A method for developing a hydrangea-type KASP marker and its application

Abstract

The present invention discloses a method for developing a hydrangea flower type KASP marker and its application. The method for developing a KASP marker of the present invention comprises the following steps: (1) obtaining typing data of SNP sites related to flower type based on resequencing data through GWAS analysis; (2) screening the data in step (1) and selecting the top 1 to 20 SNP sites with high ‑log10(P) values ​​in each Scaffold; (3) in multiple hydrangea varieties, performing genotyping verification on the SNP sites screened in step (2) through DNA sequencing and KASP methods, and finally confirming the KASP marker for hydrangea flower type identification. The present invention screens the SNP sites related to flower type obtained through GWAS analysis, and performs genotyping verification through DNA sequencing and KASP methods, so that the confirmed SNP sites have a higher correlation with the hydrangea flower type traits, providing a reference for subsequent flower type breeding.

Description

Technical Field

[0001] The invention belongs to the field of plant molecular marker assisted breeding, and particularly relates to a hydrangea type KASP marker development method and application thereof. Background Art

[0002] Hydrangea macrophylla (Thunb.) Ser. is a small deciduous shrub of the genus Hydrangea in the family Saxifragaceae, also known as hydrangea and hydrangea. Today, hydrangea has been widely used in the flower industry with its large inflorescence and rich and colorful colors (including white, pink, blue, purple, green, etc.), from fresh cut flowers produced and supplied throughout the year to potted plants, courtyards, specialty gardens, and flower landscape construction in large gardens.

[0003] Most native species of the genus Hydrangea have corymbose cymes, and the ornamental flowers are mainly composed of sepals that have evolved from leaves. In large-leaf hydrangea, there are usually two types of flowers: spherical and flat-topped. The flat-topped inflorescence is composed of non-decorative flowers in the center and decorative flowers around it, and the entire inflorescence is approximately on a plane; the spherical inflorescence is a ball, mainly composed of decorative flowers. In hydrangea, decorative flowers are obviously more ornamental, so in the process of hydrangea flower breeding, breeders are constantly trying to increase the number of decorative flowers to obtain cultivars with better ornamental value; therefore, related research on the two flower types of hydrangea, spherical and flat-topped, is also gradually being carried out, hoping to reveal their genetic laws and control genes, and lay the foundation and provide reference for subsequent flower type breeding. In addition to the inflorescence type, the double petal trait of decorative flowers is also an important breeding target, and large-leaf hydrangea can take into account both flower type and double petal traits, and is a suitable flower type breeding material.

[0004] Through hybrid breeding experiments, Uemachi et al. found that all F1 generations of flat-top inflorescence varieties and spherical varieties had flat-top inflorescences, and the ratio of flat-top to spherical flower types in the F2 generation obtained by self-pollination of the F1 generation was the expected 3:1, which was in line with Mendel's law of segregation, and it was speculated that the flower type trait might be controlled by a single major gene; Waki ​​et al. designed SSR primers through DNA sequencing technology to construct a hydrangea genetic map, and the results of linkage analysis showed that the double petals and spherical inflorescence of hydrangea were controlled by a single recessive gene, but the flower type trait segregation ratio of the F2 generation population during the experiment was 1:10, instead of the expected 1:3; Wu et al. used 5803 SNP loci to conduct whole-genome association analysis on 82 large-leaf hydrangeas, screened out a SNP locus significantly associated with flower type, and used SNP loci to identify the flower type of 18 hydrangea materials with a 100% accuracy rate. However, in the 341 F1 plants, the flower type identified by the SNP locus did not match expectations.

[0005] Although there are many studies on hydrangea flower types, they have not been able to fully explain the mechanism of spherical inflorescence formation and the genes that control this trait, so further analysis and research is needed. Summary of the invention

[0006] The purpose of the present invention is to solve the defects in the prior art, provide a method for developing a KASP marker with a high correlation with hydrangea flower type, and construct a hydrangea flower type prediction model based on the KASP marker.

[0007] In order to achieve the above object, the present invention provides a method for developing a hydrangea-type KASP marker, comprising the following steps:

[0008] (1) Based on resequencing data, GWAS analysis was performed to obtain SNP loci associated with flower type;

[0009] (2) Filter the data in step (1) and select the -log 10 (P) The top 1 to 20 SNP sites with high values;

[0010] (3) In multiple hydrangea varieties, the SNP loci screened in step (2) were genotyped and verified by DNA sequencing and KASP methods, and finally the KASP markers for hydrangea flower type identification were confirmed.

[0011] The present invention screens the SNP sites related to flower type obtained by GWAS analysis, and performs genotyping verification by DNA sequencing and KASP methods, so that the SNP sites confirmed are more highly correlated with the hydrangea flower type traits.

