Establishment and application of an SNP molecular identification system for grape cultivars

WANG Fu-qiang,FAN Xiu-cai,ZHANG Ying,SUN Lei,LIU Chong-huai,JIANG Jian-fu

Zhengzhou Fruit Research Institute,Chinese Academy of Agricultural Sciences,Zhengzhou 450009,P.R.China

Abstract We aimed to develop a set of single nucleotide polymorphism (SNP) markers that can be used to distinguish the main cultivated grape (Vitis L.) cultivars in China and provide technical support for domestic grape cultivar protection,cultivar registration,and market rights protection.A total of 517 high-quality loci were screened from 4 241 729 SNPs obtained by sequencing 304 grape accessions using specific locus amplified fragment sequencing,of which 442 were successfully designed as Kompetitive Allele Specific PCR (KASP) markers.A set of 27 markers that completely distinguishes 304 sequenced grape accessions was determined by using the program,and 26 effective markers were screened based on 23 representative grape cultivars.Finally,a total of 46 out of 48 KASP markers,including 22 markers selected by the research group in the early stage,were re-screened based on 348 grape accessions.Population structure,principal component,and cluster analyses all showed that the 348 grape accessions were best divided into two populations.In addition,cluster analysis subdivided them into six subpopulations.According to genetic distance,V.labrusca,V.davidii,V.heyneana,and V.amurensis were far from V.vinifera,while V.vinifera×V.labrusca and V.amurensis×V.vinifera were somewhere in between these two groups.Furthermore,a core set of 25 KASP markers could distinguish 95.69%of the 348 grape accessions,and the other 21 markers were used as extended markers.Therefore,SNP molecular markers based on KASP typing technology provide a new way for mapping DNA fingerprints in grape cultivars.With high efficiency and accuracy and low cost,this technology is more competitive than other current identification methods.It also has excellent application prospects in the grape distinctness,uniformity,and stability (DUS) test,as well as in promoting market rights protection in the near future.

Keywords:grape,KASP marker,variety identification,fingerprint,genetic diversity analysis

1.Introduction

The grape (VitisL.) is a perennial vine in the Vitaceae family,with at least 23 000 cultivars worldwide (http://www.vivc.de/) (Kong 2004).It includes several synonyms and homonyms owing to their strong ability for asexual reproduction,wide range of suitable planting,and frequent communication among grape accessions (Emanuelliet al.2013).In addition,the current grape seedling market management is not perfect,with random changes of variety names,hype cultivars,and other adverse effects occurring from time to time,greatly damaging the interests of breeders (Liet al.2018).Therefore,economical,efficient,and accurate variety identification is of great significance for improving management efficiency and the variety protection ability of grape accessions.

In recent years,DNA molecular markers have provided a new means for grape variety identification,owing to their short cycle,lack of environmental impact,and high-throughput detection (Emanuelliet al.2013).Among them,simple sequence repeat (SSR)and single nucleotide polymorphism (SNP) molecular markers are the two recommended marking methods,as they are both approved by the International Union for the Protection of New Cultivars of Plants (UPOV 2007) and theGeneral Guidelines for Identification of Plant Varieties Using DNA Markers(NY/T2594-2016 2016).Currently,a set of 30 SSR marker systems for the identification of Chinese grape cultivars has been developed in China (Wanget al.2020b),and eight of them have been used to distinguish 290 grape accessions in China (Liet al.2018).However,owing to the limited number of SSR markers and detection throughput,high detection cost,and time-consuming and labor-intensive data reading (Rasheedet al.2017),their use for the identification of a wider range of grape cultivars is restricted.As the latest generation of markers,SNP molecular markers provide the advantages of large numbers,dimorphism,and stable inheritance,among others (Liet al.2020),which may compensate for the technical defects of SSR markers coupled with the emergence of various high-throughput SNP detection platforms (Liuet al.2018).Moreover,they have been widely used in the fields of biology,agriculture,medicine and biological evolution (Yanget al.2019).

