and M224 share the same genetic and belong to the same group. On the other hand, population P.17, though it originated from the original population 7233.P29, is placed in a different group. In addition, three S0 populations, namely P181, P191, and 7233.P29, have been grouped together due to their genetic similarity in the evaluated traits. The results of the cluster analysis confirm the biplot findings from the principal components analysis.
Cluster analysis can be used to determine the genetic distance between groups, and one of its applications is to identify the populations that are furthest apart in terms of genetics. In this particular experiment, the largest genetic distance was found between M249 and 191 other populations. As a result, it is expected that crossbreeding between representatives of these two populations will create the highest level of heterosis due to their maximum genetic distance from each other.
According to the results of the variance analysis in Table 8, the groups obtained from the cluster analysis were significantly different in terms of leaf weight, extraction coefficient, and molasses content at 1%. Furthermore, there was a significant difference observed at a 5% for gross sugar content, sodium, potassium, white sugar content, and root dry matter. According to the mean comparison of the groups obtained from the cluster analysis presented in Table 8, the following observations were made: Group 1 had the highest leaf weight, measuring 15.6 tons per hectare. Group 2 had the highest levels of sodium, measuring at 4.7, potassium at 8.1, and molasses content at 4.4%. Lastly, Group 3 had the highest gross sugar content, measuring at 16.2 tons per hectare, nitrogen at 3.6, extraction coefficient at 73.7%, and root dry matter at 25.4%.
Table 8
Mean square and mean comparison between groups resulting from cluster analysis in the studied populations of sugar beet
trait | MS | group 1 | | group 2 | | group 3 | |
Root number | 283.9ns | 70.0 | a | 45.4 | a | 58.3 | a |
root weight (tons/ha) | 85.2 ns | 39.3 | a | 29.7 | a | 30.2 | a |
leaf weight (tons/ha) | 20.3** | 15.6 | a | 10.9 | b | 11.2 | b |
gross sugar content (%) | 2.5* | 15.1 | ab | 13.8 | b | 16.2 | a |
sodium | 0.6* | 4.5 | ab | 4.7 | a | 3.7 | b |
potassium | 0.8* | 7.3 | b | 8.1 | a | 6.7 | b |
nitrogen | 0.3 ns | 3.1 | ab | 2.7 | b | 3.6 | a |
white sugar content | 4.3* | 10.5 | ab | 8.8 | b | 12.0 | a |
white Sugar yield (tons/hectare) | 4.4 ns | 7.2 | a | 4.0 | a | 7.1 | a |
extraction coefficient (%) | 40.1** | 69.3 | c | 63.9 | b | 73.7 | a |
molasses content (%) | 0.3** | 4.0 | b | 4.4 | a | 3.6 | c |
root dry matter (%) | 2.6* | 24.1 | ab | 23.0 | b | 25.4 | a |
the average of length | 2.0 ns | 25.7 | a | 25.1 | a | 24.1 | a |
the average of diameter | 118.0 ns | 92.9 | a | 93.4 | a | 104.2 | a |
*, ** and ns: Significant at 5%, 1%, and non-significant respectively. Populations with at least one similar letter do not show significant differences according to the LSD test (P ≤ 0.05).