Development of KASP Markers for Carnation Germplasm (or Cultivar) Fingerprinting Based on Reduced-Representation Genome Sequencing Technology

TL;DR

This study develops KASP markers for carnation variety identification using genome sequencing, enabling accurate fingerprinting and resolving naming conflicts.

Contribution

The novel use of reduced-representation sequencing to develop high-accuracy KASP markers for carnation fingerprinting.

Findings

82,584 high-quality SNPs were identified from 50 carnation accessions.

45 core KASP markers achieved 99.987% distinguishing power for 309 accessions.

A fingerprint database was constructed to support variety protection and identification.

Abstract

Carnation is one of the most popular ornamental flowers worldwide. Due to its high ornamental and economic value, breeding techniques have advanced rapidly, leading to the continuous emergence of new varieties. However, this has also resulted in issues such as synonymy and homonymy. Therefore, utilizing DNA fingerprinting for rapid and accurate variety identification can play a crucial role in germplasm identification and the resolution of intellectual property disputes. In this study, we performed reduced-representation genome sequencing on 50 carnation accessions to develop single nucleotide polymorphism (SNP) markers. After filtering, 82,584 high-quality SNPs were obtained. These SNPs were used to conduct principal component analysis, population structure analysis, and cluster analysis on the 50 carnation accessions. From these high-quality SNPs, 130 SNP loci were further selected and converted into Kompetitive Allele-Specific PCR (KASP) markers. Preliminary screening using 92 carnation accessions yielded 53 KASP markers, and a subsequent screening with 217 carnation accessions identified 45 core KASP markers. Using these core markers, a fingerprint database was successfully constructed for 309 carnation accessions, achieving a distinguishing power of 99.987%. This study employed SNP fingerprinting and genetic analysis for the screening and identification of carnations, broadening the genetic basis at the molecular level and supporting subsequent variety protection efforts, thereby providing a scientific basis for carnation selection and identification.

Figures

Full text

1. Introduction

Dianthus caryophyllus L., commonly known as carnation, is a perennial herbaceous plant belonging to the family Caryophyllaceae and the genus Dianthus L. Renowned for its diverse flower colors and forms, it ranks among the world’s most important ornamental flowers, widely used as cut flowers, potted plants, and in landscaping [1]. Beyond its ornamental value, carnation also holds significant importance in culinary and medicinal applications [2]. With a cultivation history spanning over 2000 years, carnation is primarily distributed across the temperate regions of Europe and Asian countries such as China, Japan, Korea, and Malaysia, originating from the Mediterranean region [3,4]. The rapid development of the carnation industry has led to a substantial increase in new varieties. To incentivize breeding innovation, China included the genus Dianthus L. in the first batch of the “Protected List of Agricultural Plant Varieties in the People’s Republic of China” on 27 April 1999. In recent years, with growing market demand and the substantial economic benefits associated with superior new varieties, issues such as “synonymy” (different names for the same variety), “homonymy” (the same name for different varieties), and counterfeit or mislabeled varieties have become increasingly prevalent in carnation production. These practices severely infringe upon the rights of breeders and undermine motivation for independent breeding innovation.

To facilitate effective variety protection, molecular marker identification techniques have been successively established for various crops. Molecular markers are genetic markers based on nucleotide sequence variations in genetic material among individuals, directly reflecting genetic polymorphism at the DNA level. Commonly used molecular markers include Simple Sequence Repeats (SSR), Random Amplified Polymorphic DNA (RAPD), and Single Nucleotide Polymorphisms (SNP) [5]. SNP represents the third-generation DNA molecular marker technology. It offers multiple advantages, including abundant occurrence, diverse sources of variation, compatibility with database integration and data sharing, and strong associations with functional genes and even plant phenotypes. Consequently, SNPs have been widely applied in genetic diversity analysis, characterization of genetic resources, and variety identification across various crops [6,7].