High-throughput SNP detection technologies mainly include gene chips and Kompetitive Allele Specific PCR(KASP;Laboratory of the Government Chemist (LGC),TW11 0LY,UK),which are suitable for different detection needs (Liuet al.2018).Gene chip technology is mainly suitable for small samples and multi-site detection,such as the rice SNP50 (Chenet al.2014),maize 50K (Xuet al.2017),and wheat 660K (Sunet al.2020) gene chips,which were independently developed in China.Its singlesite detection volume reaches tens of thousands,but the current cost of its use is relatively high,and development is difficult.KASP technology is suitable for the detection of multiple samples and few sites,and is efficient,flexible,accurate,and low-cost (Semagnet al.2013).Based on KASP technology,researchers in China and those working internationally have constructed the core marking systems of wheat (Awaiset al.2016;Grewalet al.2020),rice (Chenet al.2014),cotton (Kuanget al.2016),cabbage (Liet al.2020),cucumber (Zhanget al.2020),and other crops to achieve rapid variety identification.

Currently,for application of SNP markers in grape variety identification,high-cost gene chips and sequencing technologies are the main detection methods.For example,Cabezaset al.(2011) constructed an SNP identification system for Eurasian grape accessions and identified a 48-SNP set for identifying cultivars.Furthermore,Laucouet al.(2018) screened out 14 highly polymorphic SNP markers based on an 18K SNPs grape gene chip on 945 materials for the identification of cultivated species in Europe.It is worth noting that the internationally selected SNP markers are only used in diploid Eurasian species,while the main grapes grown in China include more polyploid cultivars.In China,Li B Bet al.(2019) and Lianget al.(2019) used different resequencing techniques to obtain tens of thousands of SNP loci,and used them to analyze the genetic diversity of 304 and 472 grape accessions,respectively.Although there were several pairs of polyploid grapes,the cultivars were genetically analyzed,but no specific results were selected to identify grape cultivars with the least combination of SNP loci.In the actual identification of grape cultivars,researchers strive to distinguish the most samples with the least number of sites to achieve efficient,accurate,and inexpensive identification.

A comprehensive comparison shows that KASP technology is more suitable for the initial development of SNP molecular markers than other technologies.Therefore,for the first time,we attempted to use KASP technology to screen a set of SNP markers that can identify Chinese grape cultivars,based on the grape SNPs developed previously with the main domestic grape cultivars and newly bred cultivars as test materials.This technology has excellent application prospects for the grape distinctness,uniformity,and stability (DUS) test,as well as promoting variety protection in the near future.

2.Materials and methods

2.1.Test materials

The test materials were collected from the Zhengzhou Grape Garden of National Fruit Tree Accessions,Zhengzhou Institute of Fruit Research,Chinese Academy of Agricultural Sciences,in June 2019.

For KASP marker preliminary screening,23 representative grape cultivars were used,which were of different species,ploidies,and uses (Table 1).KASP marker re-screening included 348 grape accession (Table 2).

Table 1 Information regarding the 23 representative grape (Vitis L.) cultivars used in the preliminary screening of Kompetitive Allele Specific PCR (KASP) markers

2.2.DNA extraction and detection

We extracted DNA from grape leaves by using a plant genomic DNA Extraction Kit (Aisen Biotechnology Co.,Ltd.,China),and detected the DNA concentration and purity(OD260nm/OD280nm=1.8-2.0) by using a NanoDrop 1000 spectrophotometer (Thermo Fisher Scientific,USA).DNA was diluted to 50 ng μL-1with ultrapure water,and loaded into 96-well PCR plates according to the number needed for high-throughput amplification detection (Liet al.2020).

