Grape fruit transverse-diameter whole-genome selective breeding method
Through whole-genome selection breeding methods and machine learning technology, the transverse diameter traits of grape fruits can be accurately predicted, which solves the problem of low breeding efficiency in existing technologies, shortens the breeding cycle and improves fruit quality.
Patent Information
- Application Number
- CN202510746010.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-06-05
- Publication Date
- 2025-09-19
AI Technical Summary
Existing technologies have not yet fully explored the precise regulation of the transverse diameter of grape berries, resulting in low breeding efficiency and difficulty in optimizing fruit growth and development and wine quality.
A whole-genome selection breeding method was adopted, and whole-genome sequencing of multiple samples was performed using Novasek platform sequencing technology. A prediction model for grape transverse diameter phenotype was constructed in combination with machine learning methods, and high-quality variant site information was used to predict and optimize fruit transverse diameter.
It has achieved accurate prediction of the transverse diameter traits of grape berries, significantly shortened the breeding cycle, improved breeding efficiency, optimized fruit growth and development, and improved the quality and consistency of wine.
Smart Images
Figure CN120673856A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of grape fruit diameter whole genome selection breeding, and in particular to a grape fruit diameter whole genome selection breeding method. Background Art
[0002] Berry diameter, the maximum transverse diameter of a grape berry, is a key quantitative indicator of grape morphology and is crucial for assessing berry development and predicting wine quality. This parameter not only reflects the grape's growth and development, but also shows significant correlations with the accumulation of biochemical components within the berry, sensory properties, and the physicochemical characteristics of the final wine product. The performance of the diameter metric is influenced by multiple factors, including but not limited to genotypic differences, environmental variations, and human intervention. During the grapevine growth cycle, berry diameter exhibits a specific dynamic pattern of change as the phenological phase progresses. This change is closely linked to physiological processes such as carbohydrate accumulation, cell proliferation, and water transport. Berry diameter also exhibits significant differences among different grape varieties. For example, large-grain varieties (such as Kyoho and Cabernet Sauvignon) exhibit larger berry diameters, while small-grain varieties (such as Chardonnay and Pinot Noir) exhibit relatively smaller diameters. This variation in the parameter serves as an important basis for assessing grape maturity, providing a scientific basis for harvesting decisions, and facilitating precise raw material quality control, thereby optimizing the winemaking process.
[0003] In winemaking research, berry diameter parameters have a multi-dimensional impact on the sensory quality and chemical composition of the final product. Fruits with larger diameters typically exhibit higher juice content and lower skin-to-flesh ratios. This can lead to a relative dilution of secondary metabolites such as phenolics and pigments per unit volume, thus compromising the integrity of the wine's structure and the adequacy of tannin extraction. Conversely, berries with smaller diameters tend to have a higher skin-to-flesh ratio, enabling more complete material exchange during fermentation and promoting the extraction of anthocyanins, tannins, and other flavor precursors from the skin, resulting in a wine with greater structural balance and flavor complexity. Furthermore, berry diameter parameters are clearly linked to grape stress resistance and growth cycle regulation. Overly large berries can lead to imbalanced nutrient distribution within the plant, reducing the plant's overall disease resistance. On the other hand, berries with smaller diameters may be an outward manifestation of abiotic stress or growth regulation disorders, reflecting incomplete fruit development and potentially affecting the overall quality and uniformity of the grapes. By establishing a quantitative relationship model between the transverse diameter of the fruit and other quality parameters, researchers can monitor the growth and development of grapes more accurately, provide a theoretical basis for field management decisions and harvest timing, and then ensure the quality stability and flavor consistency of the final brewing product by optimizing the raw material quality control system, providing scientific support for high-quality wine production.
