Grape soluble solid whole genome selective breeding method
By combining whole genome selection with machine learning, the destructiveness and low precision of grape SSC prediction were solved, achieving efficient and accurate grape breeding, shortening the breeding cycle and reducing costs.
Patent Information
- Application Number
- CN202510626447.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-05-15
- Publication Date
- 2025-09-16
AI Technical Summary
Traditional grape SSC determination methods are destructive and have low prediction accuracy, long breeding cycles, and existing whole genome selection technologies are insufficiently applied in grape breeding.
Machine learning and whole-genome selection methods are used to screen genetic variation sites through whole-genome association analysis, and grape SSC prediction is carried out in combination with machine learning models to achieve non-destructive, high-throughput phenotypic prediction and early screening of excellent germplasm.
The accuracy of grape SSC prediction has been improved by 75%, significantly shortening the breeding cycle, reducing costs, and providing efficient and accurate grape breeding solutions.
Smart Images
Figure CN120656538A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of agricultural technology, and in particular to a grape soluble solids whole genome selection breeding method. Background Art
[0002] The sugar-acid balance of grape fruit directly impacts its quality and economic value, and soluble solids content (SSC) is a key indicator of sugar accumulation. Traditional SSC measurement methods (such as handheld refractometers) require destructive sampling, which is costly and difficult to reflect differences in fruit maturity across the plant. Furthermore, SSC is a quantitative trait regulated by multiple genes, and traditional molecular markers (such as SSR and RFLP) cannot accurately predict its inheritance. Furthermore, grape seedlings have a long juvenile period (2-3 years), resulting in lengthy breeding cycles and low efficiency. Existing whole-genome selection technologies have been applied in crops such as rice and corn, but are still in the exploratory stage for grape breeding. Summary of the Invention
[0003] (1) Technical problems solved In response to the shortcomings of existing technologies, the present invention aims to provide a grape SSC prediction method based on machine learning and whole genome selection, which solves the problems of destructive sampling, low prediction accuracy, and long breeding cycle in traditional technologies, and realizes non-destructive, high-throughput phenotypic prediction and early screening of excellent germplasm.
[0004] To achieve the above objectives, the present invention is implemented through the following technical solutions: The whole genome selection breeding method for grape soluble solids comprises the following steps: Step 1: Obtain whole genome sequencing data and soluble solids phenotypic data of grape germplasm resources; Step 2: Screening genetic variants significantly associated with soluble solids through genome-wide association analysis; Step 3: Predict the soluble solids phenotype based on the machine learning model and screen high-sugar germplasm.
[0005] Preferably, in step one, the whole genome sequencing data is obtained by second-generation sequencing technology, and the quality control of the original sequencing data is performed using Fastp software; in step two, the whole genome association analysis adopts a mixed linear model and is implemented by GEMMA software, and the screening criteria are at least a subset of the top 100 to 1,000,000 sites ranked by P value.
[0006] Preferably, in the whole genome association analysis, the sequencing data is aligned to the grape pan-genome Grapepan v.1.0 reference genome using vg software, and the variant sites are detected using GTX software; in step three, the machine learning model is selected from at least one of a ridge regression model, a lasso regression model, an elastic net regression model, a linear kernel support vector regression model, and a polynomial kernel support vector regression model.
[0007] Preferably, the input features of the machine learning model are genotype data of 10,000 genetic variation sites screened by genome-wide association analysis. In step 1, the soluble solids phenotypic data is measured by refractometer, and the measurement time covers two consecutive years of grape ripening.
[0008] Preferably, during the training process of the machine learning model, the samples are randomly divided into a training set and a test set in a ratio of 4:1, and five-fold cross-validation is used to optimize the model parameters; the method also includes using Beagle software to fill missing values and correct sequencing errors in the genotype data.
[0009] Preferably, the method further integrates genomic data of disease resistance or stress resistance traits to construct a multi-trait intelligent breeding decision system
[0010] The present invention provides a method for whole-genome selection breeding of grape soluble solids, which has the following beneficial effects: This invention, based on machine learning and whole-genome selection methods, achieves efficient phenotypic prediction of grape soluble solids (SSC) content. In a training set of 259 samples, the Ridge Regression model performed best, and the model's prediction accuracy initially increased and then decreased with the increase in variant sites. When the number of variant sites reached 10,000, the average prediction accuracy calculated through 100 cross-validations exceeded 75%, demonstrating that the model maintains high prediction accuracy while optimizing computational costs. Using the optimal model and the optimal variant site dataset, phenotypic prediction was performed on 64 test samples. The results showed that the Pearson correlation coefficient between the predicted and true values reached R=0.77 (P<2.2e-16), validating the reliability and accuracy of the method.
[0011] This method significantly improves the prediction efficiency of grape SSC content, reduces the destructiveness and errors of traditional measurement methods, and provides an accurate, efficient, and scalable technical solution for grape breeding, facilitating the rapid screening and optimization of high-sugar, high-quality grape varieties.
