Efficient corn drought-resistant breeding method integrating molecular marker development and whole genome prediction model

Abstract

The invention discloses an efficient corn drought-resistant breeding method integrating molecular marker development and a whole genome prediction model. The efficient corn drought-resistant breeding method comprises the steps of hybridization matching and DH line induction, multi-environment phenotype and genotype analysis, three-level marker set screening, optimal prediction system construction, excellent selfing line screening and the like. The method has the remarkable effects that a whole genome selection technology is applied to identification and screening of the saline-alkali-tolerant and drought-tolerant corn materials, a whole genome breeding process for selecting the saline-alkali-tolerant and drought-tolerant materials is established, and generation intervals are shortened through early selection, so that the breeding value estimation accuracy is improved, the genetic progress is accelerated, and the breeding efficiency is improved. And the method has a good prediction effect on low-heritability and difficult-to-measure complex characters, and truly realizes the breeding practice guided by a genome technology.

Description

An efficient method for drought-resistant maize breeding that integrates molecular marker development and genome-wide prediction models Technical Field

[0001] This invention relates to the field of maize molecular breeding technology, specifically to an efficient method for drought-resistant maize breeding that integrates molecular marker development and whole-genome prediction models. Background Technology

[0003] Drought resistance in maize is a complex quantitative trait, and its phenotypic variation is influenced by many factors. With the development of genomics research and biotechnology, various molecular techniques have been applied to maize breeding systems, thereby helping to accelerate the breeding process and improve breeding efficiency. Among them, genome-wide selection (GS) technology, as an emerging molecular-assisted breeding method, is receiving increasing attention from commercial breeding companies and researchers. GS estimates the effects of all markers or haplotypes across the entire genome to obtain the genomic estimated breeding value (GEBV). The biggest difference between GS and traditional marker-assisted selection (MAS) is that GS involves not only analyzing a set of significant molecular markers but also jointly analyzing all markers in the population, then using BLUP models based on phenotypic data from a reference population to predict individual breeding values, followed by artificial selection. GS breeding technology has already been implemented in plants. Many studies in recent years have shown that GS has demonstrated tangible genetic gains in maize breeding. Compared to traditional breeding methods, this significantly accelerates the maize breeding cycle, improves selection efficiency, and reduces breeding costs.

[0004] However, current research focuses primarily on QTL (quantitative trait loci) mining and gene pathway analysis, and has not yet been applied to screening parental materials. Therefore, using genome-wide selection to screen for salt-tolerant and drought-resistant materials is of great significance. Summary of the Invention

[0005] To address the shortcomings of existing technologies, the purpose of this invention is to provide an efficient method for drought-resistant maize breeding that integrates molecular marker development and whole-genome prediction models. This method applies whole-genome selection technology to the identification and screening of salt-alkali-tolerant and drought-resistant maize materials, and establishes a whole-genome breeding process for selecting salt-alkali-tolerant and drought-resistant materials.

[0006] To achieve the above objectives, the technical solution adopted by this invention is as follows: A highly efficient method for drought-resistant maize breeding that integrates molecular marker development and whole-genome prediction models, the key of which includes the following steps: Step 1, hybridization and DH line induction: Selecting highly drought-resistant inbred lines as group A, selecting inbred lines with excellent agronomic traits as group B, hybridizing to obtain F1 population, and self-pollinating F1 population to obtain F2 population; F2 tassels are induced by treatment with 0.05%-0.08% colchicine for 48-72 hours, hybridized with test varieties, and haploids are screened, and doubled to obtain DH lines; Step 2, multi-environmental phenotypic and genotypic analysis: Planting DH lines at multiple drought gradient test sites and measuring drought resistance indicators. The steps are as follows: 1. Determine the phenotypic composite value; 2. Obtain a high-quality SNP marker set using GBS sequencing, and construct the principal component matrix and phylogenetic matrix; 3. Screening of the three-level marker set: Use a mixed linear model to perform genome-wide association analysis and construct a three-level drought-resistant related significant locus marker set; 4. Construction of the optimal prediction system: Use rrBLUP, SVM, and RF models to perform genome-wide selection analysis, and combine with 5-fold cross-validation to screen the optimal genome-wide selection prediction system; 5. Screening of superior inbred lines: Estimate the genotypic breeding value of DH lines, screen candidate DH lines according to the screening threshold, and the candidate lines that meet the requirements in pot verification are the superior drought-resistant inbred lines.