2.3.KASP marker transformation and primer synthesis

To ensure the specificity of the selected SNP sites,the grape DNA sequence of 100 bp before and after each SNP site was downloaded from the Ensembl Plants website (https://plants.ensembl.org/Vitis_vinifera/Gene),which was compared in NCBI (https://blast.ncbi.nlm.nih.gov/Blast.cgi?PROGRAM=blastn&PAGE_TYPE=BlastSearch&LINK_LOC=blasthome) to remove non-specific sites.Then,SNP sites containing the specific DNA sequence were designed and synthesized by the Vegetable Improvement Center of Beijing Academy of Agriculture and Forestry Sciences,China,which established the KASP detection technology platform of the LGC Company.To ensure the quality of markers and annealing temperature consistency,the denaturation temperature of the two forward primer strands and the reverse strand should be around 62.5 and 66.0°C,respectively,the temperature difference between the front and back of the two strands should not exceed 1°C,and the GC contents should be 50-55% and ＜60%,respectively.In addition,to make full use of previous research progress,we also used the 22 high-quality KASP markers selected by the research group as part of the re-screening markers (Wanget al.2021).

2.4.PCR amplification and label detection

The PCR reaction system (10.14 μL) included 5 μL KASP Master Mix,0.14 μL KASP Marker Mix,and 5 μL template DNA (50 ng μL-1).The PCR reaction conditions are shown in Appendix A.The amplified product was detected by a fluorescent microplate detector,and detection data were read with the SNP viewer 2.0 Software developed by LGC Company (Zhanget al.2020).

2.5.Data statistics and analysis

Genetic diversity parameters were calculated based on the SNP typing result data,such as minor allele frequency(MAF),polymorphic information content (PIC) (Botsteinet al.1980;Zhanget al.2020),heterozygosity,andmissing rate,among others.All these calculations were preformed in Excel 2019 (Microsoft,USA):

wherePiandPjare the frequency of occurrence of the two SNP alleles in all tested accessions,andlis the number of samples (Zhanget al.2020).

According to the dimorphic characteristics of SNP markers,the typing data were converted into binary coding data,with the wild type (consistent with the grape genome),mutant,and heterozygous genes expressed as (1,0),(0,1),and (1,1),respectively,and missing sites recorded as (999,999)(Liet al.2020).Then,the neighborjoining algorithm of PowerMarker V3.25 Software was used to calculate the genetic distance of each grape accession(Zhaoet al.2018),and cluster maps were constructed with Figtree v1.4.4.

STRUCTURE 2.3.4 was used for Bayesian clustering (Falushet al.2003)and determining the best grouping.The allelic variation frequency characteristic number (the number of genetic populations)was set toK=1-10;the Burn-in period was 100 000,and the MCMC repeats number was set to 100 000.The mixed model and related allele frequencies were used for differentK.The value was repeated 15 times,and then the result file with the suffix ‘_f’ was compressed and uploaded to the“STRUCTURE HARVESTER”website (http://taylor0.biology.ucla.edu/struct_harvest/),according to Evannoet al.(2005).This method calculates the logarithmic function lnP(D) of ΔKand the likelihood value,modeling for the number of gene banks (K) to determine the bestKvalue.

GCTA Software was used for principal component analysis (Li Yet al.2019),MapChart 2.32 Software was used to draw a physical map (Voorrips 2002),and the Perl script provided by the Vegetable Improvement Center of Beijing Academy of Agriculture and Forestry Sciences was used to calculate the optimal combination marker (Zhanget al.2020).

3.Results

3.1.Screening of SNP

In 2015,our group used simplified genome sequencing technology (SLAF-seq) to sequence 304 grape accessions (the average sequencing depth of each sample was 8.09×),and developed a total of 466 618 SLAF tags,with 392 374 polymorphic SLAF tags.Through analysis of the polymorphic SLAF tags,a 4 241 729 SNPs population was obtained.From these,we screened out a total of 517 high-quality SNP loci.SNP sites were only conserved if (1) the DNA chain of the chromosome before and after the sequence was greater than 50 bp,and had a dimorphism and missing rate ＜0.05 (11 813 SNPs were screened out);(2) MAF＞0.05 (6 221 SNPs were screened out);(3) MAF＞0.3 (1 237 SNPs were screened out);(4) average sequencing depth ≥10× (616 SNPs were screened out);and (5) it was specific using 100 bp before and after the SNP locus by comparison with the grape reference genome (Appendix B).Through a statistical analysis of the genetic diversity information of 517 SNPs,PICs were all greater than 0.42,and 335 (64.80%) had a heterozygosity rate below 0.5 (Fig.1).