[0012] The population for GWAS analysis in step (1) is an F1 hybrid population, the female parent 'Tivoli' has a spherical inflorescence, and the male parent 'Tabe' has a flat-top inflorescence.

[0013] The specific steps of GWAS analysis are as follows:

[0014] Hydrangea plants were selected, and fresh young leaves free of pests and diseases were picked for sequencing and data analysis. Shapes with obvious phenotypic differences were selected, and observations were recorded when the plants were in full bloom. The raw data obtained from the observations were processed and outliers were eliminated. Samples with qualified DNA test results were selected for whole genome resequencing. The raw sequencing data were then subjected to sequence alignment and population SNP detection. PC values ​​and kinship analysis matrices were obtained through principal component and kinship analysis, respectively. Association analysis was performed on the phenotypic values ​​and genotypic values ​​of each trait, and significant loci were screened.

[0015] The present invention also provides a KASP marker for identifying hydrangea flower types, including one or more of snp1, snp3, snp4, snp8, snp10, snp11 and snp12; each site is as follows:

[0016] snp1 is located at position 460462 of Scaffold Hma1.2p1_0653F.1, which mutated from G to A;

[0017] snp3 is located at position 220821 of Scaffold Hma1.2p1_0669F.1, which mutated from G to A;

[0018] snp4 is located at position 949341 of Scaffold Hma1.2p1_0669F.1, which mutated from T to G;

[0019] snp8 is located at position 78611 of ScaffoldHma1.2p1_1283F.1, which mutated from C to T;

[0020] snp10 is located at position 796640 of ScaffoldHma1.2p1_0060F.1, which mutated from C to T;

[0021] snp11 is located at position 1540773 of ScaffoldHma1.2p1_0060F.1, which mutated from G to C;

[0022] snp12 is located at position 868484 of ScaffoldHma1.2p1_0653F.1, which mutated from A to T.

[0023] in,

[0024] The ratio of genotype G / G in the spherical inflorescence at the Snp1 locus was 100%;

[0025] At the Snp3 locus, 85.2% of the genotype G / G were spherical, and 100% of the genotype A / A were flat-topped.

[0026] At the Snp4 locus, 81.6% of the genotype G / G were spherical, and 87.9% of the genotype T / T were flat-topped.

[0027] The proportion of genotype C / C in the spherical inflorescence at the Snp8 locus was 97.9%;

[0028] At the Snp10 locus, 92.3% of the genotype T / T were spherical, and 88.0% of the genotype C / C were flat-topped.

[0029] At the Snp11 locus, 85.4% of the genotype C / C were spherical, and 90.3% of the genotype G / G were flat-topped;

[0030] At the Snp12 locus, 88.9% of the genotype T / T were spherical, and 89.3% of the genotype A / A were flat-topped.

[0031] The present invention also provides a KASP primer composition for identifying hydrangea flower types, comprising one or more of snp1 primer, snp3 primer, snp4 primer, snp8 primer, snp10 primer, smp11 primer and snp12 primer; the nucleotide sequence of each primer is shown in Table 1 below.

[0032] Table 1 Primer sequences used for KASP

[0033] Table 1 The primers used for KASP

[0034]

[0035] Note: The underlined part in the primer is the SNP site

[0036] Note:The underlined part of the primer is SNP site

[0037] The present invention also provides a method for predicting hydrangea flower type based on KASP markers, and the specific method is as follows: taking a combination of resequencing typing data in a GWAS population and verification typing data in varieties and hybrid offspring as a training data set, using the rfImpute function to automatically fill in a small number of missing values ​​therein; using the Gini coefficient and the Accuracy coefficient to evaluate the importance of each SNP site to the hydrangea flower type trait, and screening out candidate SNP markers; using a random forest algorithm to construct a model, selecting 3, 4 and 5 markers from the candidate SNP markers for model construction respectively, obtaining optimal models with different numbers of markers, and using the optimal model to predict the flower type of hydrangea plants.

[0038] In some embodiments, the marker combination of the 3-marker model in the optimal model is: snp12, snp8, snp3; the marker combination of the 4-marker model is: snp11, snp1, snp8, snp3; the marker combination of the 5-marker model is: snp10, snp11, snp1, snp8, snp3.

[0039] Compared with the prior art, the present invention has the following advantages:

[0040] The present invention screens the flower type-related SNP sites obtained by GWAS analysis based on resequencing data, and uses the KASP typing method for verification analysis to obtain SNP sites with a high correlation with flower type shape.

[0041] SNP markers were screened and the model was constructed using the random forest algorithm to obtain the optimal model and parameter settings for hydrangea flower type prediction, thereby realizing the prediction of hydrangea plants with unknown flower types.

