Development of Single Nucleotide Polymorphism and Phylogenetic Analysis of Rhododendron Species in Zhejiang Province, China, Using ddRAD-Seq Technology

Abstract

The genus Rhododendron presents significant challenges for systematic classification due to extensive hybridization and adaptive radiation. Here, we employed double-digest restriction site-associated DNA sequencing (ddRAD-seq) to resolve phylogenetic relationships among nine ecologically significant Rhododendron species (34 accessions) endemic to Zhejiang Province, China, a biodiversity hotspot for this genus. Using R. simsii as the reference genome, we generated 39.40 Gb of high-quality sequencing data with a Q30 score of 96.65% and a GC content of 39.63%, achieving an average alignment rate of 92.79%. Through stringent filtering (QD ≥ 2, MQ ≥ 40), we identified 14,048,702 genome-wide single nucleotide polymorphism (SNP), predominantly characterized by the mutation types T:A>C:G and C:G>T:A. The widespread R. simsii and R. simsii var. putuoense exhibited significant genetic diversity, whereas the low-altitude widespread R. molle and the endemic R. simiarum exhibited lower genetic diversity. Moderate genetic differentiation (F_st = 0.097) was observed between R. simsii and R. simsii var. putuoense, while substantial genetic differentiation was detected among the other Rhododendron species. Principal component analysis (PCA), combined with phylogenomic reconstruction, demonstrated that the Rhododendron genus can be stratified into six well-supported genetic clades. Furthermore, this study provides the first genomic validation of the sibling relationship between R. simsii and its variety, R. simsii var. putuoense, and clarifies the systematic position of R. huadingense, suggesting that it should be classified as a new subgenus. This study establishes ddRAD-seq as a cost-effective tool, providing both a theoretical framework for SNP-based phylogenetics and critical insights for conserving China’s azalea biodiversity.

1. Introduction

Rhododendron L. (Ericaceae) comprises over 1000 species globally and is widely distributed across the temperate regions of Europe, Asia, and North America. China possesses the richest resources of Rhododendron species and is also one of the primary centers of modern distribution and differentiation of the genus worldwide [1]. The country is home to approximately 571 wild species, with more than 74% endemic to China [2,3]. As a vital member of subtropical forest ecosystems, Rhododendron plants play a significant role in maintaining the stability of the structure and function of mountainous ecosystems [4]. Additionally, this genus of plants holds considerable ornamental value [5], medicinal applications, and development potential [6].

Accurate construction of phylogenetic relationship is crucial for understanding species evolution. Due to habitat diversity and reproductive modes influenced by rapid adaptive radiation and extensive natural hybridization among species, the genus Rhododendron is regarded one of the most challenging plant groups. Hybridization and genome duplication have significantly shaped the diversification of Rhododendron, driving adaptive evolution through enhanced genetic variation and novel genomic architectures [7,8]. Previous studied have employed various molecular marker techniques to investigate the phylogeny and diversity of Rhododendron species, including DNA barcoding [9,10], simple sequence repeats (EST-SSRs) [11,12], plastid genomes [13,14], and restriction-site-associated DNA sequencing (RAD-seq) [15]. Although research on the genus Rhododendron has been conducted for many years, the aforementioned markers exhibit several limitations such as low throughput, time consumption, inaccuracy, and poor cost-effectiveness, which have led to numerous unresolved issues and restricted the scientific classification and accurate identification of Rhododendron species. In recent years, the rapid advancement of genome sequencing technology, molecular marker research has presented new opportunities. Double-digest restriction-site-associated DNA sequencing (ddRAD-seq) is a reduced-representation genome sequencing (RRGS) technique developed based on next-generation sequencing (NGS), capable of developing a large number of single nucleotide polymorphism (SNP) markers at a lower cost, independent of the presence of a reference genome or chromosomal ploidy [16]. Because SNP are distributed throughout the genome, they offer greater coverage and accuracy, thus offering higher resolution and reliability compared to traditional molecular markers.

Currently, a total of 19 wild species of the genus Rhododendron have been recorded in Zhejiang Province in China. Although this number is relatively small, these species encompass four subgenera: subgen. Tsutsusi (Sweet) Pojarkova, subgen. Pentanthera (G. Don) Pojarkova, subgen. Azaleastrum Planchon ex K. Koch, and subgen. Hymenanthes (Blume) K. Koch. Furthermore, ongoing discoveries of new local species, such as Rhododendron huadingense B. Y. Ding & Y. Y. Fang [17,18] and R. simsii var. putuoense G. Y. Li & Z. H. Chen [19], have introduced new species and varieties published after the release of the Flora of China (Volume 57) in 1994. However, the descriptions of these two species was based solely on morphological descriptions, lacking robust molecular evidence. This limitation has led to controversies regarding the systematic positions and taxonomic definitions of some species.