[0007] Further, step 1, the hybridization and DH line induction, includes the following steps: Step 1.1, Parental selection and hybridization: Select groups A and B, which have significant differences in drought resistance, as parents. Produce F1 hybrids using a hybridization pattern of group A × group B. Plant 300 plants in each hybrid combination. Perform routine field management to ensure normal pollination and seed setting, and harvest F1 seeds. Step 1.2, F2 population construction: Sow F1 seeds individually. Self-pollinate during the flowering period. Harvest F2 seeds from each F1 plant to construct an F2 population containing 1000 individual plants. Step 1.3, DH line induction and identification: Select fresh tassels from F2 plants as explants and treat them with colchicine. Hybridize the induced material with test varieties, harvest hybrid ears, and identify chromosome ploidy using flow cytometry to screen haploid seeds. Soak haploid seeds in 0.1% colchicine for 24 hours before sowing, and harvest self-pollinating DH lines.

[0008] Furthermore, in step 1.3, when colchicine is used for induction treatment, the induction concentration is 0.05%-0.08%, the induction treatment time is 48-72h, and the induction temperature is 25℃.

[0009] Furthermore, the steps described in step 2, which involve planting the DH line at multiple drought gradient test sites and measuring drought resistance indicators to determine the phenotypic composite value, include: Step A1, planting 500 DH lines at 3 different drought gradient test sites using a split-plot design, with the main plot representing the water treatment and the subplot representing the DH line, with 3 replicates, 30 plants per plot, and a plant spacing of 40cm × 16cm; Step A2, drought stress treatment: stopping water supply for 20 days each during the seedling and grain-filling stages, maintaining the soil moisture content in the 0-40cm layer at 35%-45% of field capacity; normal water supply maintaining 75%-85%; Step A3, measuring key indicators, and taking the average of 3 replicates for each indicator; Step A4, calculating the multi-environment average and performing principal component analysis, extracting the first 3 principal components as the phenotypic composite value.

[0010] Furthermore, the steps in step 2, which involve obtaining a high-quality SNP marker set using GBS sequencing and constructing the principal component matrix and phylogenetic matrix, include: Step B1, Genomic DNA extraction: Extracting DNA from 500 young leaves of DH seedlings using a modified CTAB method; Step B2, GBS sequencing and SNP screening: Digesting the SNPs with PstI-HF restriction endonuclease, recovering 350-550bp fragments, and screening SNPs using the TASSEL-GBS pipeline to obtain a high-quality SNP marker set; Step B3, Population genetic matrix construction: Performing principal component analysis on the SNP marker set to obtain the principal component matrix, and calculating the phylogenetic matrix using the Centered IBS algorithm.

[0011] Furthermore, step 3 employs a mixed linear model for genome-wide association analysis to construct a set of three levels of drought-resistant significant loci markers. The steps include: Step 3.1, Association Analysis Model Construction: Using the mixed linear model in GAPIT software, the phenotypic composite value obtained in step 2 is used as the dependent variable, and SNP markers are used as independent variables. Principal component matrices and phylogenetic matrices are incorporated to construct the association analysis model. Step 3.2, Screening of the Three-Level Significant Loci Set: A significance threshold is set, and Venn diagrams are drawn using the VennDiagram package in R software to verify the overlap of each marker set loci and ensure the uniqueness of core loci. Step 3.3, Marker Set Functional Validation: By comparing with the MaizeGDB database, the overlap rate between each marker set loci and known maize drought resistance QTLs is analyzed to obtain the set of three levels of drought-resistant significant loci markers.

