Method for predicting corn ear weight heterosis by using heterosis effect
Through genome-wide association analysis and improved EGBLUP method, the genetic loci of corn ear heavy hybrid dominance was localized and integrated, and the prediction model was constructed, which solved the problem of insufficient accuracy of corn hybrid dominance prediction in the existing technology, and achieved efficient prediction of ear heavy hybrid dominance, improving breeding efficiency and reducing costs.
Patent Information
- Application Number
- CN202510595544.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-05-09
- Publication Date
- 2025-08-05
AI Technical Summary
The prior art has shortcomings in the prediction of corn hybrid advantages, and it is difficult to accurately predict the ear-heavy hybrid advantages of hybrids, resulting in low breeding efficiency and high cost.
Through genome-wide association analysis and improved EGBLUP method, the genomic molecular marker of corn hybrid populations is used to locate and integrate corn ear heavy hybrid dominance genetic loci to construct a prediction model to achieve accurate prediction of hybrid ear heavy hybrid dominance.
It significantly improves the prediction accuracy of the advantages of heavy hybrids in corn hybrids, reduces breeding costs, and improves breeding efficiency.
Smart Images
Figure BDA0005394478110000031 
Figure BDA0005394478110000041 
Figure BDA0005394478110000051
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of crop breeding, and specifically relates to a method for predicting heterosis of corn ear weight by utilizing heterosis effects. Background Art
[0002] With the development of biotechnology, the breeding methods of corn hybrids are not limited to conventional breeding. The use of efficient breeding technology has greatly promoted the breeding process of corn hybrids. Whole genome prediction is a biotechnology that infers phenotypic information of traits based on individual genotypes, and has significant application prospects in improving the efficiency of crop breeding. In theory, as long as the genotype information of corn inbred lines and the phenotypes of some hybrids (such as yield, resistance and other traits) are obtained, the phenotypes of all remaining hybrids can be predicted, and excellent hybrids can be selected based on the prediction results to enter a multi-year and multi-point evaluation process. This method does not require the evaluation of a large number of useless hybrids, can effectively reduce breeding costs, accelerate the breeding process, and help breeders make data-driven and accurate decisions without relying on experience.
[0003] Numerous studies have been conducted on genome-wide prediction of important agronomic traits in maize. Using maize genome, transcriptome, and metabolome data, researchers have utilized methods such as RR-BLUP (Ridge Regression Best Linear Unbiased Predictor), GBLUP (Genomic BLUP), machine learning, and deep learning to evaluate genome-wide prediction of maize traits such as yield, plant height, and flowering time (Massman et al., 2013; Guo et al., 2016; Azodi et al., 2020; Westhues et al., 2017; Yan et al., 2021; Montesinos-López, 2018). The prediction accuracy for grain yield reached as high as 0.87. Currently, phenotypic prediction at the genomic level primarily relies on SNP (Single Nucleotide Polymorphism) variation, often using SNP molecular marker effects.
[0004] Compared to phenotypic prediction, heterotic prediction lags behind. Most heterotic prediction studies utilize molecular markers and metabolites between parents to establish relationships with heterotic vigor or predict hybrid performance, rather than true heterotic prediction (e.g., prediction of midparent vigor). Since the measure of heterotic vigor is derived from the parents and the hybrid (i.e., midparent vigor = (hybrid - parent mean) / parent mean), hybrid vigor prediction based on models designed to predict hybrid performance is clearly inappropriate. To predict heterotic vigor, Jiang et al. (2017) cleverly integrated a conversion matrix that converts hybrid and parental performance into midparent vigor within the EGBLUP (Extended Genomic Best Linear Unbiased Prediction) model for hybrid performance, enabling the prediction of heterotic vigor in bread wheat. Although this prediction method has advanced heterotic vigor prediction, it still relies on BLUP methods and SNP markers.