Fig.1 Genetic information map of the 517 high-quality SNP loci identified in this study.There is a total of seven circles,from the outside to the inside:a,names of 27 KASP markers;b,grape chromosome numbers;c,polymorphism information content (PIC);d,minor allele frequency (MAF);e,inbreeding coefficient;f,heterozygosity rate;g,missing rate.

To analyze the reliability of the 517 SNP loci selected for the 304 accessions used for sequencing,6 221 and 517 SNP genotypes were used to construct a phylogenetic tree (Fig.2).The results showed that based on two different numbers of SNPs,the 304 grape accessions could be divided into seven populations,with populations I and II corresponding mainly toV.vinifera×V.labrusca,and the remaining five populations corresponding mainly toV.viniferaand the wild oriental species population.Thus,the 517 SNP loci have a certain degree of reliability and representativeness.

Fig.2 Neighbor-joining (N-J) tree analysis based on 6 221 (A) and 517 (B) SNP sites of 304 grape accessions.

3.2.Conversion of KASP markers

The sequences of 100 bp before and after each of the 517 high-quality SNPs were extracted,and the primer design for KASP markers was carried out.Ultimately,442 SNPs could be successfully designed for KASP markers,for a conversion rate of 85.49%.

To save costs,we screened out only 27 SNPs(Fig.1-A) from the 442 SNPs to preferentially use for the synthesis of KASP markers.These 27 SNPs were evenly distributed on 19 chromosomes,and could completely distinguish the 304 sequenced grape accessions.

3.3.KASP marker screening

The 27 KASP markers were synthesized according to the designed primer sequence,with preliminary screening by 23 representative grape cultivars.As it was suggested that the PIC and MAF values of the VIT_4_7279466 KASP marker were both 0 (Table 3),i.e.,not polymorphic,this marker was discarded.The other 26 markers had PIC and MAF values greater than 0.3 and 0.2,indicating good polymorphism and suitability for further screening.

Table 3 Genetic information after initial screening of 27 markers using 23 representative grape (Vitis L.) cultivars1)

Therefore,a total of 48 KASP markers were used to genotype 348 grape accessions and test marker reliability and stability.In the end,KASP markers VIT_15_1706929 and VIT_16_13454358 were discarded because they had no good typing results when detecting the 348 grape accessions.The remaining 46 markers had good fluorescence typing results (Fig.3).Among them,89% had a deletion rate of less than 0.05,76%had a heterozygosity rate between 0.4 and 0.6,and 96% had MAF and PIC values greater than 0.3 and 0.4,respectively,all of which indicated they are highly polymorphic (Fig.4) and evenly distributed on 19 chromosomes (Appendices C and D).

Fig.3 KASP labeled fluorescence detection results of VIT_6_1032391 (A) and VIT_13_15071171 (B).

Fig.4 Genetic information content of 46 Kompetitive Allele Specific PCR (KASP) markers detected by 348 grape accessions.PIC,polymorphism information content;MAF,minor allele frequency.

3.4.Genetic analysis of main grape accessions

Group structure analysisIn this study,whenK=2,ΔKachieved the maximum value,and it was inferred that the 348 grape accessions could be divided into two groups(Fig.5-A and B).Pop1 and Pop2 accounted for 89.37%(311) and 10.63% (37) of the total population,respectively.Pop1 contained 210V.vinifera,93V.vinifera×V.labrusca,7V.amurensis×V.vinifera,and 1V.labrusca,while Pop2 contained 24V.labrusca,7V.vinifera×V.labrusca,3V.labrusca,1V.amurensis×V.vinifera,1V.davidii,and 1V.heyneana.

Principal component analysisPrincipal component analysis of the genotypes of 348 grape accessions detected by the 46 markers completely separated 311 of the population structure analysis group Pop1 and 37 Pop2 in a two-dimensional schematic diagram of PC1 and PC2(Fig.5-C),with the PC1 value of -0.07 as the boundary.