[0012] Furthermore, the mathematical expression of the association analysis model is: Y = Xα + Qv + Ku + e, where Y is the phenotypic composite value, X is the SNP matrix, α is the labeling effect, Q is the principal component matrix, v is the structural effect, K is the kinship matrix, u is the random effect, and e is the error.

[0013] Furthermore, the steps for constructing the optimal prediction system described in step 4 include: Step 4.1, Dataset setup: Divide 500 DH lines into training and validation sets in a 7:3 ratio, and construct an equal number of random marker sets for All, Loc1, and Loc2 marker sets as controls, forming a total of 6 datasets; Step 4.2, Parallel construction of three models: Implement the specified rrBLUP, SVM, and RF statistical models in R software; Step 4.3, Model evaluation and optimization: Calculate the prediction accuracy and coefficient of determination of each model-marker set combination using 5-fold cross-validation combined with 10 repetitions; Step 4.4, Determination of the optimal system: Select the optimal whole-genome selection prediction system for subsequent DH line breeding value estimation.

[0014] Furthermore, the steps for screening superior inbred lines in step 5 include: Step 5.1, Breeding value estimation: Input the SNP data of 500 DH lines into the optimal whole-genome selection prediction system, calculate the estimated genomic breeding value of each DH line, and sort them from high to low; Step 5.2, Screening of superior materials: Set the screening threshold, initially screen out 35 candidate DH lines, conduct drought stress pot verification, measure the actual drought resistance index, and determine the candidate lines with a verification compliance rate ≥90% as superior drought-resistant inbred lines.

[0015] Furthermore, in step 5.2, when verifying the potted plants, the potted plant specifications are: 30cm in diameter × 40cm in height, the soil is garden soil: leaf mold = 2:1, and watering is stopped for 15 days during the seedling stage.

[0016] The significant effects of this invention are: 1. This invention applies whole-genome selection technology to the identification and screening of salt-alkali- and drought-resistant maize materials, establishes a whole-genome breeding process for selecting salt-alkali- and drought-resistant materials, and shortens the generation interval through early selection, improving the accuracy of breeding value estimation, accelerating genetic progress, and having a good predictive effect on complex traits with low heritability and difficult measurement, truly realizing the guidance of genomic technology for breeding practice; 2. This invention establishes an economical and effective screening process for salt-alkali- and drought-resistant materials and an efficient whole-genome selection technology. Attached Figure Description

[0017] Figure 1 is a flowchart of the method of the present invention. Detailed Implementation

[0018] The specific embodiments and working principles of the present invention will be further described in detail below with reference to the accompanying drawings.

[0019] Example: This embodiment of the invention focuses on drought-resistant maize breeding in Xinjiang, incorporating bio-breeding technologies including molecular markers, whole-genome selection, maize pollen mixing selection (25 male parents and 20 female parents, selecting male and female parent groups respectively), and haploid breeding. Pollen mixing selection and molecular markers can be added to existing technology roadmaps. Specifically, as shown in Figure 1, the specific steps of the efficient method for drought-resistant maize breeding integrating molecular marker development and whole-genome prediction models described in this invention are as follows: Step 1, Hybridization and DH line induction: Select a high drought-resistant inbred line as group A, select a superior agronomic trait inbred line as group B, hybridize to obtain F1 population, and self-pollinate the F1 population to obtain F2 population; the F2 tassels are induced by treatment with 0.05%-0.08% colchicine for 48-72 hours, hybridized with test varieties, and haploids are screened and doubled to obtain DH lines; In the specific implementation process, the specific steps of hybridization and DH line induction are as follows: Step 1.1, Parental selection and hybridization: Select group A (high drought-resistant inbred line) and group B (superior agronomic trait inbred line) with significant differences in drought resistance as parents, and prepare F1 hybrids according to the hybridization pattern of group A × group B. Plant 300 plants in each hybrid combination, and ensure normal pollination through conventional field management. F1 generation seeds were harvested after the seeds were set. Step 1.2, F2 generation population construction: F1 generation seeds were planted individually at a row spacing of 40cm and a plant spacing of 16cm. Self-pollination was carried out during the flowering period. F2 generation seeds were harvested from each F1 individual plant to construct an F2 population containing 1000 individual plants. Step 1.3, DH line induction and identification: Fresh male inflorescences of F2 generation individual plants were selected as explants and treated with colchicine. In this example, the colchicine induction concentration was 0.05%-0.08%, the induction time was 48-72h, and the induction temperature was 25℃. The induced material was hybridized with the Xinjiang native test variety (KWS9384), and the hybrid ears were harvested. Chromosomal ploidy was identified by flow cytometry, and haploid seeds were screened. The haploid seeds were soaked in 0.1% colchicine for 24h and then sown. Self-pollinating DH lines were harvested, and finally 500 DH lines were obtained for subsequent analysis.