[0005] From the perspective of quantitative genetics, since midparent vigor is the sum of the heterotic effects at all heterotic QTL loci in the genome, there is a theoretical basis for developing methods to predict heterotic effects using heterotic effects. Summary of the Invention
[0006] The purpose of the present invention is to provide a method for predicting heterosis of corn ear weight by utilizing heterosis effect, which is suitable for efficient screening of corn hybrids.
[0007] To achieve the above object, the present invention provides the following technical solutions:
[0008] A method for predicting heterosis of corn ear weight using heterosis effects comprises the following steps:
[0009] (1) Identify genetic loci for heterosis in maize ear weight using molecular markers from the maize hybrid population genome;
[0010] (2) integrating the statistically significant genetic effects of heterosis sites in step (1) into site heterosis effects and verifying the significance of the heterosis effects;
[0011] (3) Calculate the sum of heterosis effects at significant positive loci in each hybrid;
[0012] (4) constructing a model for predicting the sum of heterotic effects at significant positive sites of hybrids using the sum of heterotic effects at significant positive sites of hybrids obtained in step (3) and the genotype;
[0013] (5) performing a linear fit using the sum of heterosis effects of significant positive sites of the hybrids in step (3) and heterosis of ear weight to establish a correlation relationship;
[0014] (6) predicting the total heterosis effect of the forward sites of the hybrid to be tested using the prediction model constructed in step (4) and the genotype of the hybrid to be tested;
[0015] (7) The heterosis effect of the forward sites predicted in step (6) is used to predict the heterosis of the ear weight of the hybrid to be tested through the fitting relationship established in step (5).
[0016] Wherein, in said step (1), the genome-wide association analysis method is used to locate the dominant genetic loci for ear weight.
[0017] Wherein, in said step (2), the significant genetic effects are integrated into the heterosis effect of corn ear weight by the following formula:
[0018]
[0019] Among them, h i represents the heterotic effect of the ith heterotic locus;
[0020] R1 represents the set of loci where the parents of the hybrid have the same alleles;
[0021] R2 represents the set of loci where the parents of the hybrid have different alleles;
[0022] I A is the characteristic function of event A, which takes the value 1 if A occurs and 0 otherwise;
[0023] For the i-th QTL in R1, the alleles shared by both parents are represented by r i express;
[0024] For the jth QTL in R2, the different alleles in the parents are represented by s i and t j express;
[0025] represents the sth QTL in the i-th QTL i and tth j dominant effect of each allele;
[0026] represents the sth QTL in the i-th QTL i Alleles and the tth allele at the jth locus j Additive interaction epistatic effect between alleles;
[0027] represents the sth QTL in the i-th QTL i Alleles and the sth allele at the jth locus i , t j Additive interaction epistatic effect between alleles;
[0028] represents the sth QTL in the i-th QTL i , t i Alleles and the sth allele at the jth locus i , t j The dominant interaction epistatic effect between alleles.
[0029] Wherein, in the step (2), the heterosis effect of each site is subjected to Pearson correlation analysis with the heterosis of the middle parent of ear weight to obtain a significant heterosis QTL for ear weight.
[0030] Wherein, in the step (3), a positive contribution site is selected for each hybrid, and the corresponding heterosis effects are added together to obtain the sum of the significant positive heterosis effects of the hybrid.
[0031] Among them, in the step (4), the improved EGBLUP method is used to construct a genetic model integrating various genetic effects: dominant effect, additive-additive interaction effect, additive-dominant interaction effect, and dominant-dominant interaction effect through the genotype of SNPs to predict the sum of the heterosis effects of significant positive sites in corn hybrids.
[0032] Wherein, in the step (5), the relationship between the sum of the heterosis effects of the randomly selected hybrid ear weight and the heterosis effects of the hybrid significant positive sites is fitted by the lm function in the R language; the hybrids are divided into 1000 bins according to the order of the size of the heterosis of the hybrids, and the mean of the sum of the heterosis effects of the hybrid ear weight and the heterosis effects of the hybrid significant positive sites in each bin is calculated for relationship model construction.