Numerous methods are available for SNP genotyping. Gene chip technology is primarily suitable for small samples and multiple loci, yet it currently entails high costs and poses development challenges [8]. Whole-genome sequencing may represent the optimal method for identifying distinct varieties, offering the highest accuracy for identification; however, the cost of library construction is higher compared to other methods [9]. Genotyping by sequencing (GBS) offers lower library construction costs, but the distribution of markers is less uniform than with other approaches [10]. The TaqMan probe method is a real-time PCR detection technique utilizing fluorescent DNA probes based on the 5′ to 3′ exonuclease activity of Taq polymerase, which requires locus-specific probes [11]. In contrast, Kompetitive Allele-Specific PCR (KASP) is a fluorescence-based genotyping technology characterized by low primer synthesis costs and does not require locus-specific probes [12]. Compared to other genotyping methods, KASP technology is particularly suitable for detecting a small number of loci across a large number of samples, offering advantages in terms of labor, time, cost, distribution, and flexibility, making it more appropriate for variety identification [9,13]. Based on these advantages, researchers have widely applied KASP technology to crops such as maize [14], cigar tobacco [15], cowpea [16], rice [17], and others to achieve rapid variety identification.

Most carnation varieties are diploid, with a chromosome number of 2n = 2x = 30 [18]. Research on molecular markers in the genus Dianthus L. has primarily focused on the development of RAPD and SSR markers, and their application in kinship identification, polymorphism analysis, genetic linkage map construction, and QTL mapping. Smulders et al. [19] developed 11 pairs of SSR primers, five of which exhibited good polymorphism among 26 varieties. Kimura et al. [20] designed 13 pairs of SSR primers capable of distinguishing 39 varieties. Butiuc-Keul et al. [21] utilized five developed SSR primers to analyze the genetic diversity of five endangered, rare, and endemic carnation varieties in Romania. Wang Yongfeng [22] obtained 100 pairs of SSR primers from transcriptome sequencing of Dianthus chinensis and ultimately screened 33 highly polymorphic primers for polymorphism analysis of 28 varieties, three of which could distinguish all 28 varieties. Yagi et al. [23] constructed a genetic linkage map for carnation comprising 145 RAPD markers, 201 SSR markers, and two sequence-tagged site loci [4]. Although the application of molecular markers in carnation has proven valuable, research on SNPs as third-generation molecular markers for carnation variety identification remains insufficient. The whole-genome sequencing and assembly of the carnation cultivar ‘Scarlet Queen’ (D. caryophyllus) by Zhang et al. [1] provides a foundational genomic resource for molecular marker development in this study.

In this study, commercially available carnation varieties were collected and initially categorized into two groups based on plant type: spray type and non-spray type [24]. The spray-type varieties are all multi-flowered with relatively short plant height, while the non-spray types are taller and were further divided into single-flower and multi-flower varieties based on flower number [25]. Consequently, the collected varieties were classified into three categories: spray-type, potted varieties (code: PZ), standard single-flower varieties (code: QS), and standard spray-flower varieties (code: QM). We then employed reduced-representation genome sequencing to screen for high-quality SNP loci and used KASP technology to develop a core set of SNP markers for carnation variety identification and the construction of an SNP fingerprint database. This work is expected to enhance the management and utilization efficiency of carnation germplasm resources, provide a valuable source of genetic information, and establish a scientific basis for the screening and identification of carnation germplasm.

2. Results and Analysis

2.1. Phylogenetic, Population Structure, and PCA Analyses

To screen for suitable SNP molecular markers for conversion to KASP assays, we performed reduced-representation genome sequencing on representative varieties using GBS technology and conducted genetic diversity analysis. Reduced-representation genome sequencing of the 50 representative carnation varieties generated 36.66 Gb of raw data, with an average Q30 of 88.54% and a GC content ranging from 40.86% to 42.10%. After high-quality filtering and adapter trimming, 35.02 Gb of clean data were obtained for further analysis. Detailed statistics of the sequencing data quality are provided in Supplementary Table S1. The clean reads were aligned to the reference genome, yielding mapping rates between 92.87% and 96.34% across all samples, with an average sequencing depth of 61.47× and genome coverage ranging from 1.05% to 2.51%. Following variant calling and stringent filtering, a total of 82,584 high-quality SNP loci were identified and subsequently used for genetic diversity analysis. The phylogenetic tree (Figure 1A) revealed that the 20 spray-type varieties clustered together, and the 21 standard single-flower types formed a distinct clade. With the exception of variety QM-13, the remaining standard spray-flower types also grouped into a single cluster. Although QM-13 is classified as a standard spray-flower type, it possesses a unique and rare combination of traits—fragrance and single-petal flowers—which likely explains its distinct phylogenetic position.