[0020] Step 2, Multi-environment Phenotypic and Genotypic Analysis: DH lines were planted at multiple drought gradient test sites, and drought resistance indicators were measured to determine the phenotypic composite value; high-quality SNP marker sets were obtained using GBS sequencing, and principal component matrices and phylogenetic matrices were constructed; in specific implementation, the steps of planting DH lines at multiple drought gradient test sites and measuring drought resistance indicators to determine the phenotypic composite value include: Step A1, 500 DH lines were planted at 3 different drought gradient test sites using a split-plot design, with the main plot representing water treatments (drought stress and normal water supply), and the subplot representing DH lines, with 3 replicates, 30 plants per plot, and a plant spacing of 40cm × 16cm; Step A2 1. Drought stress treatment: Stop water supply for 20 days each during the seedling stage (three-leaf and one-heart stage) and the grain-filling stage, and maintain the soil moisture content of 0-40cm layer at 35%-45% of field capacity; maintain normal water supply at 75%-85%; Step A3: Measure key indicators, and take the average of 3 replicates for each indicator; preferably, the measured indicators include: flowering period (tasseling stage, silking stage), yield (plot yield), drought resistance index ((drought stress yield / population average drought yield) × (normal yield / population average normal yield)); Step A4: Calculate the multi-environment average and perform principal component analysis, and extract the top 3 principal components with a cumulative contribution rate ≥85% as the phenotypic composite value.

[0021] Furthermore, the steps for obtaining a high-quality SNP marker set using GBS sequencing and constructing the principal component matrix and phylogenetic matrix include: Step B1, Genomic DNA extraction: DNA was extracted from 500 young leaves of DH seedlings using a modified CTAB method. In this example, the specific steps for DNA extraction are as follows: ① Take 0.2g of leaves, grind them into powder using liquid nitrogen, add 800μL of 2% CTAB extraction buffer (containing 0.2% β-mercaptoethanol), incubate at 65℃ for 40min, and invert and mix once every 10min; ② Add an equal volume of chloroform-isoamyl alcohol (24:1), invert and mix for 15min, centrifuge at 10000r / min for 15min; ③ Take the supernatant, add 0.7 times the volume of isopropanol, precipitate at -20℃ for 40min, and centrifuge at 10000r / min for 15min; ④ Wash the precipitate twice with 75% ethanol, air dry, and dissolve in 80μL of TE buffer (pH 8.0); ⑤ Nanodrop Purity was determined by 2000 (A260 / A280=1.8-2.0, A260 / A230≥1.5), integrity was determined by 1.0% agarose gel electrophoresis, and quantified to 50 ng / μL using Qubit 4.0 for later use.