Genetic evolution cluster analysisBased on the typing results of 46 KASP markers,the neighbor-joining(N-J) evolutionary tree of 348 grape accessions was drawn (Fig.5-D),showing that the 348 accessions were highly consistent with the population structure analysis and principal component analysis,and best divided into two populations:Pop1 included 309 accessions,accounting for 88.79% of the total accessions,and Pop2 included 39,with two more accessions of ‘Zhuosexiang’(V.vinifera×V.labrusca) and ‘Kangzhen5’ (American population) than the Pop2 group based on just the population structure analysis.

Furthermore,to analyze the genetic relationship of the 348 grape accessions in more detail,we divided them into six subpopulations according to clustering results (Fig.5-D).

Fig.5 Population structure,cluster and principal component analysis of 348 grape accessions.A,the distribution of K-values with ΔK.B,population structure analysis when K=2.C,a PCA model score plot of the 348 grape accessions.D,neighbor-joining(N-J) tree cluster analysis.

Subpopulation I:Pop2 group,included 24 American population accessions such as ‘5BB’,‘SO4’,‘Kangzhen3’,which were rootstock accessions,except for ‘Concord’.The accessions ‘Gongzhuhong’,‘Zhuosexiang’,‘Compell Early’,‘Niagara’,and ‘Reliance’,beingV.vinifera×V.labrusca,could all be traced back to the genetic relationship with ‘Concord’ (V.labrusca),while the parental information of ‘Guizhoushuijing’ is unknown,but it showed the same genotype as ‘Nicaragua’,and is suspected to be synonymous.‘Huiliangciputao’ and‘Ziqiu’ (V.davidii),‘Guiheizhenzhu4’ (V.heyneana),and ‘Beibinghong’ (V.amurensis×V.vinifera) were also clustered in this group as populations that are distantly related toV.vinifera.

Subpopulation II-1 contained 6V.amurensis×V.viniferaand 69V.vinifera×V.labrusca.The cluster diagram showed that the 6V.amurensiswere far fromV.vinifera×V.labruscaaccording to genetic distances,and they each had a genetic relationship withV.amurensis.‘Gongniang1’,‘Beichun’,‘Beihong’,and‘Beimei’ are all hybrid offspring of ‘Muscat Hamburg’ andV.amurensis,which were closely clustered into a small branch in the cluster diagram.There were 63 accessions inV.vinifera×V.labruscagenetically related to ‘Kyoho’.Among them,‘Chunguang’ and ‘Miguang’ were crossbred from ‘Kyoho’ and ‘Zaoheibao’;‘Wuhezaohong’and ‘Yanpu1’ were cross-bred from ‘Kyoho’ and‘Zhengzhouzaohong’;‘Xiangyue’ and ‘Zizhenxiang’ were selected from the cross of ‘Shenyangmeiguixiangyabian’and ‘Zizhenxiangyabian’;‘Shennongshuofeng’ and‘Shennongxiangfeng’ were selected from the self-progeny of ‘Zizhenxiang’;‘Jingyou’,‘Jingya’,and ‘Hutai1’ were selected from the seedlings of ‘Black Olympia’,and the triploid ‘Zaoxiaxiang’,‘Sanbentiputao’,‘Chunxiangwuhe’,and ‘Runbaozaoxia’ sprouted from ‘Summer Black’.They were all closely related and clustered together separately.However,‘Huapuguixiang’,‘Huapuheifeng’,and ‘Huapumeigui’ of the Huapu series,and ‘Shenfeng’,‘Shenshuo’,‘Shenhua’,‘Shenxiu’,and ‘Shenyi’ of the Shen series were not exactly the same,resulting in different genetic distances among them.

Subpopulation II-2 contained 1 ‘Heiguixiang’ ofV.vinifera×V.labruscaand 32V.vinifera.The 32V.viniferawere all wine-making accessions,among which the clonal accessions of ‘Cabernet Sauvignon’,‘Syrah’,‘Merlot’,‘Cabernet Franc’,and ‘Riesling’ were all concentrated in this group.

