The Relation between Genetic Difference for Parents of Hybrid Rice and Heterosis ()
1. Introduction
Rice is an important food crop; nearly half of the world’s population depends on rice as the main food. Breeding rice varieties with high yield and high quality are necessary to ensure global food security. In order to achieve this goal, a large number of rice varieties (combinations) with high yield and high quality have been bred, greatly increasing rice yield per unit area. However, the genetic background is relatively narrow for the parents of rice varieties (combinations) promoted in the current production, which limits the utilization of intervarietal heterosis, and increasing rice production has met a bottleneck. The hybrid F1 of indica-japonica subspecies contains huge biology heterosis [1] [2] . Exploitation of inter-subspecific heterosis in rice is recognised as an effective approach to further improving the rice yield [3] [4] . There was high degree genetic differentiation between indica-japonica subspecies genome [5] [6] , and has closely relation with the strong heterosis [7] [8] . But because of a farther genetic relationship, there were four major problems for hybrid F1: low seed-set rate, high plant height, long growth period, and grain not full [9] [10] , so the degree of difficulty for breeding hybrid rice combinations with strong heterosis was very larger.
The inter-subspecies hybrid fertility was controlled by interaction gene of a set of multiple alleles on S5 site, and found that Ketan Nangka, CPSLO17 and Dular crossed with type typical indica or japonica varieties, seed setting rate of their offspring was normal. They put forward the theory of wide compatibility, the wide compatibility gene and are multiple alleles with indica and japonica subspecies hybrid sterility genes, their genetic fits a pattern of unit point sporophyte - gametophyte interactions, and located the wide compatibility gene [11] [12] . This theory was supported by follows a string of research [13] [14] [15] , and found multiple hybrid sterility sites except S5 site, and hybrid abortion was caused by female gamete abortion [16] [17] [18] [19] and also pollen abortion [19] [20] [21] [22] .
“Building a bridge between indica and japonica” could partially solve the above problems [23] . The breeding goals of indica-compatible japonica lines was proposed [24] [25] Mixed with japonica in indica restorer line in southern China and mixed with indica in japonica in northern China. Up to now, a number of indica-japonica varieties with both wide compatibility gene and restoration gene have been developed [26] [27] [28] . Yu Yahui et al. studied the heterosis between indica-japonica subspecies by using japonica two-line hybrid rice parents and combinations as materials, and concluded that the difference of indica-japonica composition of parents was significantly and high significantly positively correlated with the yield and yield heterosis [29] . However, it was rarely for the report of successful matched indica type (or partial indica type) sterile lines with japonica type (or partial japonica type) restorer lines so far. The purpose of this study was aimed to explore the key technology of inter-indica-japonica subspecies combination to provide theoretical and practical reference for the breeding of inter-indica hybrid rice.
2. Materials and Methods
2.1. Test Materials
Sterile lines: Minyuan A, Guangzhan 63S, 33S, Shen 97A, Hengfeng A, Denoted by A1, A2, A3, A4 and A5 respectively. Provide by Hunan Liusan seed industry Co. Ltd.
Restorer line: R1, R2, R3, R4, R5, R6, R7, R8, R9, R10, R11, R12, R13, R14, R15, R16, R17 and R18, all of them was bred by research group. The control combination was provided by Long Ping seed industry.
2.2. Test Design
A total of 114 materials (including 5 sterile lines, 18 restorer lines, 90 combinations and control combination Y Liangyou 1) were planted in the breeding base of Chang’an Village, Gansha Town, Changsha County. All combinations and parents material planted according to the method of contrast test design. Except for control combination, the other material planted more than 100 plants, and with a spacing of 16.7 cm × 16.7 cm with single seedling per hill, and the distance among materials was 33 cm. Field management was accorded to the production requirements of conventional hybrid rice planting.
2.3. Sampling
Selected 2 spots for each material in the middle of plot, 10 plants for each plot to count effective panicle of unit area, sampled according to the effective panicle of unit area each plot. Inspected plant height, panicle length, total grain number of per panicle, filled grains per panicle, setting rate, 1000 grain weight in doors, 50 plants were harvested from each plot and dried and weighed, converted to the actual yield of unit area.
2.4. Measure Cheng’s Index
Identify the indica-japonica attribute of parental materials by Cheng’s index method. The total score ≤ 8 is divided into indica, 9 - 13 divided into indicalinous, 14 - 17 divided into japonicalinous, 18 to 24 divided into japonica. The specific method was as below Table 1:
1) Glume hairiness: take ten grains randomly, and observe the morphological characters of glume hairness;
![]()
Table 1. The scores and identification character according to Cheng’s index.