[0022] Step B2, GBS Sequencing and SNP Screening: PstI-HF restriction endonuclease was used for digestion, and 350-550bp fragments were recovered. Paired-end sequencing was performed using the Illumina NovaSeq 6000 platform, and SNPs were screened using the TASSEL-GBS pipeline in conjunction with the maize B73 reference genome to obtain a high-quality SNP marker set. The reaction system for enzyme digestion in this example was as follows: DNA 100ng, enzyme 1μL, 10×CutSmart Buffer 2μL, ddH2O added to 20μL, digestion at 37℃ for 3h, and inactivation at 80℃ for 20min. Add the barcode adapter for ligation, and PCR amplify after overnight ligation at 16℃ (primers: F: 5'-AATGATACGGCGACCACCGAGATCTACAC-3', R: 5'-CAAGCAGAAGACGGCATACGAGAT-3'). Amplification conditions: 95℃ pre-denaturation for 5 min; 95℃ denaturation for 30 s, 62℃ annealing for 30 s, 72℃ extension for 40 s, 20 cycles; 72℃ extension for 5 min.

[0023] In this example, the criteria for screening SNPs using the TASSEL-GBS pipeline are: sequencing depth ≥8×, integrity ≥85%, minor allele frequency ≥0.05, and Hardy-Weinberg equilibrium P-value ≥0.01.

[0024] Step B3, Population Genetic Matrix Construction: Principal component analysis was performed on the SNP marker set using Admixture software. The K value was set to 5 to obtain a principal component matrix containing the first three principal components. The kinship matrix was calculated using the Centered IBS algorithm in TASSEL software, ensuring that the diagonal values ​​of the matrix were 1 and the off-diagonal values ​​ranged from -0.1 to 0.9 for subsequent association analysis.

[0025] Step 3, Tertiary Marker Set Screening: A genome-wide association analysis (MLM) was performed using a mixed linear model to construct a tertiary drought-resistant associated significant locus marker set. The specific steps for constructing this set are as follows: Step 3.1, Association Analysis Model Construction: A mixed linear model (MLM) was used with the phenotypic composite value obtained in Step 2 as the dependent variable and SNP markers as independent variables. Principal component matrices and kinship matrices were incorporated, with the principal component matrix controlling for population relationships and the kinship matrix controlling for kinship relationships. In this example, the mathematical expression of the association analysis model is: Y = Xα + Qv + Ku + e, where Y is the phenotypic composite value, X is the SNP matrix, α is the marker effect, Q is the principal component matrix, v is the structural effect, K is the kinship matrix, u is the random effect, and e is the error.

[0026] Step 3.2, Screening of the three-level significant locus set: Set a significance threshold for screening, use the VennDiagram package in R software to draw a Venn diagram, verify the overlap of loci in each marker set, and ensure the uniqueness of the core loci; In this example, the significance threshold includes: ① All marker set: all high-quality SNP markers obtained in step 2.2; ② Loc1 marker set: significantly associated loci with P < 0.01; ③ Loc2 marker set: core significantly associated loci with P < 0.001; At the same time, a set of intermediate threshold loci with P < 0.005 is added for comparative analysis; Step 3.3, Verification of marker set function: Analyze the overlap rate between each marker set locus and known maize drought resistance QTLs through comparison with the MaizeGDB database, where the overlap rate of Loc1 and Loc2 marker sets should be ≥ 30% and ≥ 50% respectively, to ensure the association between the marker set and drought resistance traits, and obtain the three-level drought resistance-related significant locus marker set.