Subpopulation II-3 included all diploids,containing 5V.vinifera×V.labruscaand 31V.vinifera.Among them,‘Jingxiangyu’,‘Ruiducuixia’,‘Ruiduhongmei’,and‘Ruiduxiangyu’ were all hybrid offspring of ‘Jingxiu’ and‘Xiangfei’,while ‘Fengbao’ and ‘Guifeimeigui’ were hybrid offspring of ‘Queen of Vineyard’ and ‘Hongxiangjiao’.They had close genetic distances and clustered together.In addition,‘Bixiangwuhe’ and ‘Jingyu’,both of which were related to the ‘Queen of Vineyard’,clustered together with the ‘Queen of Vineyard’ in this group.

Subpopulation II-4 contained 10V.vinifera×V.labrusca,1V.amurensis×V.vinifera,and 65V.vinifera.Among theV.vinifera×V.labrusca,‘Feicuimeigui’,‘Heixiangjiao’,and‘Hongshuangwei’ were the hybrid offspring of ‘Queen of Vineyard’ and ‘Hongxiangjiao’,while ‘Hongxiangjiao’ and‘Jixiang’,had ‘White Banana’ as one of the parents,being closely related.However,theV.viniferawere mainly divided into nucleated and non-nucleated varieties with‘Muscat Hamburg’ and ‘Thompsons Seedless’,respectively,as the backbone parent.‘Zhengzhouzaohong’,‘Zizhenzhu’,‘Xiangfei’,and ‘Zaomeigui’ (hybrid offspring of ‘Muscat Hamburg’ and ‘Pearl of Csaba’) and ‘Jingfeng’,‘Jingzijing’,and ‘Jingzaojing’ (related to ‘Pearl of Csaba’) clustered together.Moreover,the genotypes of ‘Zexiang’ and ‘Zeyu’were exactly the same and they clustered closely together.One of the parents was ‘Muscat Hamburg’,and ‘Lingzhen1’ofV.amurensis×V.viniferawas also clustered with them.

Subpopulation II-5 contained 7V.vinifera×V.labruscaand 82V.vinifera.Among theV.vinifera×V.labrusca,‘Qingfeng’ and ‘Zhengyanwuhe’ clustered together as the hybrid offspring of ‘Bronx Seedless’ and ‘Jingxiu’.TheV.viniferamainly had ‘Muscat Hamburg’,‘Red Globe’,‘Thompsons Seedless’,‘Pearl of Csaba’,and ‘Manicure Finger’ as their backbone parents.In particular,the five ‘bao’ series selected from the backbone parents of ‘Guibao,’ ‘Thompsons Seedless’,and ‘Muscat Hamburg’ were clustered in this group.‘Lümunage’ and ‘Hongmunage’ were both clones of‘Munage’;‘Baiwuhe’,‘Centennial Seedless’,‘Thompsons Seedless (Afghanistan)’,‘Thompsons Seedless(Ningxia)’ and ‘Zaowuhebai’ were all cloned accessions of ‘Thompsons Seedless’;and ‘Niunai’,‘Xuanhuaniunai’and ‘Changlibainianmanai’ were all cloned accessions of ‘Manai,’ and were clustered together.In addition,this group also clustered 23 Chinese local accessions in the last branch,indicating that the genetic distances of these Chinese local accessions were close toV.vinifera,and far from theV.labruscaand oriental wild populations.

3.5.Application of KASP markers in variety identification

To improve the identification ability of the KASP markers,the genotype data of 348 accessions tested by 46 KASP markers were analyzed for identification efficiency.The comparison of genotype data indicated that 11 groups(26 accessions) of the 348 accessions had the same genotype.Among them,five groups were synonymous germplasm (‘Niagara’ and ‘Guizhoushuijing’;‘Zexiang’and ‘Zeyu’;‘Xiying’,‘Baiwuhe’ and ‘Thompsons Seedless’;‘Zaomoli’ and ‘Molixiang’;‘Changlibainianmanai’,‘Xuanhuaniunai’,and ‘Manai’),three groups belonged to clonal accessions (‘Riesling’ and ‘Riesling237’;‘Cabernet Sauvignon ISV-FV5’,‘Cabernet Sauvignon R5’,and‘Cabernet Sauvignon A1’;‘110♀(Lanzhou)’ and ‘101-14♀’),two groups were budding types (‘Zaoxiaxiang’,‘Runbaozaoxia’,and ‘Sanbentiputao’ budding by ‘Summer Black’;‘Luopuzaosheng’ budding by ‘Jingya’),and one group had genetic characteristics of ‘Kyoho’ (‘Takatsuma’and ‘Pioneer’).After removing the consistent genotypic accessions,we completely distinguished 333 accessions with only 25 markers,and the identification efficiency reached 95.69% (Fig.6).