[0033] Compared with the prior art, the present invention has the following beneficial effects:
[0034] The present invention's method for predicting heterosis in corn ear weight using heterosis effects not only guides the prediction of heterosis in corn ear weight but also enables accurate prediction of the total magnitude of positive heterosis effects in corn hybrids. Predicting heterosis in corn ear weight using the method provided by the present invention can significantly improve hybrid breeding efficiency and reduce breeding costs. DETAILED DESCRIPTION
[0035] The following is a clear and complete description of the technical solutions in the embodiments of the present invention. Obviously, the embodiments described are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making any creative efforts are within the scope of protection of the present invention.
[0036] A method for predicting heterosis of corn ear weight using heterosis effects comprises the following steps:
[0037] (1) Identification of heterotic genetic loci for corn ear weight
[0038] Correlation analysis was performed using high-quality genomic SNP information from the publicly available maize CUBIC (Complete-diallel plus Unbalanced Breeding-derived Inter-Cross) population and ear weight data from five environments (Xiao et al., Genome Biology, 2021;22:148). After data filtering, BLUP data for ear weight of 30 paternal parents, 206 maternal parents, 5,706 corresponding hybrids, and 40,448 SNPs were used for the study. The analysis showed that parental heterosis for ear weight ranged from 36.1% to 419.4%, with an average of 146.8%, with significant differences between hybrids.
[0039] The improved whole-genome association analysis method of Jiang et al. (Nature Genetics, 2017; 49(12): 1741) was used to locate the heterosis loci for ear weight. A total of 16,889 significant additive-additive interaction effect loci and 343 significant dominant-dominant interaction effect loci for ear weight were detected (q < 0.05; Benjamini-Hochberg, 1995). No significant dominant effect loci or additive-dominant interaction effect loci were found.
[0040] (2) Obtaining heterosis effects in corn ear weight
[0041] The insignificant genetic effects of heterosis for corn ear weight were set to 0, and the significant genetic effects were integrated into the heterosis effect for corn ear weight using the following formula:
[0042]
[0043]
[0044] Among them, h i represents the heterotic effect of the ith heterotic locus; R1 represents the set of loci where the parents of the hybrid have the same alleles; R2 represents the set of loci where the parents of the hybrid have different alleles; I A is the characteristic function of event A, which takes the value 1 if A occurs and 0 otherwise. For the i-th QTL in R1, the alleles shared by the parents are represented by r i Indicates that for the jth QTL in R2, the different alleles in the parents are represented by s i and t j express; represents the sth QTL in the i-th QTL i and tth jdominant effect of each allele; represents the sth QTL in the i-th QTL i Alleles and the tth allele at the jth locus j Additive interaction epistatic effect between alleles; represents the sth QTL in the i-th QTL i Alleles and the sth allele at the jth locus i , t j Additive interaction epistatic effect between alleles; represents the sth QTL in the i-th QTL i , t i Alleles and the sth allele at the jth locus i , t j The dominant interaction epistatic effect between alleles.
[0045] (3) Significance analysis of heterosis effect on ear weight
[0046] Using 5,706 maize hybrids, a Pearson correlation analysis was performed between the heterotic effect at each locus and the heterosis of the parent in ear weight. A total of 5,996 significant QTLs for heterosis in ear weight were identified (q < 0.05; Benjamini-Hochberg, 1995). A single significant locus for heterotic effect in ear weight could explain up to 40.7% of the heterotic variation in ear weight.
[0047] (4) Calculation of the sum of significant positive heterotic effects on ear weight in corn hybrids
[0048] The 5,996 significant loci heterotic effects obtained contributed differently to ear weight heterosis in each maize hybrid, with some contributing positively, some contributing negatively, and some contributing nothing (statistically insignificant). For each hybrid, loci with positive contributions were selected and their corresponding heterotic effects were summed to obtain the total significant positive heterotic effect for the hybrid.