[0012] This invention uses whole-genome selection and machine learning techniques to achieve non-destructive, high-throughput prediction of grape SSC, with a prediction accuracy exceeding 75%, significantly shortening the breeding cycle and reducing field management costs. This technology can also simultaneously integrate data on traits such as disease resistance and stress tolerance, providing efficient technical support for intelligent grape breeding. BRIEF DESCRIPTION OF THE DRAWINGS
[0013] Figure 1 is a flow chart of the present invention; Figure 2 This is a diagram showing the results of genome-wide association analysis of grape soluble solids of the present invention; Figure 3 This is a graph of the full gene selection and prediction results based on machine learning in the present invention. DETAILED DESCRIPTION
[0014] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.
[0015] See also Figure 1-3 , the present invention provides a technical solution: The whole-genome selection breeding method for soluble solids in grapes was as follows: 323 grape germplasm resources were selected and planted in the grape germplasm resource nursery of the Zhengzhou Fruit Research Institute of the Chinese Academy of Agricultural Sciences (34.79°N, 113.73°E). The SSC content was measured (refractometer method) for two consecutive years during the ripening period to ensure data stability.
[0016] Next-generation sequencing was then performed using the Illumina platform to obtain whole-genome data. The raw data was quality controlled using Fastp software, low-quality sequences were filtered out, and the sequencing data were mapped to the grape pan-genome Grapepan v.1.0 reference genome using vg software. Variant sites were detected using GTX software to generate a GVCF file.
[0017] After merging the GVCF files into a VCF file, a low-missing-rate dataset was obtained using basic filtering with PLINK (--geno 0.2 --vcf-min-gq 20). Beagle software was then used to fill missing values in the genotype data and correct sequencing errors to generate a high-quality VCF file. A genome-wide association analysis was performed using a mixed linear model with GEMMA software to identify genetic variants significantly associated with SSC. The top 10,000 loci were extracted as input to the machine learning model.
[0018] The 323 samples were randomly divided into a training set (259 samples) and a test set (64 samples) in a 4:1 ratio. Based on the scikit-learn library, a ridge regression model was selected, and the genotype data of 10,000 genetic variation sites screened by whole-genome association analysis were input. The model parameters were optimized through five-fold cross-validation, and the Pearson correlation coefficient between the predicted and true values of the test set reached 0.77 (P<2.2e-16). Finally, it was deployed in an intelligent breeding decision-making system to achieve early screening of high-sugar content germplasm.
[0019] It should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or device comprising a series of elements includes not only those elements, but also other elements not explicitly listed, or elements inherent to such process, method, article, or device. In the absence of further limitations, an element defined by the phrase "comprising a..." does not exclude the presence of additional identical elements in the process, method, article, or device comprising the element.
[0020] 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 whole-genome selection breeding of grape soluble solids, characterized in that: The following steps are involved: Step 1: Obtain whole genome sequencing data and soluble solids phenotypic data of grape germplasm resources; Step 2: Screening genetic variants significantly associated with soluble solids through genome-wide association analysis; Step 3: Predict the soluble solids phenotype based on the machine learning model and screen high-sugar germplasm.
2. The grape soluble solids whole genome selection breeding method according to claim 1, characterized in that: In step 1, the whole genome sequencing data is obtained by second-generation sequencing technology, and the quality of the original sequencing data is controlled by Fastp software.
3. The grape soluble solids whole genome selection breeding method according to claim 1, characterized in that: In step 2, the genome-wide association analysis was performed using a mixed linear model using GEMMA software, with the screening criteria being at least a subset of the top 100 to 1,000,000 sites ranked by P value.
4. The grape soluble solids whole genome selection breeding method according to claim 3, characterized in that: In the genome-wide association analysis, sequencing data were aligned to the grape pan-genome Grapepan v.1.0 reference genome using vg software, and variant sites were detected using GTX software.
5. The grape soluble solids whole genome selection breeding method according to claim 1, characterized in that: In step three, the machine learning model is selected from at least one of a ridge regression model, a lasso regression model, an elastic net regression model, a linear kernel support vector regression model, and a polynomial kernel support vector regression model.
6. The grape soluble solids whole genome selection breeding method according to claim 5, characterized in that: The input features of the machine learning model are genotype data of 10,000 genetic variation sites screened by genome-wide association analysis.
7. The grape soluble solids whole genome selection breeding method according to claim 1, characterized in that: In step 1, the soluble solids phenotypic data is measured by refractometer, and the measurement time covers two consecutive years of grape ripening.
8. The grape soluble solids whole genome selection breeding method according to claim 1, characterized in that: During the training process of the machine learning model, the samples were divided into training set and test set in a ratio of 4:1, and five-fold cross-validation was used to optimize the model parameters.
9. The grape soluble solids whole genome selection breeding method according to claim 1, characterized in that: The method also includes using Beagle software to fill missing values and correct sequencing errors in genotype data.