Fig.6 Cumulative KASP molecular marker efficiency for 333 grape accessions.

4.Discussion

As of the end of 2020,China’s grape planting area(785 000 ha) ranked the third worldwide (http://www.oiv.int/en/).As a concentrated distribution area of East Asian populations,China has at least 38 species and preserves more than 3 000 grape accessions (Kong 2004;Duanet al.2019),and the number of newly bred grape cultivars in China is increasing at a rate of at least 20 per year (Menget al.2017).The development of DNA fingerprinting is of great significance for rapid and accurate variety and purity identification,and serving the interests of breeders (Liet al.2020).

Recently,researchers have developed fingerprints based on SNP core markers in corn (Maet al.2019),cotton (Kuanget al.2016),cabbage (Liet al.2020),and other crops.For grape crops,some researchers have also developed a set of core SNP markers to identify grape cultivars (Cabezaset al.2011).However,the cost of the SNP detection technology they used is generally high.In addition,there are manyV.vinifera×V.labruscaand local grape accessions in China that are polyploid.Therefore,foreign standards are not fully applicable to the identification of grape cultivars in China.Additionally,there is no known benefit of using SNP markers to construct fingerprints in Chinese grape research (Wanget al.2020a).

To develop an SNP fingerprint,first,we must have an SNP population suitable for developing the SNP markers.This study was based on the 4 241 729 SNPs population by sequencing 304 grape accessions with SLAF-seq,and by strictly controlling the screening conditions,we successfully screened 517 high-quality SNPs.Based on these 517 SNPs and 304 grape accessions,the phylogenetic trees constructed using 6 221 SNPs were highly consistent,which fully guaranteed the quality and representativeness of the SNP sites used for the identification of the constructed grape accessions.

Second,choosing the appropriate SNP typing technique is essential for improving identification efficiency and reducing detection costs.Currently,there are at least 20 SNP typing technologies.Among them,Semagnet al.(2013) compared and analyzed KASP and Illumina GoldenGate technologies from the perspectives of conversion rate,accuracy,cost,and flexibility using corn as the crop material.KASP has greater advantages when a small number of sites is needed to detect a large sample population.Zhaoet al.(2017) used four SNPs to compare the three typing methods of sequencing,cleaved amplified polymorphic sequence (CAPS),and KASP based on 53 soybean materials,and also stated that KASP technology is more flexible and practical.Liuet al.(2018) also concluded that KASP technology can achieve the purpose of high-throughput detection,and is more suitable for crop variety identification than other technologies.Moreover,Li P Ret al.(2019),Liuet al.(2019),Maet al.(2019) and Liet al.(2020) selected 50,48,22 and 59 core KASP markers based on KASP technology for identifying cabbage,tomato,maize and cabbage cultivars,respectively.

Therefore,based on the several advantages of KASP technology,this study also used KASP as technical support for grape accession genotyping.However,unlike previous studies,it ingeniously greatly reduced the cost of developing KASP tags.Instead of synthesizing the 442 KASP markers successfully designed to verify the typing effect,based on the genotypes of 304 sequenced accessions corresponding to 442 markers,a set of only 27 that could completely distinguish the 304 accessions was calculated using Perl program priority verification of marker combinations,which directly reduced the cost of synthetic primers by 94%.In addition,to make full use of previous research progress,22 KASP markers selected from 61 high-quality SNPs selected by a research group from abroad were used in the KASP marker screening library.Finally,we selected 46 KASP markers from the 49 KASP markers for grape variety identification using 23 representative cultivars and 348 grape accessions as test samples.