[0042] The SNP loci screened by the present invention show polymorphism in hydrangea varieties and are closely related to the trait of inflorescence type. At the same time, the hydrangea flower type prediction model constructed using the SNP loci has high accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 Some hydrangea varieties used in Example 2 of the present invention are shown;

[0044] Figure 2 The hybrid offspring of some hydrangeas in Example 2 of the present invention are shown;

[0045] Figure 3 The KASP genotype diagram of the three SNP sites snp10, snp11 and snp12 in Example 2 of the present invention;

[0046] Figure 4 The three SNP loci snp10, snp11 and snp12 in Example 2 of the present invention are genotype diagrams in hydrangea varieties and hybrid populations;

[0047] Figure 5 The importance evaluation of each point to the hydrangea flower shape in Example 3 of the present invention;

[0048] In the figure,

[0049] Hma1.2p1_0060F.1_1540773 is snp11,

[0050] Hma1.2p1_0653F.1_460462 is snp1,

[0051] Hma1.2p1_1283F.1_78611 is snp8,

[0052] Hma1.2p1_0669F.1_220821 is snp3,

[0053] Hma1.2p1_0653F.1_868484 is snp12,

[0054] Hma1.2p1_0669F.1_949341 is snp4,

[0055] Hma1.2p1_0744F.1_396693 is snp5,

[0056] Hma1.2p1_0060F.1_796640 is snp10,

[0057] Hma1.2p1_0744F.1_678323 is snp6,

[0058] Hma1.2p1_2149F.1_31384 is snp9,

[0059] Hma1.2p1_0744F.1_875534 is snp7;

[0060] Figure 6 It is a classification diagram of the optimal model with different numbers of markers in Example 3 of the present invention; A is the optimal model with 3 markers, B is the optimal model with 4 markers, and C is the optimal model with 5 markers;

[0061] In the figure, red dots represent flat-top inflorescences, and blue dots represent spherical inflorescences. DETAILED DESCRIPTION

[0062] The present invention is described in detail below in conjunction with specific embodiments and drawings.

[0063] Example 1

[0064] 285 flower type-related SNP loci obtained from previous GWAS analysis

[0065] 1. Plant materials

[0066] The selected population was the F1 hybrid population, the female parent 'Tivoli' had a spherical inflorescence, and the male parent 'Tabe' had a flat-top inflorescence. 299 hydrangea plants were selected, fresh disease-free and insect-free young leaves were picked, immediately placed in liquid nitrogen, and sent to Beijing Biomike Biotechnology Co., Ltd. (Qingdao) on dry ice for sequencing and data analysis.

[0067] 2. Phenotypic observation items and methods

[0068] Select traits with obvious phenotypic differences, and observe and record them when the plants are in full bloom according to the DUS test index. Use EXCEL2016 to process the original data obtained from the observation and eliminate outliers.

[0069] 3. Plant DNA extraction and resequencing

[0070] DNA extraction and whole genome resequencing were commissioned to BioMed. A plant genome DNA extraction kit was used to extract total DNA, and 299 samples with qualified DNA test results were selected for whole genome resequencing.

[0071] 4. Sequence alignment and population SNP detection

[0072] The raw sequencing data filtered by the company were aligned to the Hydrangea macrophylla 'Aogashima' (download address: https: / / www.ncbi.nlm.nih.gov / nuccore / 1861644412?report=fasta). The alignment software used was BWA, and the alignment results were deduplicated using SAMTOOLS software. The final data obtained was clean data, and the file was saved as VCF (Variant Call Format) for subsequent analysis.

[0073] 5. SNP Filtering

[0074] GATK software was used to further filter the SNP and small InDel (small Insertion and Deletion) data. The screening criteria were as follows: (i) Based on the subroutine vcfutils.pl (varFilter-w 5-W 10) in bcftools, SNPs within 5 bp near INDEL and SNPs within 10 bp of adjacent INDEL were filtered out; (ii) SNP sites with Phred values ​​lower than 30 (i.e., error rate greater than 1‰) were removed; (iii) Sequencing depth: SNPs with too low or too high sequencing depth were removed, and the sequencing depths of all sites and all individuals were counted, and the screening threshold was set according to the distribution; (iv) Minor Allele Frequency (MAF): MAF with a frequency of too low (0.05-0.1) would affect the SNP identification efficiency and the judgment of the population structure, and it needed to be filtered before analysis; (iv) Missing data: Sequencing data will be missing during the sequencing and SNP identification processes, affecting subsequent analysis.

