Ⅰ. INTRODUCTION

Oat (Avena sativa L.) is a widely cultivated cereal crop grown around the world, valued for its versatile use as a grain, forage, and industrial resource (Kaur et al., 2024;Marshall et al., 2013). In Korea, oat has traditionally been used as a forage crop, and continuous domestic breeding efforts have led to the development of superior forage oat cultivars. As a result, the self-sufficiency rate of forage oat seed has more than doubled, rising from 21% in 2020 to 42.4% in 2023 (Park et al., 2026). However, as the domestic forage oat seed market has expanded, providing reliable molecular tools to support cultivar identification and seed quality control has emerged as critical challenges (Campuzano-Duque et al., 2022).

Conventional oat cultivar identification has primarily relied on morphological and agronomic traits such as heading date, plant height, leaf morphology, and seed characteristics. However, these phenotypic traits can be easily influenced by environmental conditions and require considerable labor and time to evaluate. In particular, they have fundamental limitations in distinguishing cultivars with similar genetic backgrounds or in achieving rapid and clear identification during the early seedling stage. Therefore, the establishment of an objective and reproducible molecular identification system is essential to enhance the reliability of the seed distribution system and to prevent violation of plant variety protection rights (Azizi et al., 2021;Endo et al., 2022;Yu and Chung, 2021).

DNA-based molecular markers are unaffected by environmental factors and can be applied at any growth stage, including the seedling stage, making them an excellent alternative to morphological identification. Among the various molecular marker systems, insertion-deletion (InDel) markers offer the advantages of co-dominant inheritance and high reproducibility. In addition, compared to simple sequence repeat (SSR) or single nucleotide polymorphism (SNP) markers, which often require expensive equipment or complex procedures, InDel markers provide a more economical and efficient tool for cultivar identification (Noda et al., 2021;Tian et al., 2026). Although InDel marker-based cultivar identification and genetic diversity assessment have already been successfully applied in several major crops (Endo et al., 2022;Tian et al., 2026), an optimized InDel-based DNA profiling system has not yet been systematically established for the various domestic and foreign oat cultivars currently circulating in the Korean market.

In addition to accurate cultivar identification, elucidating the genetic diversity and relationships among cultivars is essential for understanding the genetic background and genetic relationships of domestic and foreign oat cultivars. In particular, multivariate analytical approaches such as unweighted pair group method with arithmetic mean (UPGMA), principal coordinate analysis (PCoA), and analysis of molecular variance (AMOVA) can quantitatively visualize the clustering patterns and the degree of differentiation among lineages, providing an objective basis for the systematic classification and management of germplasm beyond simple cultivar identification (Li et al., 2000;Sulima et al., 2018;Wight et al., 2010).

To this end, the objectives of this study were as follows: First, to develop an efficient and highly reproducible InDel marker-based molecular identification system capable of clearly distinguishing 26 major domestic and foreign oat cultivars. Second, to evaluate the polymorphism levels and marker utility of the selected core InDel marker set. Third, to elucidate the genetic relationships and clustering patterns among the tested oat cultivars using multivariate cluster analysis. The results of this study are expected to serve as a reliable molecular tool for establishing a rigorous seed purity management system in the domestic forage oat market and for supporting future oat breeding.

Ⅱ. MATERIALS AND METHODS

1. Plant materials and Genomic DNA extraction

A total of 26 oat (Avena sativa L.) cultivars, comprising major domestic cultivars and international introduced accessions (Table 1), were used in this study. All seeds were germinated and grown in a growth chamber/incubator maintained at 22°C for 7 to 10 days. Fresh young leaf tissues were harvested from the seedlings for DNA extraction. To ensure robust genotyping and assess uniformity within each cultivar, DNA was individually extracted from 3 to 5 seedlings per cultivar. Furthermore, all following PCR amplifications and capillary electrophoresis steps were performed in triplicate. After confirming the genetic uniformity among the individuals within each cultivar, a consensus genotype profile was established for each of the 26 cultivars to be used in the downstream statistical analyses. Genomic DNA (gDNA) was extracted using the DNeasy Plant Mini Kit (250) (Qiagen, Germany) according to the manufacturer’s instructions. The concentration, purity, and integrity of the extracted gDNA were evaluated using a NanoDrop 2000 Spectrophotometer (Thermo Fisher Scientific, USA) by measuring absorbance ratios at A 260 / A 280 and A 260 / A 230. The extracted gDNA stock samples were stored at −80°C until further PCR analysis.