To verify the reliability of the 46 KASP markers,we conducted genetic diversity analysis.First,based on the detection of 348 grape accessions,the MAF and PIC values of the 46 KASP markers all reached the standard of high polymorphism (Liet al.2020).Second,population structure analysis,principal component analysis,and genetic evolution analysis showed that the 348 grape accessions can be classified into two major populations.The classification results were highly consistent,fully demonstrating that the 46 KASP markers have a certain degree of reliability.

Furthermore,to verify marker accuracy,we compared the genetic evolution analysis results with the pedigree relationships of the 348 grape accessions.In this study,the 348 grape accessions could be divided into six subpopulations.Among them,allV.labrusca(25),V.davidii(3),V.heyneana(1),andV.amurensis(1)belonged to cluster I;77.00% ofV.vinifera×V.labrusca(77) and 87.50% ofV.amurensis×V.vinifera(7)belonged to cluster I and II-1,and allV.vinifera(210)were clustered in II-2,II-3,II-4,and II-5.According to genetic distance,V.labrusca,V.davidii,V.heyneana,andV.amurensiswere far fromV.vinifera,whileV.vinifera×V.labruscaandV.amurensis×V.viniferawere somewhere in between,which was consistent with the conclusions of Li B Bet al.(2019) and Lianget al.(2019)using different resequencing techniques to analyze the population relationships of 304 and 472 grape accessions,respectively.In addition,23 local accessions cultivated in China belonged to cluster II-5 with the grape accessions ofV.vinifera,indicating that some local accessions in China were closely related to the accessions ofV.vinifera.This confirmed the conclusion of Liet al.(2017) and the grape accession research on 61 local accessions in China and 33V.viniferabased on SSR molecular markers that table grapes are continuously spreading from Europe to China,which is the same as the historical hypothesis(Kong 2004).Therefore,it was fully proven that the 46 KASP markers were reliable,effective,and accurate in detecting 348 grape accessions.

To further improve the identification efficiency and accuracy of the markers and reduce detection cost,we tested the genotypes of 348 accessions based on the 46 KASP markers and analyzed their identification efficiency.Among the 11 groups of accessions with the same genotype,five groups belonged to clonal accessions,two groups were budding types,and one group had genetic characteristics of ‘Kyoho’.The genetic difference of its own accession was very small,and the phenotypes also showed small differences,so it was not possible to achieve identification with only 46 markers.After removing the accessions of the same genotype,only 25 markers could completely distinguish the remaining 333 accessions,and the identification efficiency reached 95.69%,which basically met the criteria for a set of core KASP markers.

These 333 grape accessions are only a small part of the total grape accessions.To distinguish between wild populations with rich genetic characteristics and the large numbers ofV.viniferapopulations more SNP markers are necessary,with greater discriminatory ability.We only verified 27 of the 442 KASP markers,although verifying the remaining markers could have screened out more high-quality markers.However,these 333 accessions basically include table,wine,raisin,and rootstock grape accessions that are the main cultivars grown internationally,as well as cultivars newly bred in China in recent years.Furthermore,if SSR and SNP molecular markers are used comprehensively,the identification capabilities would be improved effectively,and this enhancement needs to be further studied.

5.Conclusion

A set of SNP molecular identification system was constructed for grape cultivars,including 46 high-quality KASP markers,which could provide a scientific basis and technical support for the protection of new grape cultivars and variety identification.

Acknowledgements

The financial support for this research provided by the National Key R&D Program of China (2019YFD1001401),the China Agriculture Research System of MOF and MARA (CARS-29-yc-1),and the Agricultural Science and Technology Innovation Program of Chinese Academy of Agricultural Sciences (CAAS-ASTIP-2017-ZFRI) is gratefully appreciated.

Declaration of competing interest

The authors declare that they have no conflict of interest.

Appendicesassociated with this paper are available on http://www.ChinaAgriSci.com/V2/En/appendix.htm