[0075] 6. Linkage disequilibrium analysis

[0076] Linkage disequilibrium analysis is the theoretical basis of association analysis, and the size of the LD value determines the accuracy of GWAS analysis. In this study, plink2 software was used to calculate the LD value between markers. In natural populations, D' or r2 is usually used to represent the strength of linkage disequilibrium. The closer r2 is to 1, the stronger the linkage relationship is, and the closer r2 is to 0, the weaker the linkage relationship is. By fitting r2 with the SNP spacing obtained by genome sequencing, the closer the distance between SNPs is, the more SNP sites are in the same fragment length, the larger r2 is, and the farther the SNP interval is, that is, the fewer SNP sites are detected in the same fragment length, and the smaller r2 is. Generally, data with r2 greater than 0.1 are used to evaluate the decay of LD values ​​with the physical distance between markers (LD decay). Generally speaking, the longer the LD decay distance, the fewer the number of SNPs per unit length, the weaker the variation of the species, and the lower the probability of recombination between SNPs; the shorter the LD decay distance, the more SNPs are in the unit length, the stronger the variation of the species, and the higher the probability of recombination between SNPs. The R language package was used to compile the LD decay graph and analyze the calculation results.

[0077] 7. Principal component and kinship analysis

[0078] Population structure refers to the existence of different subpopulations in the population under study. Species in the same subpopulation are closely related, while species between subpopulations are more distantly related. It is a widely used population clustering analysis method and an important genetic relationship analysis tool that helps to understand the evolution of materials. This study used two methods to infer the number of subpopulations in the material, Admixture analysis and Principal Component Analysis (PCA).

[0079] Based on SNP, the admixture software was used to analyze the population structure of the research material. For the research population, the number of subpopulations (K value) was set in advance, the K value was set to 1 to 10, and the length of burnin period was set to 10000. The optimal number of clusters was determined according to the maximum likelihood principle, and the population structure analysis results corresponding to the K value were plotted. The clustering results were cross-validated, and the optimal number of clusters was determined according to the valley value of the cross-validation error rate.

[0080] Principal component analysis is mainly based on the clustering of SNP differences obtained through sequencing of different research materials into different subgroups, and its results can be used to verify the results of other clustering methods such as population structure analysis. This study used EIGENSOFT software to perform principal component analysis based on SNP data, and obtained principal component values ​​as covariates for subsequent whole genome association analysis models.

[0081] GCTA version 1.92.1 software was used to calculate the kinship K (kinship) between pairs of samples to assist in evolutionary analysis.

[0082] 8. Genome-wide association analysis

[0083] Based on SNP, the mixed linear model (MLM) of Fastlmm software and GEMMA software was used, and the PC value in the principal component analysis and the kinship analysis matrix (Kinship) were used as covariates to perform association analysis on the phenotypic value and genotypic value of each trait to obtain the association value. The calculation formula is as follows:

[0084] Y=Xα+PCβ+Kμ+e

[0085] Among them, X is genotype, Y is phenotype, SNP markers significantly associated with phenotypic traits are determined, and the contribution rate of different SNP sites to the phenotype is calculated. The Manhattan plot and "qq" plot are drawn using the "qqman" package in R language.

[0086] A total of 369 significant loci (-log10(P)>8) were screened through association analysis using two methods, of which 294 were screened using Fastlmm software, and 360 were screened using the mixed linear model of GEMMA. A total of 285 significant loci were screened using the two models together, and their explanation rate for phenotypic variation was between 11.80% and 60.54%.

[0087] Example 2

[0088] SNP site screening

[0089] 1. Test materials

[0090] The experimental materials included three parts: (1) 34 plants in the GWAS population (‘Tivoli’בTabe’ hybrid offspring), of which 16 plants showed flat-top inflorescence and 18 plants showed round inflorescence; (2) 52 hydrangea varieties, of which 26 varieties showed flat-top inflorescence and 26 varieties showed round inflorescence (see Figure 1 , Table 2-1).

[0091] Table 2-152 List of Hydrangea Varieties

[0092] Table 2-1 The list of the 52 Hydrangea cultivars

[0093]

[0094] Table 2-1 continued

[0095]

[0096] The third part of the test material is 7 F 1 Generation group and 1 F 2 Two F1 generation groups were spherical, and two F 1 The first generation group is flat-topped, and the other three F 1 Generation groups and F 2 In the first generation of plants, there are both spherical and flat-topped types (see Figure 2 , Table 2-2). The hydrangea varieties and hybrid offspring plants were planted in the resource nursery of Shanghai Chenshan Botanical Garden and the nursery of Hangzhou Landscape Greening Co., Ltd.