[0004] Previously, optimal traits for berry diameter were typically manually selected through continuous hybridization and phenotypic selection. Currently, whole-genome selection breeding methods based on machine learning have been widely applied to improve grape varieties for traits such as seed abortion, pest resistance, berry size, and flavor, significantly improving the efficiency of offspring screening. However, precise control of grape berry diameter (Tdf) has been understudied. Therefore, leveraging whole-genome variant loci information for predicting, screening, and optimizing berry diameter has the potential to shorten breeding cycles and improve breeding efficiency. Precise control of berry diameter can optimize grape berry growth and development, thereby influencing berry quality, maturity, and the flavor, structure, and color of the wine. Furthermore, this method can help identify grape varieties that exhibit stable performance across diverse environmental conditions, expand the range of suitable planting areas and the supply period, and thus bring significant economic benefits to the grape industry. Summary of the Invention
[0005] To achieve the above objectives, the present invention is implemented through the following technical solutions: A method for whole-genome selection breeding of grape fruit diameter, comprising the following steps:
[0006] Whole genome sequencing of multiple samples was performed using Novasek platform sequencing technology;
[0007] Use fastp software to perform quality control on the raw data, eliminate low-quality reads and adapter contamination in the samples, and obtain high-quality resequencing data;
[0008] The resequencing data were mapped to the pan-genome Grapepan v.1.0 constructed based on the complete PNT2T genome using the giraffe command in the vg software for variant typing and obtained bam files;
[0009] Use GTX software to process the bam file to obtain the gvcf file, and use the GTX join program to integrate the variant data set to obtain the vcf file containing single nucleotide polymorphism and insertion and deletion variant information;
[0010] The vcf file is processed through PLINK software to complete the quality control of the variant data set, and Begale software is used to fill the missing values in the vcf file, finally obtaining a high-quality variant site information file;
[0011] After obtaining grape diameter phenotypic data of 323 samples and a high-quality pan-genome variation dataset, a genome-wide association analysis of the phenotypes was performed using a mixed linear model in the GEMMA software. The association results were as follows: Figure 2 As shown;
[0012] According to the p-values generated by the Wald test in the GWAS results, the associated sites were ranked, and the information of the top 100, 500, 1000, 5000, 10000, 50000, 100000, 500000, and 1000000 sites were extracted as candidate significantly associated variant sites.
[0013] Using the high-quality variant sites and sample phenotypic data obtained by sequencing, machine learning methods were used to construct prediction models for multiple grape diameter phenotypes.
[0014] Preferably, the plurality of samples are natural populations consisting of 323 domestic grape germplasm resources provided by Zhengzhou Fruit Trees, Chinese Academy of Agricultural Sciences.
[0015] Preferably, the version of the fastp software is v. 0.23.2.
[0016] Preferably, the version of the vg software is v.1.51.0.
[0017] Preferably, the version of the GTX software is v.2.2.1.
[0018] Preferably, the version of the PLINK software is v.1.90b4.6.
[0019] Preferably, the version of the software is v. 21Apr21.304.
[0020] Preferably, the version of the GEMMA software is v. 0.98.3.
[0021] It has the following beneficial effects:
[0022] This whole-genome selection breeding method for grape fruit diameter utilizes a machine learning-based whole-genome selection breeding method for grape fruit diameter traits to successfully predict the phenotype of grape fruit diameter traits. Based on the optimal model, the ridge regression model, the WGS variation data of any grape sample can be used to predict the phenotypic value of grape diameter. When five models were evaluated using 259 training set samples, the ridge regression model performed best, and the model prediction accuracy reached a peak when the number of associated variation sites was 5,000. In the 64 test set samples of this patent, the Pearson correlation coefficient between the model prediction value and the true value reached R = 0.95 (P < 2.2e-16), proving that this method can accurately and stably predict the grape fruit diameter phenotype.
[0023] This method, when applied to actual grape breeding, can significantly shorten the breeding cycle, reduce resource and cost inputs, and achieve standardized and precise seedling cultivation. This has significant application value in the future of grape breeding and in the international market. BRIEF DESCRIPTION OF THE DRAWINGS
[0024] Figure 1 This is a flow chart for genomic prediction of grape fruit diameter traits of the present invention;
[0025] Figure 2 The results of genome-wide association analysis of transverse diameter traits of 323 grape samples of the present invention are as follows;
[0026] Figure 3 This is the prediction accuracy evaluation and verification result of the whole genome selection model based on machine learning in the present invention. DETAILED DESCRIPTION
[0027] 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.
[0028] See also Figure 1-Figure 3 The present invention provides a technical solution: a method for whole-genome selection breeding of grape fruit diameter, characterized by comprising the following steps:
[0029] Whole genome sequencing of multiple samples was performed using Novasek platform sequencing technology;
[0030] Use fastp software to perform quality control on the raw data, eliminate low-quality reads and adapter contamination in the samples, and obtain high-quality resequencing data;
[0031] The resequencing data were mapped to the pan-genome Grapepan v.1.0 constructed based on the complete PNT2T genome using the giraffe command in the vg software for variant typing and obtained bam files;
[0032] Use GTX software to process the bam file to obtain the gvcf file, and use the GTX join program to integrate the variant data set to obtain the vcf file containing single nucleotide polymorphism and insertion and deletion variant information;
[0033] The vcf file is processed through PLINK software to complete the quality control of the variant data set, and Begale software is used to fill the missing values in the vcf file, finally obtaining a high-quality variant site information file;
[0034] After obtaining grape diameter phenotypic data of 323 samples and a high-quality pan-genome variation dataset, a genome-wide association analysis of the phenotypes was performed using a mixed linear model in the GEMMA software. The association results were as follows: Figure 2 As shown;
[0035] According to the p-values generated by the Wald test in the GWAS results, the associated sites were ranked, and the information of the top 100, 500, 1000, 5000, 10000, 50000, 100000, 500000, and 1000000 sites were extracted as candidate significantly associated variant sites.