2. Whole-genome resequencing, InDel Discovery, and filtering

Whole-genome resequencing (WGRS) data from five oat accessions (Swan, S126, Daeyang, Highspeed, and Samhan) were utilized for candidate InDel marker discovery. These five accessions were specifically selected to represent contrasting genetic backgrounds and agronomic traits to maximize the discovery of broadly applicable polymorphic markers: 'Daeyang' as a representative naked oat cultivar; 'Highspeed' and 'Samhan' representing early- and late-maturing domestic hulled oats, respectively; and 'Swan' as a major foreign-introduced hulled oat cultivar. Notably, S126 is an elite domestic breeding line scheduled for official cultivar registration. Although it is not included in the commercial 26-cultivar panel (Table 1), it was included in the discovery phase to ensure the developed markers capture the latest genetic variations in current breeding lines. Raw reads generated from the Illumina sequencing platform were pre-processed using Trimmomatic (v0.39) to remove adapter sequences and low-quality bases (Phred score ≥ 21, read length > 50 bp). Clean reads were aligned to the oat reference genome (Avena sativa v2, PepsiCo_OT3098_V2) using the BWA-mem algorithm integrated into the Sentieon DNAseq pipeline (v202503).

Short variant calling and genotyping were conducted using GATK (v4.4.0.0) and vcftools (v0.1.17). To select highly robust, polymorphic InDel markers suitable for reliable oat cultivar identification, a rigorous 8-step variant filtering pipeline was executed and validated via automated capillary electrophoresis (QIAxcel):

3. Primer design for candidate InDel markers

Specific primer pairs flanking each target InDel locus were designed to amplify the selected InDel regions. Detailed information on the primer sequences and amplification characteristics of the developed InDel markers is provided in Table S1.

4. PCR amplification and capillary electrophoresis

PCR amplification was performed in a total volume of 20 μL containing 1 μL of template gDNA, 10 μL of 2X HiSense Taq PCR Master Mix (dye-free formulation; CellSafe, Korea), 1 μL of forward primer (10 pmol/μL), 1 μL of reverse primer (10 pmol/μL), and 7 μL of nuclease-free water.

Thermal cycling was conducted using thermal cyclers, including the C1000 Touch Thermal Cycler (Bio-Rad, USA) and the ProFlex PCR System (Applied Biosystems, CA, USA), under the following conditions: initial denaturation at 95°C for 5 min; 35 cycles of denaturation at 95°C for 30 s, annealing at 60°C for 30 s, and extension at 72°C for 1 min; followed by a final extension at 72°C for 5 min.

To confirm PCR products, automated capillary electrophoresis was performed using an eGene HDA-GT12™ system (QIAGEN, Germany) with a QIAxcel DNA High Resolution Kit. Amplicons were separated using the OM500 analysis method. Fragment size calibration and allele sizing were determined using a QIAxcel DNA size marker (50 – 800 bp) and Alignment marker (15 bp/1,000 bp). Digital electropherograms and virtual gel images were processed and analyzed using QIAxcel ScreenGel Software (v2.1).