[0097] Table 2-2 F 1 and F 2 Parents and flower phenotype

[0098] Table 2-2 The parents of the F 1 and F 2 and the inflorescence types of the hybrids

[0099]

[0100] 2. Plant genomic DNA extraction

[0101] Fresh and healthy leaves of target population plants were collected from June to August 2022, and the inflorescence type was recorded when flowering. The total genomic DNA was extracted using the Polysaccharide Polyphenol Plant Genome Extraction Kit (purchased from Tiangen Biochemical Technology (Beijing) Co., Ltd.), and the extracted DNA was measured using a Nano Drop 2000 / 2000c spectrophotometer and 1.0% agarose gel electrophoresis to determine the DNA concentration and purity. The DNA extraction steps are as follows:

[0102] (1) Weigh about 120 mg of fresh plant tissue and grind it thoroughly in liquid nitrogen;

[0103] (2) Quickly add 400 μL of GPS buffer and 10 μL of RNase A solution (10 mg / ml) to the ground powder, vortex quickly to mix, and place the centrifuge tube in a 65°C water bath for 15 min. Invert the centrifuge tube several times during the water bath to mix the sample.

[0104] (3) Add 100 μL of buffer GPA, vortex for 1 min, centrifuge at 12,000 rpm for 5 min, transfer the supernatant to the filter column CS, then centrifuge at 12,000 rpm for 1 min, and transfer the filtrate to a new centrifuge tube;

[0105] (4) Add an equal volume of anhydrous ethanol and mix thoroughly;

[0106] (5) Transfer the solution obtained in the previous step to the RNase-Free adsorption column CR2, centrifuge at 12000 rpm for 1 min, discard the waste liquid, and place the RNase-Free adsorption column CR2 into a collection tube;

[0107] (6) Add 550 μL of deproteinized solution RD to the RNase-Free adsorption column CR2, centrifuge at 12000 rpm for 1 min, discard the waste liquid, and place the RNase-Free adsorption column CR2 in a collection tube;

[0108] (7) Add 700 μL of rinse solution PW (anhydrous ethanol has been added before use) to the RNase-Free adsorption column CR2, centrifuge at 12,000 rpm for 1 min, discard the waste liquid, place the RNase-Free adsorption column CR2 in the collection tube, and repeat once;

[0109] (8) Place the RNase-Free adsorption column CR2 back into the collection tube, centrifuge at 12,000 rpm for 2 min, discard the collection tube, and then transfer the RNase-Free adsorption column CR2 to a new centrifuge tube and dry it at room temperature for 5-10 min;

[0110] (9) Add 80 μL of elution buffer TB to the RNase-Free adsorption column CR2, place at room temperature for 4 min, and then centrifuge at 12,000 rpm for 2 min. Collect the solution into a centrifuge tube and store it at -20°C.

[0111] 3. SNP site screening

[0112] The 285 flower type-related SNP loci obtained from the previous GWAS analysis were screened based on the resequencing data typing data, and the -log 10 The top 1 to 20 SNP sites with high (P) values ​​(renamed snp1, snp2, snp3, ..., snp20, respectively) were resequencing genotyping and compared with the observed phenotypes, and the sites with high correlation with the phenotypes were selected for further verification and analysis in the validation population.

[0113] 4. SNP typing

[0114] (1) KASP-PCR

[0115] Competitive allele-specific PCR (KASP): DNA sequences of 200 bp upstream and downstream of the SNP site were obtained from the hydrangea reference genome. A set of primers was designed for each SNP site, including two specific forward primers and a universal reverse primer. The 3' end bases of the two specific forward primers were complementary to the bases before and after the mutation of the SNP site, and the two base types were labeled with VIC and FAM fluorescence, respectively. The primer sequences were synthesized by LGC (see Table 2-3 for some primer sequences).

[0116] The reaction system was 10 μL, including 1 μL DNA template (concentration 60 ng / μL), 5 μL 2×KASP Master mix solution, 0.14 μL Primer assay mix (a mixture of 3 primers), and 4 μL ddHO. 2 O. Amplification reaction and data reading were performed on an ABIStepOne Plus fluorescent quantitative PCR instrument. The reaction procedure was divided into three stages: the first stage: pre-denaturation at 94°C for 15 min; the second stage: denaturation at 94°C for 20 s, extension at 61°C for 1 min, and then the temperature was reduced by 0.6°C in each cycle, for a total of 10 cycles; the third stage: denaturation at 94°C for 20 s, extension at 55°C for 1 min, for a total of 30 cycles.