[0027] Step 4, Construction of the Optimal Prediction System: Three models—rrBLUP, SVM, and RF—were used for genome-wide selection analysis, combined with 5-fold cross-validation, to select the optimal genome-wide selection prediction system. In this embodiment, the specific steps for constructing the optimal prediction system are as follows: Step 4.1, Dataset Setup: 500 DH lines were divided into training and validation sets in a 7:3 ratio. This division was repeated 10 times to eliminate random errors. Equal numbers of random marker sets were constructed for the All, Loc1, and Loc2 marker sets as controls, resulting in a total of 6 datasets. Step 4.2, Parallel Construction of the Three Models: The specified rrBLUP, SVM, and RF statistical models were implemented in R software. The parameter ranges for each model were set as follows during model construction: ① rrBLUP model (rrBLUP package version 4.6.1), parameter: trait = phenotypic composite value, geno = SNP matrix, K = NULL; ② SVM model (e1071 package version 1.7-13), kernel function is radial basis function, grid search optimization parameters, parameter optimization range: cost = 1-15, gamma = 0.001-0.1; ③ RF model (randomForest package version 4.7-1.1), parameters: ntree = 1500, mtry = √n (n is the number of labels).

[0028] Step 4.3, Model Evaluation and Optimization: Using 5-fold cross-validation with 10 replicates, the prediction accuracy (rGEBV, Pearson correlation coefficient between predicted breeding values ​​and actual phenotypic composite values) and coefficient of determination (R²) of each model-marker set combination were calculated. 2 The calculation results show that the Loc1 labeled set + SVM model has the highest prediction accuracy (≥0.83), R0 2≥0.69; All marker set + rrBLUP model is the next best; Loc2 marker set has insufficient prediction stability due to too few loci; Step 4.4, Determination of the optimal system: Based on the calculation results, the optimal whole genome selection prediction system is selected, that is, Loc1 marker set (P<0.01) + SVM model is the optimal whole genome selection prediction system, which is used for subsequent DH line breeding value estimation.

[0029] Step 5: Screening of superior inbred lines: Estimate the breeding value of the DH line genome, screen candidate DH lines according to the screening threshold, and obtain the candidate lines that meet the requirements in pot verification as superior drought-resistant inbred lines.

[0030] In specific implementation, the steps for screening superior inbred lines in this process include: Step 5.1, Breeding Value Estimation: Input the SNP data of 500 DH lines into the optimal whole-genome selection prediction system, calculate the estimated genomic breeding value (GEBV) of each DH line (the higher the GEBV, the stronger the overall drought resistance), and sort them from high to low GEBV; Step 5.2, Screening of Superior Materials: Set a screening threshold, initially screen out 35 candidate DH lines, conduct drought stress pot verification, measure the actual drought resistance index, and determine the candidate lines with a verification compliance rate ≥90% as superior drought-resistant inbred lines.

[0031] In this example, the screening thresholds included: GEBV ≥ 0.9 (the control mean was 0), drought resistance index ≥ 1.2, and yield loss rate under drought stress ≤ 18%. When conducting potted plant verification, the potted plant specifications were: diameter 30cm × height 40cm, the soil was garden soil: leaf mold = 2:1, and watering was stopped for 15 days during the seedling stage.

[0032] In summary, this invention, through steps such as hybridization and DH line induction, multi-environmental phenotypic and genotypic analysis, three-level marker set screening, construction of an optimal prediction system, and screening of superior inbred lines, has enabled the application of whole-genome selection technology to the identification and screening of salt-alkali-tolerant and drought-resistant maize materials. This establishes a whole-genome breeding process for selecting salt-alkali-tolerant and drought-resistant materials, shortens generation intervals through early selection, improves the accuracy of breeding value estimation, accelerates genetic progress, and demonstrates good predictive effects for complex traits with low heritability and difficulty in measurement. It truly realizes the guidance of genomic technology in breeding practice; furthermore, it establishes an economical and effective screening process for salt-alkali-tolerant and drought-resistant materials and an efficient whole-genome selection technology.

[0033] The technical solution provided by this invention has been described in detail above. Specific examples have been used to illustrate the principles and implementation methods of this invention. The descriptions of the embodiments above are only for the purpose of helping to understand the method and core ideas of this invention. It should be noted that those skilled in the art can make several improvements and modifications to this invention without departing from the principles of this invention, and these improvements and modifications also fall within the protection scope of the claims of this invention.