Principal component analysis (PCA) yielded results largely consistent with the phylogenetic tree. The first three principal components (PC1, PC2, and PC3) explained 17.78%, 10.00%, and 6.69% of the total genetic variance, respectively, with a cumulative contribution of 34.47%. This reflects the complex genetic structure of the population and the presence of multiple sources of variation, indicating that the first three PCs did not fully capture all genetic variation (Figure 1B). Regarding clustering patterns, the PZ group formed a tight cluster in the PCA space, suggesting the lowest level of intra-group genetic differentiation and the highest genetic homogeneity. With the exception of one extreme outlier, the samples from the QM group showed a high degree of overlap in distribution with the QS group, indicating considerable genetic similarity between these two groups. In contrast, the QS group exhibited the widest distribution along the PCA axes, reflecting the highest level of intra-group genetic diversity. At the inter-group level, the PZ group was clearly separated from both the QM and QS groups in the PCA plot, demonstrating significant genetic differentiation between the PZ group and the latter two. Conversely, the closer genetic relationship observed between the QM and QS groups suggests that they share a more recent common ancestor or have experienced frequent historical gene flow.

Population structure was analyzed using ADMIXTURE with K values ranging from 1 to 10. The optimal number of subpopulations (K) was determined as K = 2, which exhibited the lowest cross-validation error (CV error) (Figure 1C), indicating that the 50 varieties can be primarily divided into two subgroups. In the corresponding population structure bar plot at K = 2 (Figure 1D), the 50 accessions were clearly assigned to two major genetic clusters (represented by blue and pink), demonstrating a concise and well-defined population division. The PZ group was predominantly associated with the blue ancestral component, while the QM and QS groups were primarily associated with the pink component. This genetic clustering aligns perfectly with the phenotypic classification based on plant type (spray type/non-spray type). Therefore, the genetic diversity analysis confirms the accurate classification of our collected 50 varieties. Furthermore, these 50 varieties exhibit rich phenotypic variation in key traits such as plant type, flower form, and color series (see Table 1). Collectively, these results demonstrate that the selected 50 varieties are representative, and the SNP markers identified from their sequencing data are suitable for KASP marker development in carnation.

2.2. Development of KASP Markers

To successfully convert SNP markers into KASP markers, we used 309 carnation varieties for marker screening and validation. Based on the established criteria for KASP marker development (detailed in Section 4.3.1), 130 high-quality and highly polymorphic SNP loci were selected from the 82,584 SNPs used for the genetic diversity analysis. The initial screening yielded 53 pairs of KASP primers. The initial screening yielded 53 KASP primer pairs. The genotyping results of these 53 SNP loci in the 92 varieties showed clear clustering into three distinct groups—representing two homozygous clusters and one heterozygous cluster (Figure 2). These 53 KASP primer pairs were then subjected to a secondary screening. Eight primer pairs with unstable amplification, high missing data rates, or low PIC values were excluded, resulting in 45 KASP primer pairs retained for variety identification (primer information is provided in Supplementary Table S3). These 45 final markers, distributed across all 15 chromosomes (Figure 3), exhibited the following characteristics in the 309 accessions (including all varieties from both screening stages): Major Allele Frequency ranging from 0.49 to 0.74, heterozygosity from 0.14 to 0.57, polymorphism information content (PIC) from 0.32 to 0.41, and missing rates between 0% and 5.50% (detailed parameters are provided in Supplementary Table S4).

2.3. Application of KASP Markers

To evaluate the effectiveness of the 45 KASP markers, we constructed a molecular fingerprinting database and analyzed their distinguishing power. The genotyping results of the 45 KASP loci across 309 carnation varieties were visualized to construct a DNA fingerprint map (Supplementary Table S5). In this map, nucleotides C/C, A/A, T/T, and G/G are represented by yellow, green, blue, and purple, respectively; missing data are displayed in gray; and heterozygous sites are shown in white. A total of 47,586 pairwise combinations were generated from the 309 varieties. Among these, only 6 pairs showed zero differential sites, resulting in a distinguishing rate of 99.987%. When considering combinations with one or fewer differential sites, the number increased to 8 pairs, corresponding to a distinguishing rate of 99.983%. For combinations with two or fewer differential sites, 14 pairs were identified, yielding a distinguishing rate of 99.971%. The distribution of pairwise combinations according to the number of differential sites is shown in Figure 4.