[0117] Table 2-3 Some primer sequences used for KASP

[0118]

[0119] Note: The underlined part in the primer is the SNP site

[0120] Note:The underlined part of the primer is SNP site

[0121] (2) GWAS population and parent sample typing results

[0122] In order to verify the reliability of the resequencing typing data and the screened SNP sites, snp10, snp11, and snp12 were used as examples to verify the three SNP sites in some GWAS populations and their parents using the KASP typing method. The results showed that among the 34 GWAS populations and their parents, Tivoli and Tabe, 36 samples could be accurately genotyped by the KASP method ( Figure 3). The results of snp10 showed that among the 17 flat-top inflorescence plants, 15 were of C / T genotype (2 plants had no results), and all 19 plants with spherical inflorescences were of T / T genotype (Table 2-4). The parents 'Tivoli' and 'Tabei' were of T / T and C / T genotypes, respectively. That is, all heterozygous genotype C / T showed flat-top inflorescences, and all homozygous genotype T / T showed spherical inflorescences. The results of the locus snp11 showed that among the 17 flat-top plants, 15 were of C / G genotype, and 1 was of C / C genotype (1 plant had no results). Among the 19 spherical plants, 18 were of C / C genotype, and 1 was of C / G genotype. The parents 'Tivoli' and 'Tabei' were of C / C and C / G genotypes, respectively. The results of the locus snp12 showed that among the 17 flat-top plants, 16 were of A / T genotype, 1 was of T / T genotype, among the 19 spherical plants, 18 were of T / T genotype, and 1 was missing. The genotypes of the parents 'Tivoli' and 'Tabe' at this locus were T / T and A / T, respectively. The verification of some SNP loci in the GWAS population showed the reliability of the resequencing typing data and that the above three loci were closely related to the hydrangea type traits. In the typing results, each locus had only one homozygous and one heterozygous genotype, and its parents were also heterozygous and homozygous genotypes, which was in line with the DNA genetic law, and the selected loci could be further verified and analyzed.

[0123] Table 2-4 KASP genotyping results

[0124] Table 2-4 The genotyping results of KASP

[0125]

[0126] (3) Hydrangea varieties and F 1 、F 2 Group typing results

[0127] In 50 hydrangea varieties, 12 SNPs screened in the early stage were genotyped and verified by DNA sequencing and KASP. The results showed that the results of three biological replicates of each hydrangea variety were consistent. Among them, the genotyping results of snp6, snp7, and snp8 showed no polymorphism in the hydrangea varieties, only one or a very small part of two genotypes, and the three genotypes in the resequencing were speculated to be due to the single nucleotide polymorphism of the parents at the site, resulting in the hybrid F 1The remaining 9 SNP loci all showed polymorphism in the 50 hydrangea varieties. Different from the verification results in the GWAS population, the results of the four SNP loci snp10, snp11, snp12, and snp4 showed that the spherical and flat-topped inflorescences corresponded to a homozygous genotype, while the heterozygous genotype had both spherical and flat-topped inflorescences ( Figure 4 ), and the ratio of the two flower types is close to 1:1 (Table 2-5). For example, the result of Snp10 is that in the spherical inflorescence, 3 varieties are C / C homozygous genotypes, 9 varieties are T / T homozygous genotypes, and 13 varieties are C / T heterozygous genotypes, while in the flat-topped inflorescence, 14 varieties are C / C homozygous genotypes, 11 varieties are C / T heterozygous genotypes, and 1 variety is T / T homozygous genotypes, that is, the homozygous genotype T / T is basically a spherical inflorescence, the homozygous genotype C / C is basically a flat-topped inflorescence, and the heterozygous genotype C / T has both spherical and flat-topped inflorescences. Among them, the homozygous genotypes corresponding to the spherical inflorescences at the three loci snp11, snp12, and snp4 are C / C, T / T, and G / G, respectively. It is also worth noting that the three loci snp10, snp11, and snp12 have a certain consistency in genotyping, that is, when a hydrangea plant is heterozygous at one locus, the other two loci are usually also heterozygous. Although the remaining SNP loci show polymorphism, they are not strongly associated with the trait of inflorescence type.

[0128] Table 2-550 hydrangea varieties typing results

[0129] Table 2-5 The genotyping results of the 50 hydrangea cultivars

[0130]

[0131] A: homozygous genotype 1; B: homozygous genotype 2; H: heterozygous genotype

[0132] A: homozygous genotype 1; B: homozygous genotype 2; H: heterozygous genotype

[0133] Table 2-6F 1 、F 2 The results of hydrangea plant typing

[0134] Table 2-6 The genotyping results of F1, F2 generation Hydrangea plants

[0135]

[0136] A: homozygous genotype 1; B: homozygous genotype 2; H: heterozygous genotype

[0137] A: homozygous genotype 1; B: homozygous genotype 2; H: heterozygous genotype