[0049] (5) Construction of a prediction model for the sum of heterosis effects on ear weight at significant positive sites in hybrids
[0050] The EGBLUP method for predicting heterosis proposed by Jiang et al. (2017) Make improvements ( Using an improved method, a genetic model integrating various genetic effects (dominance (D), additive-additive interaction (AA), additive-dominant interaction (AD), and dominant-dominant interaction (DD)) was constructed using the genotypes of 40,448 SNPs to predict the sum of heterotic effects at significant positive loci in 5,706 maize hybrids. 80% of the hybrids served as the training set, and 20% of the hybrids served as the validation set. Prediction accuracy was calculated as the mean Pearson correlation coefficient between the predicted and measured values from 100 cross-validations. The results showed that the genetic effect model accurately predicted the sum of heterotic effects at significant positive loci in hybrids, with a prediction accuracy of 0.96 to 0.98 (Table 1).
[0051] Table 1 Prediction accuracy of genetic models for the sum of heterosis effects at significant positive loci in hybrids
[0052] Prediction Model Prediction accuracy D 0.96 AA 0.97 D+AA 0.98 D+AA+AD 0.98 D+AA+AD+DD 0.98
[0053] (6) Fitting the linear relationship between the sum of heterotic effects of ear weight at significant positive sites in hybrids and heterotic effects of ear weight
[0054] The relationship between the heterosis of ear weight and the heterosis effects of significant positive sites in hybrids of 4,565 randomly selected hybrids (i.e., 80% of the materials in 5,706 corn hybrids) was fitted using the lm function in R language. The fitting model was lm (the heterosis of ear weight ~ the heterosis effects of significant positive sites in hybrids), that is, y = ax + b. Among them, y is the heterosis of ear weight, x is the heterosis effects of significant positive sites in hybrids, a is the coefficient, and b is the intercept. In order to reduce the influence of similar data, the hybrids were divided into 1000 bins according to the order of heterosis of hybrids. The mean of the heterosis of ear weight and the heterosis effects of significant positive sites in hybrids in each bin was calculated for relationship model construction. 80% of hybrids were randomly selected 10 times to construct the relationship model. The model R 2 The average value of is 0.26. The detailed information of the fitted model is shown in Table 2.
[0055] Table 2 Details of the fitted model
[0056] Model coefficient intercept <![CDATA[R 2 ]]> p Model 1 0.010 129.671 0.262 0 Model 2 0.010 129.549 0.256 0 Model 3 0.011 128.940 0.279 0 Model 4 0.010 129.589 0.258 0 Model 5 0.009 130.610 0.240 0 Model 6 0.010 129.572 0.247 0 Model 7 0.008 131.121 0.228 0 Model 8 0.009 130.233 0.259 0 Model 9 0.010 129.585 0.272 0 Model 10 0.009 130.031 0.251 0
[0057] (7) Prediction of the sum of heterotic effects of ear weight at significant positive sites in the tested hybrids
[0058] 20% of the 5,706 maize hybrids were used as test hybrids (1,141 hybrids). The sum of heterotic effects of ear weight at significant positive loci was predicted using the model constructed in step (5) and the genotypes of 40,448 SNPs in the 1,141 hybrids.
[0059] (8) Prediction of heterosis of ear weight of tested hybrids
[0060] The sum of heterotic effects of ear weight at the significant positive loci of the tested hybrid obtained in (7) and the fitting model established in (6) can be used to predict heterotic effects of ear weight of the tested hybrid. The prediction accuracy of heterotic effects of ear weight of the tested hybrid is determined by the average of the Pearson correlation coefficients between the predicted values of heterotic effects of the parent in ear weight obtained from 10 predictions and the measured values of heterotic effects of the parent in ear weight of the tested hybrid.
[0061] Through analysis, the prediction accuracy of heterosis for ear weight using the sum of heterosis effects for ear weight at predicted significant positive sites can reach 0.31. To further evaluate the prediction results, the heterosis for ear weight at the sum of heterosis effects for ear weight at the measured significant positive sites of the tested hybrids was predicted using the fitting model established in (6), and the prediction accuracy was 0.32.
[0062] Comparison of prediction accuracy shows that the prediction accuracy of the method of the present invention for predicting heterosis of ear weight by predicting the sum of heterosis effects of ear weight at predicted significant positive sites is close to that of the prediction using the sum of heterosis effects of ear weight at measured significant positive sites, with a difference of only 3.1%.