To validate the effectiveness of our core markers in cluster analysis, a phylogenetic tree of the 309 carnation varieties was constructed using the unweighted pair group method with arithmetic means (UPGMA) based on Nei’s genetic distance. The 309 varieties were divided into two major groups (Figure 5). Group I consisted of 141 spray-type varieties, while Group II comprised 168 non-spray varieties. Within Group II (non-spray varieties), most standard spray-flower types and standard single-flower types formed distinguishable clusters, with only a few varieties scattered between them. Furthermore, varieties originating from the same breeder or institution tended to cluster together, indicating close genetic relationships. Examples include Zilan and Ziwu (accession codes PZ21 and P18), Fen Feige and Feige (QM3 and QM7), and P21394 and P21308 (P20 and PZ38). Varieties with minor and indistinguishable phenotypic differences in the field also clustered closely, such as Ruibu and Zibian (QS10 and Q105), and P21430 and P21131 (P43 and P50).

3. Discussion

GBS is a genetic sequencing approach that utilizes SNPs for genotyping studies. As a form of reduced-representation sequencing, it offers advantages such as rapid library preparation for large sample sets and reduced genome complexity [26]. To date, high-quality genomic resources for carnation remain relatively scarce, and the publicly available carnation genomes are all derived from single individuals, which limits their ability to represent the full spectrum of genetic diversity within the species [27]. Super-genotyping by sequencing (Super-GBS), an advanced development based on GBS, demonstrates superior performance over conventional GBS in terms of SNP marker reproducibility, stability, and genotyping accuracy [28]. Therefore, in this study, we applied Super-GBS to perform reduced-representation genome sequencing on 50 representative carnation varieties, successfully identifying 82,584 high-quality SNP loci for subsequent genetic diversity and population structure analysis.

The genetic structure of carnation germplasm can effectively reveal the genetic relationships among tested materials. In the SNP-based phylogenetic tree constructed in this study, with the exception of the single-petal and fragrant variety QM-13, the remaining non-spray types (standard spray-flower and standard single-flower) and spray-type varieties formed distinct clusters, which shows some correlation with their phenotypic traits, particularly plant type. This clustering pattern was further supported by principal component analysis. Interestingly, all standard single-flower varieties are large-flowered carnations, while all standard spray-flower varieties are small-flowered carnations. Moreover, the standard spray-flower varieties were genetically closer to the spray-type varieties than to the standard single-flower types. This is likely attributed to the fact that spray-type varieties are dwarf variants derived from standard spray-flower types, a finding consistent with the classification based on flower types reported by Jiang et al. [29]. Furthermore, population structure analysis revealed that the tested carnations could be divided into two distinct subpopulations: spray-type and non-spray type.

The development of molecular markers for variety identification requires careful selection of a representative sample set of appropriate size. For instance, Wang et al. [15] developed KASP markers based on sequencing data from 113 cigar tobacco accessions; Li et al. [30] identified 50 core KASP markers using resequencing data of 50 cabbage inbred lines; and Li et al. [31] developed 143 KASP markers based on whole-genome resequencing of 121 sweet osmanthus cultivars. The selection of samples for marker development depends on the complexity of the species’ genome and the type of markers targeted. In this study, the 50 varieties used for KASP marker development encompassed diverse cultivation types (spray, standard single-flower, and standard spray-flower), flower forms (double, semi-double, and single), and color series (red, orange, green, pink, purple, white, and yellow), to ensure representativeness. Furthermore, the phylogenetic tree constructed for 309 carnations using the 45 core KASP markers exhibited a clustering pattern consistent with that generated from the SNP markers derived from the sequencing data of the 50 varieties. Collectively, this confirms the suitability of the 50 accessions for KASP marker development.