2) Phenol reaction: take ten grains randomly, and soak them into 2% phenol solution in the petri dish for about 24 h, then observe the glume color;
3) Interval between the 1st and 2nd nodes of panicle axis: measure ten panicles randomly between the 1st and 2nd nodes;
4) Glume color at heading: take ten plances randomly, and identify the glume color at the heading stage;
5) Leaf pubescence: take ten leaves randomly, and judge how many leaf pubescences at the heading stage;
6) Shape of grain: take five grains randomly, measure their lengths and widths, then calculate the length-width ratio.
2.5. Data Processing
The data was calculated in Excel 2010, and the correlation analysis and Euclidean distance was calculated in DPS 9.5.
Super male parents heterosis =
Control heterosis =
3. Results
3.1. Analysis on Heterosis of Test Combinations
3.1.1. Super Male Parent Heterosis of Test Combinations
Super male parent heterosis of all eight traits were positive, among of them, the strongest super male parent heterosis was filled grains per panicle, was 22.00%, followed by total grains number per panicle, was 17.42%, the smallest was effective panicles per unit area, was 1.16% (Table 2); The number of test combinations with negative super male parent heterosis for plant height trait was 8.89%. The number of combinations with positive super male parent heterosis for effective panicles per unit area, filled grains per panicle, seed setting rate, 1000 weight,
![]()
Table 2. The over male heterosis for agronomic trats and yield of test combinations (Unit: %).
Note: PH: Plant height, PP: Effective panicles per unit area, PL: Panicle leight, NGP: total number of grains per panicle, FGP: Filled grains per panicle, SR: setting rate, GW: 1000 weight, YP: Actual yield unit area, the same bellow.
and the actual yield unit area were 60.00%, 74.44%, 66.67%, 71.11%, 67.78%, 65.56% and 65.56% respectively (Table 2). The results showed that the combinations with positive super parent heterosis was common, It is possible for selecting out the combinations with the stronger super male parent heterosis.
3.1.2. Control Heterosis of Test Combinations
The control heterosis mean value was positive for effective panicles per unit area, total grain number per panicle and actual production. The control heterosis of effective panicles per unit area was strongest, was 4.69%, followed by total number of grain per panicle, was 3.27%; The control heterosis of plant heitht was weakest, was −11.75%. Combinations with negative control heterosis occupied 93.33%. It indicated that the plant height was shorter than control combination common. The number of combinations with positive control heterosis were effective panicles per unit area, filled grains per panicle, 1000 weight and the actual yield traits were 65.56%, 17.78%, 51.11%, 43.33%, 10.00%, 23.33%, 51.11% respectively. The results showed it was possible for selecting combinations with good comprehensive traits and strong control heterosis (Table 3).
3.2. Genetic Difference between the Parents of Combinations
3.2.1. Euclidean Distance between the Parents
The Euclidean distance between parents of the test combinations was in the interval 1.4 - 5.19. The range of Euclidean distance between the parents of the test combination was 3.72. Among of them, the number of combinations was 5 for the Euclidean distance greater than or equal to 5; 11 for greater than or equal to 4 and less than 5; 41 for greater than or equal to 3 and less than 4; 27 for greater than or equal to 2 and less than 3, 8 for less than 2 (Table 4). This result indicating the parents of the test combinations had certain genetic differences.
3.2.2. The Cheng’s Index Value and the Difference Value between the Parents of Combinations
Among 5 sterile lines, Cheng’s index value of A1 and A5 less than 8, was indica,
![]()
Table 3. The control heterosis for agronomic trats and yield of test combinations. Unit: %.
![]()
Table 4. The Euclidean distance between two parents.
Cheng’s index value of A2, A3 and A4 was in the interval 9 - 13, were the indicalinous. Among of 18 restorers lines, Cheng’s index value of R4, R12, R14 less than 8, were indica, Cheng’s index value of R3, R5, R6, R8, R9, R10, R13, R16, R17 and R18 was in the interval 9 - 13, were the indicalinous, Cheng’s index value of R1, R2 and R15 was in the interval 14 - 18, were japonicalinous. The number of combinations were 4, 17, 19, 17, 13, 3, 5, 6, 4 and 2 for the parent Cheng’s index difference value 9, 8, 7, 6, 5, 4, 3, 2, 1 and 0 respectively (Table 5).