[0063] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.
Claims
1. A method for predicting heterosis of corn ear weight using heterosis effect, characterized in that: The following steps are involved: (1) Identify genetic loci for heterosis in maize ear weight using molecular markers from the maize hybrid population genome; (2) integrating the statistically significant genetic effects of heterosis sites in step (1) into site heterosis effects and verifying the significance of the heterosis effects; (3) Calculate the sum of heterosis effects at significant positive loci in each hybrid; (4) constructing a model for predicting the sum of heterotic effects at significant positive sites of hybrids using the sum of heterotic effects at significant positive sites of hybrids obtained in step (3) and the genotype; (5) performing a linear fit using the sum of heterosis effects of significant positive sites of the hybrids in step (3) and heterosis of ear weight to establish a correlation relationship; (6) predicting the total heterosis effect of the forward sites of the hybrid to be tested using the prediction model constructed in step (4) and the genotype of the hybrid to be tested; (7) The heterosis effect of the forward sites predicted in step (6) is used to predict the heterosis of the ear weight of the hybrid to be tested through the fitting relationship established in step (5).
2. The method for predicting heterosis of corn ear weight using heterosis effect according to claim 1, characterized in that: In the step (1), a genome-wide association analysis method is used to locate the dominant genetic loci for ear weight.
3. The method for predicting heterosis of corn ear weight using heterosis effect according to claim 1, characterized in that: In step (2), the significant genetic effects are integrated into the heterosis effect of corn ear weight by the following formula: Among them, h i represents the heterotic effect of the ith heterotic locus; R1 represents the set of loci where the parents of the hybrid have the same alleles; R2 represents the set of loci where the parents of the hybrid have different alleles; I A is the characteristic function of event A, which takes the value 1 if A occurs and 0 otherwise; For the i-th QTL in R1, the alleles shared by both parents are represented by r i express; For the jth QTL in R2, the different alleles in the parents are represented by s i and t j express; represents the sth QTL in the i-th QTL i and tth j dominant effect of each allele; represents the sth QTL in the i-th QTL i Alleles and the tth allele at the jth locus j Additive interaction epistatic effect between alleles; represents the sth QTL in the i-th QTL i Alleles and the sth allele at the jth locus i , t j Additive interaction epistatic effect between alleles; represents the sth QTL in the i-th QTL i , t i Alleles and the sth allele at the jth locus i , t j The dominant interaction epistatic effect between alleles.
4. The method for predicting heterosis of corn ear weight using heterosis effect according to claim 1, characterized in that: In the step (2), the heterosis effect of each site is subjected to Pearson correlation analysis with the heterosis of the middle parent of ear weight to obtain a significant heterosis QTL for ear weight.
5. The method for predicting heterosis of corn ear weight using heterosis effect according to claim 1, characterized in that: In the step (3), a positive contribution site is selected for each hybrid, and the corresponding heterosis effects are added together to obtain the sum of the significant positive heterosis effects of the hybrid.
6. The method for predicting heterosis of corn ear weight using heterosis effect according to claim 1, characterized in that: In the step (4), the improved EGBLUP method is used to construct a genetic model integrating various genetic effects: dominant effect, additive-additive interaction effect, additive-dominant interaction effect, and dominant-dominant interaction effect through the genotype of SNPs to predict the sum of heterosis effects at significant positive sites of corn hybrids.
7. The method for predicting heterosis of corn ear weight using heterosis effect according to claim 1, characterized in that: In the step (5), the relationship between the sum of the heterosis effects of the randomly selected hybrid ear weight and the heterosis effects of the hybrid significant positive sites is fitted by the lm function in the R language; the hybrids are divided into 1000 bins according to the order of the size of the heterosis of the hybrids, and the mean of the sum of the heterosis effects of the hybrid ear weight and the heterosis effects of the hybrid significant positive sites in each bin is calculated for relationship model construction.