5. Genetic diversity analysis and cultivar identification system

1) InDel genotyping and core marker selection

Amplified InDel band patterns were scored for each cultivar based on allele sizes (e.g., Allele A, Allele B, and Heterozygote H). Monomorphic bands were excluded from further analysis. To evaluate the polymorphism of each InDel locus, key genetic parameters including Major Allele Frequency (MAF), Number of Alleles (N_a), and Polymorphism Information Content (PIC) were calculated using R software (v4.5.2). MAF was determined as the frequency of the most common allele at each locus, and N_a represented the total number of detected alleles per locus. PIC was calculated to measure the informative power of each InDel marker (Botstein et al., 1980). To select the optimized core marker set, a step-by-step selection procedure was applied. Candidate markers were evaluated not only by their PIC values, but mainly by their ability to distinguish the 26 cultivars. Markers showing redundant genotype patterns across the cultivars were removed to minimize panel size. In contrast, markers needed to separate specific, closely related cultivars were retained, regardless of their overall PIC values. To visualize the genotypic patterns of the selected InDel markers across the 26 oat cultivars, a genotypic heatmap was generated using the R packages ggplot2 and reshape2.

2) Genetic distance and cluster analysis

For genetic diversity and cluster analysis, the genotypic matrix was converted into a genind object to calculate pairwise genetic distances based on Nei’s genetic distance (Nei et al., 1983). A UPGMA dendrogram was constructed with 100 bootstrap replicates using the poppr and ape packages.

3) Principal Coordinate Analysis (PCoA)

To complement the UPGMA clustering and examine continuous genetic relationships, Principal Coordinate Analysis (PCoA) based on Euclidean distance was performed using cmdscale. The variation explained by the first two axes (PCoA1 and PCoA2) was calculated, and the resulting distribution of cultivars was visualized using ggplot2 and ggrepel.

4) Analysis of Molecular Variance (AMOVA)

To examine genetic differentiation between and within lineages, Analysis of Molecular Variance (AMOVA) was conducted using the R package poppr. Genotypic data were loaded into a genind object and grouped by lineage. We estimated degrees of freedom (df), variance components, and the proportion of total genetic variation at each level. Statistical significance was determined by 999 Monte Carlo permutations using the randtest function in ade4.

Ⅲ. RESULTS AND DISCUSSION

1. InDel marker profiling and polymorphism analysis

To establish an efficient molecular profiling system for oat (Avena sativa L.) cultivars, we evaluated the genetic polymorphism of selected InDel markers across 26 oat varieties (Table 1). Genomic DNA was extracted from young oat seedlings cultivated under controlled conditions (Fig. 1a). To assess genetic variation across these accessions, we evaluated the polymorphism of 12 candidate InDel markers, with their statistical parameters detailing allele distribution and informativeness (Table 2). Automated capillary-based electrophoresis successfully differentiated homozygous and heterozygous genotypes, as exemplified by the OID_48 marker displaying distinct polymorphic bands: 437 bp for A, 501 bp for B, and co-dominant bands for 'Heterozygous' H patterns (Fig. 1b). Among the evaluated loci, a core panel of nine optimized InDel markers (OID_11, OID_28, OID_31, OID_35, OID_39, OID_43, OID_44, OID_48, and OID_167) was selected to maximize genotype differentiation efficiency while minimizing marker redundancy. Although PIC values generally indicate marker informativeness, the primary criterion for the core panel selection was the capability to provide unique alleles for each cultivar. For instance, OID_41 (PIC = 0.350) was excluded from the final set because its identification pattern was completely redundant and already covered by other selected markers. In contrast, OID_35 (PIC = 0.132) was strictly retained despite its relatively low polymorphism, because it possessed a unique allele essential for separating specific cultivars that could not be differentiated by any other markers in the panel. High-throughput genotyping using this streamlined panel generated a comprehensive genotype database heatmap across all 26 varieties, clearly color-coded by allele pattern (A: Blue, B: Red, H: Green) (Fig. 1c).

While conventional InDel markers are typically bi-allelic with a maximum theoretical PIC of 0.500, the OID_44 and OID_48 loci had three alleles (N_a=3), showing higher PIC values of 0.576 and 0.559, respectively (Table 2). This multi-allelic nature improved the overall resolution of the core set. Importantly, this nine core InDel panel generated unique multilocus genotypes for all 26 cultivars evaluated in this study. These results demonstrate that when conducted via individual plant genotyping, the optimized nine-marker panel reliably distinguishes contaminating cultivars among circulating commercial varieties, showing strong potential applicability for seed purity management and protecting plant variety rights in oat breeding programs, especially with the growing forage oat market and seed imports.