DNA fingerprinting, characterized by its high polymorphism, serves as a powerful tool for variety identification and has been widely applied in the authentication of variety resources across numerous crops. Commonly used molecular markers include SSRs and SNPs. However, SSR markers present limitations in practical applications, such as restricted marker numbers and detection throughput, certain mutation rates at loci, time-consuming and labor-intensive data interpretation, and limited effectiveness in distinguishing somatic mutants [32]. In contrast, SNPs offer higher polymorphism levels and, since they do not rely on fragment length variations like SSRs, are more amenable to standardization across different laboratories and platforms [33]. Currently, researchers have established SNP-based fingerprinting systems with core marker sets in various crops, including maize [14], grape [34], and cowpea [16]. In this study, we employed KASP technology to genotype SNP loci in carnation and thus constructed a fingerprint database for 309 carnation accessions using 45 core SNP markers. Among these, two pairs of suspected synonymous varieties—Ruibu and Zibian (QS10 and Q105), and P21430 and P21131 (P43 and P50)—exhibited zero differential sites, suggesting they may be closely related or represent cases of synonymy (different names for the same genotype). Conversely, three pairs of varieties sharing the same name—Hua’erzi (P75 and QM37), Lücha (P72 and QM60), and Xiangbin (QS35 and P73)—showed considerable molecular differences, along with distinct phenotypic traits and plant types, indicating they likely represent instances of homonymy (the same name for different genotypes). Therefore, the developed set of 45 KASP markers effectively identifies genetically similar and phenotypically analogous varieties and can be directly utilized for variety genotyping and identification.

The scope of material collection and the number of developed KASP markers were relatively limited, which imposes certain constraints on the genetic analysis. In the cluster analysis of 309 carnation varieties based on the 45 KASP loci, most varieties could be clearly clustered according to their PZ, QM, and QS groups. This demonstrates the effectiveness of this marker set in reflecting varietal genetic differentiation and the genetic rationality of grouping based on the phenotypic trait of plant type. However, cross-group clustering was observed in a few varieties. For example, the standard single-flower variety QS4 clustered within the spray-type group. This phenomenon may be attributed to insufficient marker coverage failing to capture the unique genetic characteristics of certain samples [35]. Undeniably, the current common principle for selecting molecular markers for variety identification remains to adopt a minimal marker set to maximize sample discrimination. For instance, Yang et al. [36] developed 41 core KASP markers to distinguish 329 cauliflower varieties, and Xing et al. [37] constructed a molecular fingerprint database for 356 radish varieties using 32 core SNP markers. In the future, as breeding efforts yield an increasing number of varieties, incorporating more KASP markers associated with phenotypic traits could be beneficial. This expansion is expected to enhance the identification rate to meet the demands of researchers [9].

Under special circumstances, molecular marker identification can be combined with phenotypic characterization. In practical carnation production, where propagation is mainly carried out through cuttings, somatic mutants are most likely to be detected during this process. Color mutations are the most easily noticeable, and in practice, such mutants are often discovered by parties other than the original breeder. If it can be demonstrated that such a mutant is clearly distinct from all existing varieties, including the original one, the discoverer may apply for plant breeders’ rights [38]. Carnation varieties that are genetically very close and only differ in flower color may qualify as essentially derived varieties (EDV). Such varieties might show zero differential sites across the 45 KASP markers. Judging whether a variety is essentially derived requires the use of high-density molecular markers; the inability to distinguish EDV with only 45 loci is expected. For varieties that cannot be differentiated using SNP markers, DUS (Distinctness, Uniformity, and Stability) testing through field investigation of phenotypic traits may still allow differentiation based on certain characteristics. Therefore, combining both methods can effectively improve the accuracy of variety identification. Furthermore, with advances in breeding techniques and the increasing number of bred varieties, increasing the number of SNP markers—particularly those closely linked to important agronomic traits—may be necessary to enable more accurate and efficient variety identification in the future.

This study employed reduced-representation genome sequencing to develop SNP markers from 50 carnation accessions, yielding 82,584 high-quality SNPs. Based on these markers, principal component analysis, population structure analysis, and cluster analysis were conducted. A subset of 130 SNPs was selected and converted into KASP markers, from which 45 core KASP markers were ultimately identified through initial and repeated screening processes, demonstrating a distinguishing power of 99.987%. Using these core markers, a fingerprint database was successfully constructed for 309 carnation accessions, enabling effective variety identification. This work not only provides a reliable molecular basis for genetic analysis, selection, and identification of carnations but also lays an important foundation for subsequent germplasm conservation and breeding efforts.