[0138] In the verification of 108 hybrid offspring, the verification results of 12 SNP loci were generally consistent with those in the variety verification, and the results of snp10, snp11, snp12, and snp4 were still better. For example, in the verification of hybrid offspring, 3 plants with spherical inflorescences were homozygous G / G genotypes, 21 plants with C / G heterozygous genotypes, and 19 plants with C / C homozygous genotypes; in the flat-top inflorescences, 20 plants were homozygous G / G genotypes, 20 plants were heterozygous C / G genotypes, and 5 plants were homozygous C / C genotypes. The homozygous genotypes G / G and C / C corresponded to the spherical and flat-top inflorescence types, respectively. That is, in these four SNP loci, two homozygous genotypes corresponded to the spherical and flat-top inflorescences, respectively, and the heterozygous genotypes contained two inflorescence types, and the number was close. In addition, the snp1 locus in Scaffold Hma1.2p1_0653F.1 showed rich polymorphism in the genotyping results of varieties and hybrid offspring, among which the G / G homozygous genotype appeared most frequently, and there were also four genotypes: A / G, A / A, G / T, and T / T. Except for G / G, the other four genotypes all showed flat-top inflorescences, and the G / G genotype included two flower types: spherical and flat-top (Table 2-7). In the typing results of all varieties and hybrid offspring, a total of 12 hydrangea plants were genotypes other than G / G. The overall verification results showed that the best results were obtained at sites on the three scaffolds Hma1.2p1_0060F.1 (snp10, snp11), Hma1.2p1_0653F.1 (snp1, snp2, snp12) and Hma1.2p1_0669F.1 (snp3, snp4) (i.e., including snp10, snp11, snp1, snp12).

[0139] Table 2-7 Snp1 typing results

[0140] Table 2-7 The genotyping results of Snp1

[0141]

[0142] Example 3

[0143] Construction of hydrangea flower type prediction model based on random forest algorithm

[0144] In this embodiment, the genotype data of the screened SNP sites are used as classification features, and the classification model is constructed using the random forest algorithm in the R language package. The resequencing typing data in the GWAS population and the validation population typing data are combined as a data set for subsequent analysis. First, the importance of a single marker is evaluated based on the matching degree of genotype and phenotype using the Gini coefficient and the Accuracy coefficient. Then, SNPs are used as candidate markers, and i markers are selected from the candidate markers according to the random forest algorithm in the R language to construct a model, and the accuracy of model predictions for different numbers of markers is analyzed respectively. The parameter is set to the number of trees as 1000 (ntree), and the mtry parameter is added to optimize the model.

[0145] The specific method is as follows:

[0146] The combination of resequencing typing data in the GWAS population and validation typing data in varieties and hybrid offspring was used as the training data set, and the rfImpute function was used to automatically fill in a small number of missing values. The Gini coefficient and Accuracy coefficient were used to evaluate the importance of each SNP site for hydrangea flower type traits. The results are shown in Figure 5 As shown, the Accuracy coefficient shows that the three loci snp11 (Hma1.2p1_0060F.1_1540773), snp1 (Hma1.2p1_0653F.1_460462), and snp8 (Hma1.2p1_1283F.1_78611) are more correlated with flower shape, that is, they are more important, while the Gini coefficient results show that the three SNP loci snp11 (Hma1.2p1_0060F.1_1540773), snp12 (Hma1.2p1_0653F.1_868484), and snp10 (Hma1.2p1_0060F.1_796640) are more correlated with flower shape traits.

[0147] The construction function screened these 11 SNP markers, and finally selected seven SNP markers, snp12, snp10, snp5, snp11, snp1, snp8, and snp3, for the subsequent hydrangea type prediction model construction based on the importance of each SNP site. For the seven selected SNP markers, the random forest algorithm was used to build the model, and 3 markers, 4 markers, and 5 markers were selected for model construction, respectively, to obtain the optimal model and parameter settings. Among them, the optimal model combination in the 3-marker model is snp12, snp8, and Snp 3, with an overall error rate of 14.78%, and the parameter settings are mtry=1, ntree=200. The classification diagram is shown in the figure. Figure 6 As shown in A. The optimal model combination results in the 4-marker model are the four SNP loci snp11, snp1, snp8, and snp 3. The overall error rate is 12.61%, and the model parameters are set to mtry=2, ntree=300. The model classification diagram is shown in Figure 6 B. The optimal model combination of the 5-marker model is snp10, snp11, snp1, snp8, and snp3. The overall error rate is 12.17%, and the model parameters are set to mtry=2, ntree=300. The optimal model classification diagram is shown in Figure 6 C.

[0148] From the results, it can be seen that the more SNP markers there are in the model, the higher the accuracy of the hydrangea flower type prediction. However, in actual application, the model with more markers will be more time-consuming and laborious to use. After the model is trained using a large amount of data as a training set, it will be able to output the predicted hydrangea flower type according to the different genotypes at different loci. In this study, the 3, 4, and 5 marker models have little difference in accuracy. In the subsequent use process, different numbers of models can be selected according to different situations and conditions to predict hydrangea plants with unknown flower types. The model relies on the model's own learning and judgment. After learning a large number of data sets, the algorithm will give the predicted hydrangea flower type based on the constructed model, without having to rely on the researchers' own judgment.