2. Genetic relationship and cluster analysis

To analyze the genetic relationships and clustering among the 26 oat cultivars, UPGMA cluster analysis and principal coordinate analysis (PCoA) were conducted using the InDel genotyping matrix. Both methods consistently divided the cultivars into two distinct lineages (Fig. 2).

The UPGMA dendrogram resolved the 26 cultivars into two distinct clusters (Fig. 2a): Lineage 1 (orange cluster), which included 13 varieties (Hi-Early, Rayeon, Antigo, Jungmo2513, Highspeed, Choyang, Ilhan, Donghan, Seonyang, Daeyang, Cayuse, Joyeon, and Jopung), and Lineage 2 (blue cluster), which comprised the remaining 13 varieties (Dajo, Samhan, Punghan, Pallinup, Taehan, Shinhan, Janghan, Dahan, Johan, Gwanghan, Swan, Okhan, and Dakyeong).

The two-dimensional PCoA plot corroborated this binary structure, with PCoA 1 (30.26%) and PCoA 2 (18.83%) explaining a combined 49.09% of the total genetic variance (Fig. 2b). Spatial separation along the PCoA 1 axis perfectly aligned with the clustering patterns observed in the UPGMA dendrogram, highlighting the clear genomic divergence between the two lineages. This clear genetic separation may reflect potential associations with the underlying breeding histories, pedigrees, and geographic origins of the cultivars. Notably, all domestic naked oat cultivars (Daeyang, Choyang, and Seonyang) were tightly co-clustered within lineage 1, suggesting possible genetic inheritance from shared parental lines and common ancestry in domestic naked oat breeding programs. Furthermore, the limited number of foreign introduced varieties evaluated in this study showed a tendency to group by their geographic origins: American cultivars (Cayuse and Antigo) belonged to lineage 1, whereas Australian cultivars (Swan and Pallinup) were classified into lineage 2. This demonstrates that the core InDel panel effectively captures lineage-specific genomic divergence across international accessions.

Analysis of Molecular Variance (AMOVA) was further conducted to quantify the partitioning of genetic diversity between and within these two defined lineages (Table 3). The statistical significance of these variance components (σ², p < 0.001 based on 999 Monte Carlo permutations) confirms both highly significant inter-group genetic differentiation (32.11%) and rich intra-group genetic diversity (51.10%). This significant genetic partitioning may reflect underlying pedigree backgrounds and selection histories, suggesting that the selected InDel marker set has practical utility for germplasm classification, parental selection, and molecular breeding in oat.

Ⅳ. CONCLUSIONS

With the domestic forage oat market continuing to expand alongside the growth of the livestock industry, securing a reliable molecular tool for seed purity management and cultivar protection has become increasingly important. In response to this need, this study developed a reliable InDel marker-based identification system capable of preventing unauthorized cultivar mixing and supporting plant breeders rights.

Using 26 domestic and foreign-introduced oat cultivars, we evaluated 12 InDel markers and performed genetic similarity and UPGMA cluster analysis to characterize their genetic relationships. The evaluated markers exhibited 2 to 3 alleles per locus, with PIC values ranging from 0.132 to 0.576, indicating their utility for distinguishing genetically similar cultivars. Although certain domestic and foreign-introduced cultivars showed close genetic similarity and formed mixed clusters, an optimized core combination of nine InDel markers (OID_11, OID_28, OID_31, OID_35, OID_39, OID_43, OID_44, OID_48, and OID_167) successfully discriminated all 26 cultivars with complete identification efficiency, indicating that a compact and efficient marker set can provide an effective tool for cultivar identification.

When applied at the individual plant level, this system enables complete identification of varietal contamination among all 26 commercial cultivars evaluated. These findings provide a practical molecular framework for potential applicability for seed quality control in the domestic forage oat industry. By offering a robust and reproducible tool for intellectual property protection, this InDel-based identification system is expected to enhance market transparency and strengthen trust in the domestic forage oat seed market.