3.3. Analysis on the Relation between Difference of Parents and Heterosis
3.3.1. The Correlation Analysis between Differences of Parents and Heterosis
From the phenotypic value, the phenotype values of plant height, total number of grains per panicle and filled grains per panicle were high significantly positive correlation with the Euclidean distance between the parents. The phenotype values of total number of grains per panicle and filled grains per panicle were significantly or high significantly positively correlation with the cheng’s index difference value between the parents (Table 6). From the super male parent
![]()
Table 5. The difference value between two parents.
![]()
Table 6. The correlation analysis between the difference of parents and heterosis.
Note: *means a = 0.05, r = 0.2072, 0.05 significant level, **means a = 0.01, r = 0.2702, 0.01 significant level.
heterosis. The super male parents heterosis of panicle length, total number of grains per panicle and filled grains per panicle were significantly or high significantly positive correlation with the Euclidean distance between the parents, while the super male parent heterosis of effective panicle unit area was significantly negatively correlation with the Euclidean distance between the parents. Panicle length traits was significant positive correlation between super male parents heterosis and the difference value of cheng’s index of the parents, while for other traits, there was no significant correlation between super male parents heterosis and the difference value of cheng’s index between parents (Table 6).
From the control heterosis, the control heterosis of plant height, panicle length, total number of grains per panicle, filled grains per panicle and seed setting rate trait were significantly or high significantly positively correlation with the Euclidean distance between the parents. The control heterosis of effective panicle per unit area, panicle leight, total number of grains per panicle, filled grains per panicle trait were significantly or high significantly positively correlation with Cheng’s index difference value of parents, it was significantly negatively correlated between control heterosis and Cheng’s index difference value of parents for plant height (Table 6).
The above results showed there were significant or high significant positive correlation between parental Euclidean distance and phenotype value, super male parent heterosis and control heterosis of total grain of per panicle and filled grains per panicle. The positive significant or high significant correlation between parental Cheng’s index difference value and phenotype value and control heterosis of total number grain of per panicle and filled grains per panicle, but it was no significant correlation for yield. Therefore, it was not larger for Euclidean distance and Cheng’s index difference value of parents, the stronger for actual yield heterosis of combinations.
3.3.2. The Correlation Analysis between the Different Parental Genetic Difference Interval and the Heterosis
To further understanding the relationship between parental genetic difference and heterosis, we analyzed the correlation between different intervals of parental genetic differences and heterosis, hoped to find out the parental genetic difference interval closely related with heterosis. The parental Euclidean distance in the interval of 3.41 - 3.52, it was high significant negative correlation between the parental Euclidean distance and the phenotype value of plant height. it was significant or high significant positive correlation between the parental Euclidean distance and the phenotype value of effective panicles per unit area, filled grains per panicle, 1000 - grain weight and actual yield and parental (Table 7). It was possible for selecting the combinations with good comprehensive performance and plant height is shorter, When the Euclidean distance of the parents in the interval of 3.41 - 3.52.
In the interval of 3 - 5 for Chen’s index difference value, there was a significant negative correlation between the phenotype value of plant height and
![]()
Table 7. The correlation analysis on interval of the difference for parents and the heterosis for agronomic trats and yield of combinations.
Note: *means 0.05 significant level, **means 0.01 significant level.
parental Cheng’s index difference value. the phenotype values of effective panicles per unit area, total number of grains per panicle, filled grains per panicle, setting rate and the actual yield was significant or highly significant positive correlation with Cheng’s index difference value of parents, that when Cheng’s index difference value of the parents in the range of 3 - 5, it was possible for screening out combinations with good comprehensive properties, and plant height is short.
When the Euclidean distance of parents was in the interval 3.45 - 3.58, the super male parent heterosis of plant height was negatively correlated with the Euclidean distance of parents, there was significant or highly significant positive correlation between the super male parent heterosis of parents effective panicle per unit area, total number of grains per panicle, seed setting rate, 1000 - grain weight and actual yield and the parental Euclidean distance. This indicated that when the Euclidean distance of parents was 3.45 - 3.58, it was completely possible to screen out the combinations with better comprehensive properties and shorter plant height than the male.
When Cheng’s index difference value of parents was in the interval of 3 - 4, the super male parent heterosis of plant height was negatively correlated with the Cheng’s index difference value of parents, the super male parents heterosis of effective panicle per unit area, total number of grains per panicle, setting rate, 1000 - grain weight and actual yield was significant positive correlated with the Cheng’s index difference value of parents. It showed when Cheng’s index difference value is in the interval of 3 - 4, it is possible to screen out the combinations with stronger super male parent heterosis and shorter plant height than the parent.