Claims

1. A method for developing a hydrangea-type KASP marker, characterized in that: The following steps are involved: The typing data of SNP loci related to flower type were obtained based on resequencing data through GWAS analysis; Filter the data in step (1) and select the -log 10 (P) The top 1 to 20 SNP sites with high values; In multiple hydrangea varieties, the SNP loci screened in step (2) were genotyped and verified by DNA sequencing and KASP methods, and finally the KASP markers for hydrangea flower type identification were confirmed; The KASP markers finally confirmed for hydrangea flower type identification are: snp1, snp3, snp8, snp10, snp11, and snp12; The snp1 is located at position 460462 of Scaffold Hma1.2p1_0653F.1, and the site is mutated from G to A; The snp3 is located at position 220821 of Scaffold Hma1.2p1_0669F.1, and the site is mutated from G to A; The snp8 is located at position 78611 of Scaffold Hma1.2p1_1283F.1, and the site is mutated from C to T; The snp10 is located at position 796640 of Scaffold Hma1.2p1_0060F.1, and the site is mutated from C to T; The snp11 is located at position 1540773 of Scaffold Hma1.2p1_0060F.1, and the site is mutated from G to C; The snp12 is located at position 868484 of Scaffold Hma1.2p1_0653F.1, and the site is mutated from A to T.

2. Application of a KASP primer combination in identifying hydrangea flower types, wherein the KASP primer combination uses a combination of snp12 primer, snp8 primer and snp3 primer, or a combination of snp11 primer, snp1 primer, snp8 primer and snp3 primer, or a combination of snp10 primer, snp11 primer, snp1 primer, snp8 primer and snp3 primer; The nucleotide sequences of the upstream primers F1 and F2 of the snp1 primer are shown in SEQ ID No. 1-2, and the nucleotide sequence of the downstream primer is shown in SEQ ID No. 3; The nucleotide sequences of the upstream primers F1 and F2 of the snp3 primer are shown in SEQ ID No. 4-5, and the nucleotide sequence of the downstream primer is shown in SEQ ID No. 6; The nucleotide sequences of the upstream primers F1 and F2 of the snp8 primer are shown in SEQ ID No. 10-11, and the nucleotide sequence of the downstream primer is shown in SEQ ID No. 12; The nucleotide sequences of the upstream primers F1 and F2 of the snp10 primer are shown in SEQ ID No. 13-14, and the nucleotide sequence of the downstream primer is shown in SEQ ID No. 15; The nucleotide sequences of the upstream primers F1 and F2 of the snp11 primer are shown in SEQ ID No. 16-17, and the nucleotide sequence of the downstream primer is shown in SEQ ID No. 18; The nucleotide sequences of the upstream primers F1 and F2 of the snp12 primer are shown in SEQ ID Nos. 19 to 20, and the nucleotide sequence of the downstream primer is shown in SEQ ID No.

21.

3. A hydrangea flower type prediction method, characterized in that: The specific method is as follows: the combination of resequencing typing data in the GWAS population and verification typing data in varieties and hybrid offspring is used as the training data set, and the rfImpute function is used to automatically fill in a small number of missing values; the Gini coefficient and Accuracy coefficient are used to evaluate the importance of each SNP site for hydrangea flower type traits, and candidate SNP markers are screened; the random forest algorithm is used to construct the model, and 3, 4, and 5 markers are selected from the candidate SNP markers for model construction, and the optimal model with different marker numbers is obtained, and the optimal model is used to predict the flower type of hydrangea plants; The marker combination of the 3-marker model in the optimal model is: snp12, snp8, snp3, the marker combination of the 4-marker model is: snp11, snp1, snp8, snp3, and the marker combination of the 5-marker model is: snp10, snp11, snp1, snp8, snp3; The snp1 is located at position 460462 of Scaffold Hma1.2p1_0653F.1, and the site is mutated from G to A; The snp3 is located at position 220821 of Scaffold Hma1.2p1_0669F.1, and the site is mutated from G to A; The snp8 is located at position 78611 of Scaffold Hma1.2p1_1283F.1, and the site is mutated from C to T; The snp10 is located at position 796640 of Scaffold Hma1.2p1_0060F.1, and the site is mutated from C to T; The snp11 is located at position 1540773 of Scaffold Hma1.2p1_0060F.1, and the site is mutated from G to C; The snp12 is located at position 868484 of Scaffold Hma1.2p1_0653F.1, and the site is mutated from A to T.

Images