Claims

1. A highly efficient method for drought-resistant maize breeding that integrates molecular marker development and genome-wide prediction models, characterized in that, The process includes the following steps: Step 1, Hybridization and DH line induction: High drought-resistant inbred lines are selected as group A, and inbred lines with excellent agronomic traits are selected as group B. Hybridization yields the F1 population, and self-pollination of the F1 population yields the F2 population. The F2 tassels are induced by treatment with 0.05%-0.08% colchicine for 48-72 hours, hybridized with test varieties, and haploids are screened and doubled to obtain the DH line. Step 2, Multi-environment phenotypic and genotypic analysis: The DH line is planted at multiple drought gradient test sites, and drought resistance indicators are measured to determine the comprehensive phenotypic value. High-quality SNP marker sets were obtained using GBS sequencing, and principal component matrices and phylogenetic matrices were constructed. Step 3: Tertiary marker set screening: Genome-wide association analysis was performed using a mixed linear model to construct a tertiary drought-resistant related significant locus marker set. Step 4: Construction of the optimal prediction system: Genome-wide selection analysis was performed using three models: rrBLUP, SVM, and RF. Combined with 5-fold cross-validation, the optimal genome-wide selection prediction system was screened. Step 5: Screening of superior inbred lines: The estimated breeding value of the DH line genome was estimated, and candidate DH lines were screened according to the screening threshold. Candidate lines that met the requirements in pot validation were obtained as superior drought-resistant inbred lines.

2. The efficient method for drought-resistant maize breeding integrating molecular marker development and whole-genome prediction models according to claim 1, characterized in that: Step 1, the hybridization and DH line induction, includes the following steps: Step 1.1, Parental selection and hybridization: Select groups A and B, which have significant differences in drought resistance, as parents. Produce F1 hybrids using a hybridization pattern of group A × group B. Plant 300 plants per hybrid combination. Perform routine field management to ensure normal pollination and seed setting, and harvest F1 seeds. Step 1.2, F2 population construction: Sow F1 seeds individually. Self-pollinate during the flowering period. Harvest F2 seeds from each F1 plant to construct an F2 population containing 1000 individual plants. Step 1.3, DH line induction and identification: Select fresh tassels from F2 plants as explants and treat them with colchicine. Hybridize the induced material with test varieties, harvest hybrid ears, and identify chromosome ploidy using flow cytometry to screen for haploid seeds. Soak haploid seeds in 0.1% colchicine for 24 hours before sowing, and harvest self-pollinating DH lines.

3. The efficient method for drought-resistant maize breeding integrating molecular marker development and whole-genome prediction models according to claim 2, characterized in that: When colchicine is used for induction treatment in step 1.3, the induction concentration is 0.05%-0.08%, the induction treatment time is 48-72h, and the induction temperature is 25℃.

4. The efficient method for drought-resistant maize breeding integrating molecular marker development and whole-genome prediction models according to claim 1, characterized in that: Step 2, which describes planting the DH line at multiple drought gradient test sites and measuring drought resistance indicators to determine the phenotypic composite value, includes the following steps: Step A1: Planting 500 DH lines at 3 different drought gradient test sites using a split-plot design, with the main plot representing the water treatment and the subplot representing the DH line, with 3 replicates. Each plot contains 30 plants, with a plant spacing of 40cm × 16cm. Step A2: Drought stress treatment: Water supply was stopped for 20 days each during the seedling and grain-filling stages, maintaining the soil moisture content in the 0-40cm layer at 35%-45% of field capacity; normal water supply was maintained at 75%-85%. Step A3: Measuring key indicators, and taking the average of 3 replicates for each indicator. Step A4: Calculating the multi-environment average and performing principal component analysis, extracting the first 3 principal components as the phenotypic composite value.