[0036] Using high-quality variant sites and sample phenotypic data obtained through sequencing, we constructed multiple prediction models for grape diameter phenotypes using machine learning methods.
[0037] We selected five representative models from the scikit-learn library: elastic net regression, ridge regression, lasso regression, and support vector regression with linear and polynomial kernels. We divided 323 grape samples into a training set (n=259) and a test set (n=64) with an 80% / 20% split. Within the training set, we evaluated the performance of each model through five-fold cross-validation on the input data, resulting in the optimal prediction model and the optimal number of associated variant datasets.
[0038] Based on the optimal model and variant dataset, we performed phenotypic prediction on the test set samples and further evaluated the model's effectiveness in predicting grape diameter phenotypes for actual samples. By calculating the Pearson correlation coefficient between the true Tdf values of the test set samples and the model's predicted values, we assessed the accuracy and effectiveness of the model in practical applications. Comparative analysis found that when 5,000 significantly associated variant sites were selected and the ridge regression model was the machine learning model, the variant information performed best in predicting grape diameter, with high accuracy, making it suitable for phenotypic prediction.
[0039] Obviously, the embodiments described are only some of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field and related fields without making creative efforts should fall within the scope of protection of the present invention. Structures, devices, and operating methods not specifically described and explained in the present invention shall be implemented in accordance with conventional means in the field unless otherwise specified or limited.
Claims
1. A method for whole-genome selection breeding of grape fruit diameter, characterized in that: The steps include: Whole genome sequencing of multiple samples was performed using Novasek platform sequencing technology; Use fastp software to perform quality control on the raw data, eliminate low-quality reads and adapter contamination in the samples, and obtain high-quality resequencing data; The resequencing data were mapped to the pan-genome Grapepan v.1.0 constructed based on the complete PNT2T genome using the giraffe command in the vg software for variant typing and obtained bam files; Use GTX software to process the bam file to obtain the gvcf file, and use the GTX join program to integrate the variant data set to obtain the vcf file containing single nucleotide polymorphism and insertion and deletion variant information; The vcf file is processed through PLINK software to complete the quality control of the variant data set, and Begale software is used to fill the missing values in the vcf file, finally obtaining a high-quality variant site information file; After obtaining grape diameter phenotypic data for 323 samples and a high-quality pan-genome variation dataset, a genome-wide association analysis of the phenotypes was performed using a mixed linear model in the GEMMA software. The association results are shown in Figure 2. According to the p-values generated by the Wald test in the GWAS results, the associated sites were ranked, and the information of the top 100, 500, 1000, 5000, 10000, 50000, 100000, 500000, and 1000000 sites were extracted as candidate significantly associated variant sites. Using the high-quality variant sites and sample phenotypic data obtained by sequencing, machine learning methods were used to construct prediction models for multiple grape diameter phenotypes.
2. The grape fruit diameter whole genome selection breeding method according to claim 1, characterized in that: The multiple samples are natural populations consisting of 323 domestic grape germplasm resources provided by Zhengzhou Fruit Tree, Chinese Academy of Agricultural Sciences.
3. The grape fruit diameter whole genome selection breeding method according to claim 1, characterized in that: The version of the fastp software is v. 0.23.
2.
4. The method for whole-genome selection and breeding of grape fruit diameter according to claim 1, characterized in that: The version of the vg software is v.1.51.
0.
5. The grape fruit diameter whole genome selection breeding method according to claim 1, characterized in that: The version of the GTX software is v.2.2.
1.
6. The method for whole-genome selection and breeding of grape fruit diameter according to claim 1, characterized in that: The version of the PLINK software is v.1.90b4.
6.
7. The method for whole-genome selection and breeding of grape fruit diameter according to claim 1, characterized in that: The version of the Begale software is v. 21Apr21.
304.
8. The method for whole-genome selection and breeding of grape fruit diameter according to claim 1, characterized in that: The version of the GEMMA software is v. 0.98.3.