4. Materials and Methods

4.1. Plant Materials

A total of 309 carnation varieties, covering spray types, standard single-flower types, and standard spray-flower types (Supplementary Table S2), were selected for SNP marker development. Among them, 50 representative varieties, comprising 20 spray types, 21 standard single-flower types, and 9 standard spray-flower types, were used for sequencing (Table 1) (Figure 6). The expression states of these 50 varieties for key grouping traits, including plant type, flower form, and primary petal color, encompassed all corresponding states described in the “Guidelines for the conduct of tests for distinctness, uniformity and stability: Carnation (D. caryophyllus)” [24]. Initial screening of SNP loci was conducted using 92 varieties, which included the 50 sequenced ones. The remaining 217 varieties were then employed for primer re-screening and validation. Following the identification of core primers, a DNA fingerprint database was established for all 309 varieties. In 2023, seedlings of 309 carnation varieties collected from the Flower Research Institute of Yunnan Academy of Agricultural Sciences, Kunming Binfen Horticulture Co., Ltd. (Kunming, China), and Kunming Dounan Flower Market were cultivated in Kunming, Yunnan Province. The plants were grown in pots measuring 17 cm × 20 cm, containing a 1:2 mixture of red soil and substrate. Planting spacing was set at 15–20 cm between plants and 20 cm between rows for standard types, and 20–25 cm between plants and 30 cm between rows for spray types. At 3–4 weeks after transplanting, a single pinching was carried out, retaining 4–6 nodes. For single-flower types, the main bud was retained and lateral buds were removed; for spray-flower types, the main bud was removed when it reached 0.3 cm in diameter, while lateral buds were retained. The optimal growth temperature was maintained at 15–25 °C, with nighttime temperatures not falling below 10 °C; the suitable temperature during the flowering period was 18–25 °C. A photoperiod of 14–16 h per day was ensured, and the relative air humidity was maintained at 60–70%.

4.2. Genetic Diversity Analysis

4.2.1. DNA Extraction and Library Preparation

Genomic DNA was extracted from young leaf tissues of the 50 carnation varieties using a Plant Genomic DNA Kit (Tiangen Biotech (Beijing) Co., Ltd., Beijing, China). To comprehensively represent the population genotype and minimize the influence of potential off-type or contaminated individuals, leaves from five individual plants per variety were pooled as one sample for DNA extraction. The quality and concentration of the DNA were assessed by electrophoresis and a high-throughput ultra-micro spectrophotometer (Implen NanoPhotometer^®^ N120, München, Germany). DNA samples passing quality control were used for sequencing library construction. The DNA was digested with the restriction enzymes PstI-HF and MspI, and the digested products were purified using a magnetic bead-based method. Fragments ranging from 300 to 700 bp were selected and recovered. The purified products were then subjected to PCR for adapter ligation and subsequently sequenced on an Illumina Nova platform (PE150) (Illumina, San Diego, CA, USA).

4.2.2. Genome Assembly and Genetic Data Analysis

After obtaining the raw sequencing data, the demultiplexing of reads by barcode and restriction site was performed using the process_radtags module in Stacks software (version 2.4), yielding raw reads for each sample [39]. To mitigate the impact of sequencing errors, quality control and filtering of the raw reads were conducted using fastp software (version 0.20.0) to obtain clean reads [40]. The key parameters for fastp were: —n_base_limit 5, —cut_window_size 4, —cut_mean_quality 20, —length_required 75, —qualified_quality_phred 15. The clean reads were then aligned to the reference genome using BWA software (version 0.7.17), and the mapping rate was calculated to assess the similarity between each sample and the reference genome [41]. For the reference genome, the access link is: “https://www.ncbi.nlm.nih.gov/datasets/genome/GCA_023091065.1/ (accessed on 1 July 2023)”. Based on the alignment results, variant calling for SNPs and InDels was performed using the HaplotypeCaller module in GATK software (version 4.1.3.0) [42]. The key steps and parameters for the variant calling pipeline were as follows: BaseRecalibrator (default parameters), ApplyBQSR (default), HaplotypeCaller —emit-ref-confidence GVCF, CombineGVCFs (default), GenotypeGVCFs (default), SelectVariants —select-type SNP, SelectVariants —select-type INDEL, VariantFiltration —filter-expression “QD < 2.0” —filter-name “Filter”. Subsequently, the initial SNP set was filtered in one step using vcftools software (version 0.1.16) with the following key parameters: —maf 0.01, —max-missing 0.8, —minDP 4, —min-alleles 2, —max-alleles 2.