5. The efficient method for drought-resistant maize breeding integrating molecular marker development and genome-wide prediction models according to claim 1 or 4, characterized in that: The steps described in step 2 for obtaining a high-quality SNP marker set using GBS sequencing and constructing the principal component matrix and phylogenetic matrix include: Step B1, Genomic DNA extraction: Extracting DNA from 500 young leaves of DH seedlings using the modified CTAB method; Step B2, GBS sequencing and SNP screening: Digesting with PstI-HF restriction endonuclease to recover 350-550bp fragments, and screening SNPs using the TASSEL-GBS pipeline to obtain a high-quality SNP marker set; Step B3, Population genetic matrix construction: Performing principal component analysis on the SNP marker set to obtain the principal component matrix, and calculating the phylogenetic matrix using the Centered IBS algorithm.

6. The efficient method for drought-resistant maize breeding integrating molecular marker development and whole-genome prediction models according to claim 1, characterized in that: Step 3, employing a mixed linear model for genome-wide association analysis, involves constructing a set of three levels of drought-resistant significant loci markers. The steps include: Step 3.1, Association Model Construction: Using the mixed linear model in GAPIT software, the phenotypic composite value obtained in Step 2 is used as the dependent variable, and SNP markers are used as independent variables. Principal component matrices and phylogenetic matrices are incorporated to construct the association analysis model. Step 3.2, Screening of the Three-Level Significant Loci Set: A significance threshold is set, and Venn diagrams are drawn using the VennDiagram package in R software to verify the overlap of each marker set loci and ensure the uniqueness of core loci. Step 3.3, Marker Set Functional Validation: By comparing with the MaizeGDB database, the overlap rate between each marker set loci and known maize drought resistance QTLs is analyzed to obtain the set of three levels of drought-resistant significant loci markers.

7. The efficient method for drought-resistant maize breeding integrating molecular marker development and genome-wide prediction models according to claim 6, characterized in that: The mathematical expression of the association analysis model is: Y = Xα + Qv + Ku + e, where Y is the phenotypic composite value, X is the SNP matrix, α is the labeling effect, Q is the principal component matrix, v is the structural effect, K is the kinship matrix, u is the random effect, and e is the error.

8. The efficient method for drought-resistant maize breeding integrating molecular marker development and whole-genome prediction models according to claim 1, characterized in that: The steps for constructing the optimal prediction system described in step 4 include: Step 4.1, Dataset setup: Divide 500 DH lines into training and validation sets in a 7:3 ratio, and construct equal numbers of random marker sets for All, Loc1, and Loc2 marker sets as controls, forming a total of 6 datasets; Step 4.2, Parallel construction of three models: Implement the specified rrBLUP, SVM, and RF statistical models in R software; Step 4.3, Model evaluation and optimization: Calculate the prediction accuracy and coefficient of determination for each model-marker set combination using 5-fold cross-validation combined with 10 replicates; Step 4.4, Determination of the optimal system: Select the optimal whole-genome selection prediction system for subsequent DH line breeding value estimation.

9. The efficient method for drought-resistant maize breeding integrating molecular marker development and whole-genome prediction models according to claim 1, characterized in that: The steps for screening superior inbred lines in step 5 include: Step 5.1, Breeding value estimation: Input the SNP data of 500 DH lines into the optimal whole genome selection prediction system, calculate the estimated genomic breeding value of each DH line, and sort them from high to low; Step 5.2, Screening of superior materials: Set the screening threshold, initially screen out 35 candidate DH lines, conduct drought stress pot verification, measure the actual drought resistance index, and determine the candidate lines with a verification compliance rate ≥90% as superior drought-resistant inbred lines.

10. The efficient method for drought-resistant maize breeding integrating molecular marker development and whole-genome prediction models according to claim 9, characterized in that: When verifying the potted plants in step 5.2, the potted plant specifications are: 30cm in diameter × 40cm in height, the soil is garden soil: leaf mold = 2:1, and watering is stopped for 15 days during the seedling stage.

Images