When the Euclidean distance of parents was in the interval 3.41 - 3.46, the control heterosis of plant height traits was highly significant negative correlation with the Euclidean distance of parents, the control heterosis of effective panicles per unit area, filled grains per panicle, setting rate and the actual yield was significant or highly significant positive correlation with Euclidean distance of parents. This indicated when the Euclidean distance of parents was in the interval of 3.41 - 3.46, it was completely possible to select combinations with better comprehensive traits and shorter plant height than the control combination.
When Cheng’s index differences of parents was in the interval of 3 - 4, the control heterosis of plant height was significant negative correlation with the Cheng’s index difference value of parents, The control heterosis of effective panicles per unit area, filled grains per panicle, setting rate and the actual yield were significantly or highly significant positive correlation with Cheng’s index difference value of parents. When Cheng’s index difference value of parents was in the interval of 3 – 4, this indicated that it was completely possible to select combinations with better comprehensive traits and shorter plant height than the control combination, when Cheng’s index difference value between parents was in the intervals 3 and 4.
Based on the above analysis, we could find the Phenotypic value, super parent and control heterosis of effective panicles per unit area, filled grains per panicle, seed setting rate, 1000 - grain weight and the actual yield traits were significantly or highly significant positive correlation with Euclidean distance and Cheng Shi index difference value of parents, when Euclidean distance of parents was in the interval 3.41 - 3.46 and Cheng’s index difference value of parents was in the interval 3 - 4. This indicated that the greater the genetic difference was, the stronger the heterosis was, as long as the genetic difference of the parents was in the proper range.
4. Conclusions and Discussion
The hybrid F1 of indica-japonica subspecies contains huge biology heterosis [28] [30] , which is closely related to the genetic differences between the parents, but the former research results were not consistent. Some scholars believed that the seed setting rate of indica-japonica hybrid F1 was significantly negatively correlated with the Cheng’s index difference of parents [30] [31] [32] [33] . The fertility index of F1 would show a peak value, and the indica-japonica intermediate type parents with moderate genetic differentiation could make up a high heterosis combination within a certain range of genetic differentiation of parents [34] . It was relatively moderate that the cheng’s index difference value of parents was in the interval 6 - 13. At this interval, the biological and economic heterosis was stronger, the parental Cheng’s index difference value was significantly negatively correlated with the setting rate of F1, while was significantly positive correlated with the biological yield of F1 (Yang et al., 1991. Liu et al., 1992.) [31] [35] . When Cheng’s index difference value of parents was in the interval 1 - 6, the performance was excellent for main yield characters of combinations including effective panicle per plant, grain number per panicle, seed setting rate and yield per plant, and the super male and control heterosis were strong [36] . However, some people believed that the genetic difference between parents of three-line hybrid rice was not closely related to heterosis, and the yield of hybrid rice was closely related to the yield of both parents, especially the average yield of both parents [37] .
In this study, 18 restorer lines with different degrees of indica and japonica were crossed with 5 indica sterile lines, and 90 combinations were prepared. The Euclidean distance between parents ranged from 1.40 to 5.19, and the difference in Cheng’s index ranged from 0 to 9. The relationship between the Euclidean distance and Ching’s index between parents and heterosis was explored. Results showed that Euclidean distance and Cheng’s index difference value of parents were significantly or highly significantly positively correlated with the phenotypic value and control heterosis of total number grains and filled grains per panicle of test combinations. The parental Euclidean distance were significantly or highly significant positive correlation with super male parent heterosis of total number grains and filled grain per panicle, significantly negative correlation with super male parent heterosis of effective panicles per unit area, significantly positive correlation with control heterosis of setting rate, but not significantly correlation with the phenotype of value actual yield, super parent and control heterosis.
Further analysis showed that the Euclidean distance and Cheng Shi index difference value were in the interval 3.41 - 3.46 and 3 - 4 respectively, effective panicles per unit area, total grain number per panicle, filled grain per panicle, seed setting rate, 1000 - grain weight and the actual yield of the phenotypic value, super parent and contrast heterosis were significantly or high significantly positive correlated with Euclidean distance and Cheng’s index difference value of parents. The results indicated that the greater the genetic difference, the stronger the heterosis, when the genetic difference of parents was in the proper range.
Acknowledgements
The research was supported by the National Key Research and Development Project (2017YFD301500) of China.
Authors’ Contributions
Z. X., and Z. Q. conceived and designed the study. Z. X., T. X., and Z. J., performed the experiments. Z.X. wrote the paper.