For each individual, SNP markers were concatenated end-to-end. Missing data at corresponding sites were represented by a gap (“-”). The resulting sequences were used to construct a phylogenetic tree using the neighbor-joining method. Genetic distances were calculated with treebest software (version 1.9.2), and the robustness of the tree topology was assessed using bootstrap analysis with 1000 replicates [43]. The resulting phylogenetic tree was visualized using the ggtree package in R. To determine the optimal number of populations (K) and population structure, SNP genotyping data were analyzed using ADMIXTURE software (version 1.3.0). Principal component analysis (PCA) was performed on the SNP dataset using PLINK software (version 1.9) [44].

4.3. Development of Molecular Marker Identification System

4.3.1. KASP Marker Development and Genotyping

SNP loci obtained from Section 4.2.2 were filtered based on the following criteria: (1) locus call rate = 100%, minor allele frequency (MAF) > 0.15, heterozygosity rate < 0.4, and removal of redundant markers; (2) no other SNP within 30 bp upstream or downstream of the selected SNP, and absence of ≥8 consecutive mononucleotide repeats; (3) no homologous sequences within 50 bp upstream or downstream of the locus or elsewhere in the genome; (4) GC content between 40% and 60% within the 150 bp flanking region; (5) PIC ≥ 0.2 and locus polymorphism ≥ 0.4.

KASP primers were designed using Primer3Plus “https://www.primer3plus.com/ (accessed on 1 September 2023)”. A 20 bp sequence immediately upstream of the SNP was designated as the allele-specific forward primer. The two alternative SNP alleles were then incorporated as the 3′-terminal base to create two distinct allele-specific forward primers. Using Primer3Plus with these fixed forward primers, a common reverse primer (R) was designed to ensure PCR amplicon sizes between 60–100 bp, followed by a Primer-BLAST “https://www.ncbi.nlm.nih.gov/ (accessed on 1 September 2023)” check to confirm specificity. The designed allele-specific forward primers were subsequently tagged at their 5′ ends with the FAM (GAAGGTGACCAAGTTCATGCT) or HEX (GAAGGTCGGAGTCAACGGATT) universal adapter sequences, respectively. All final primers were synthesized by Tsingke Biotechnology Co., Ltd. (Kunming, China).

Initial screening was performed on 92 carnation varieties (including the 50 used for sequencing). Following the exclusion of primers with poor genotyping performance, a preliminary set of single-copy, unlinked KASP primers was obtained for further study. A secondary screening was then conducted on an additional 217 carnation varieties. Primers demonstrating unstable amplification, high missing data rates, or low PIC values were discarded, resulting in the final set of core primers for variety identification. Genotyping and signal detection were performed using the GeneMatrix platform (GeneMatrix (Hanchen Guangyi Technology Co., Ltd., Chengdu, China)). The KASP reaction was carried out in a 5 µL mixture containing: 2.5 µL of 2 × Master Mix for ASPCR V1 (Hanchen Guangyi Technology Co., Ltd., Chengdu, China), 0.005 µL of each allele-specific forward primer (100 µM), 0.006 µL of the common reverse primer (100 µM), and 2 µL of template DNA (20 ng·µL^−1^). The PCR amplification protocol was as follows: initial denaturation at 95 °C for 10 min; 10 touchdown cycles of 95 °C for 20 s, 61–55 °C for 60 s (decreasing by 0.6 °C per cycle); followed by 35 cycles of 95 °C for 20 s and 55 °C for 40 s.

4.3.2. Data Analysis

Based on the genotyping results, genetic diversity parameters—including Major Allele Frequency, PIC, observed heterozygosity, and missing rate—were calculated using PowerMarker v3.25. Nei’s genetic distance among the 309 carnation accessions was also computed with PowerMarker v3.25 based on the allele frequencies of the selected markers. Cluster analysis was performed using the unweighted pair-group method with arithmetic means (UPGMA). The number of differential sites between varieties was calculated using the R language (version 4.5.1), and the DNA fingerprint map of the 309 carnation varieties was visualized.