This study employs ddRAD-seq technology to perform reduced-representation genome sequencing of representative Rhododendron plants from Zhejiang, with the aim of identifying high-quality SNP markers at the genomic level. The objectives are as follows: (1) to evaluate the genetic diversity and differentiation among these species and (2) to reconstruct robust phylogenetic relationships, thereby providing evidence and solutions for the challenges associated with classification Rhododendron species in Zhejiang.

2. Materials and Methods

2.1. Plant Materials

A total of 34 samples representing nine species and one variety of Rhododendron were used for ddRAD-seq analysis. The species included R. simsii Planch., R. farrerae Sweet, R. huadingense, R. molle (Blume) G. Don, R. ovatum (Lindl.) Planch. ex Maxim., R. championiae Hook., R. latoucheae Franch., R. simiarum Hance, R. fortunei Lindl., and R. simsii var. putuoense. Two to five individuals were sampled from each taxon, encompassing four subgenera and seven sections following the Flora of China (2005) (

Table 1. Data characteristics of nine species and one variety of Rhododendron by ddRAD sequencing.

No.	Sample	Species	Subgenus	Section	Reads Number	Total Bases/bp	Mapping Rate/%	Sequence Depth/×
1	DJ_1	R. simsii	subgen. Tsutsusi	sect. Tsutsusi	7,842,910	1,123,802,645	96.84	2.13
2	DJ_2	R. simsii	subgen. Tsutsusi	sect. Tsutsusi	8,447,614	1,209,526,268	96.58	2.29
3	DJ_3	R. simsii	subgen. Tsutsusi	sect. Tsutsusi	6,872,262	983,964,165	96.91	1.86
4	DJ_4	R. simsii	subgen. Tsutsusi	sect. Tsutsusi	8,227,636	1,178,483,283	91.14	2.23
5	DJ_5	R. simsii	subgen. Tsutsusi	sect. Tsutsusi	7,468,622	1,069,413,995	93.78	2.02
6	PTDJ_1	R. simsii var. putuoense			14,464,694	2,071,406,861	95.01	3.92
7	PTDJ_2	R. simsii var. putuoense			14,516,774	2,079,276,661	83.70	3.93
8	PTDJ_3	R. simsii var. putuoense			18,210,666	2,607,244,029	93.94	4.93
9	PTDJ_4	R. simsii var. putuoense			9,344,540	2,106,199,392	94.93	3.98
10	PTDJ_5	R. simsii var. putuoense			7,083,544	1,337,742,896	94.50	2.53
11	DXDJ_1	R. farrerae	subgen. Tsutsusi	sect. Brachycalyx	7,887,388	1,129,332,945	95.82	2.14
12	DXDJ_2	R. farrerae	subgen. Tsutsusi	sect. Brachycalyx	7,065,322	1,011,327,506	94.25	1.91
13	DXDJ_3	R. farrerae	subgen. Tsutsusi	sect. Brachycalyx	7,406,124	1,060,945,460	95.96	2.01
14	HDDJ_1	R. huadingense	subgen. Tsutsusi	sect. Brachycalyx	7,345,694	1,051,336,075	92.74	1.99
15	HDDJ_2	R. huadingense	subgen. Tsutsusi	sect. Brachycalyx	7,588,046	1,085,558,978	93.62	2.05
16	HDDJ_3	R. huadingense	subgen. Tsutsusi	sect. Brachycalyx	8,763,824	1,254,026,187	94.01	2.37
17	HDDJ_4	R. huadingense	subgen. Tsutsusi	sect. Brachycalyx	7,032,948	1,006,765,697	93.28	1.90
18	YZZ_1	R. molle	subgen. Pentanthera	sect. Pentanthera	6,872,262	954,249,606	89.29	1.81
19	YZZ_2	R. molle	subgen. Pentanthera	sect. Pentanthera	8,227,636	1,071,341,272	88.80	2.03
20	YZZ_3	R. molle	subgen. Pentanthera	sect. Pentanthera	7,468,622	935,457,901	90.03	1.77
21	MYH_1	R. ovatum	subgen. Azaleastrum	sect. Azaleastrum	7,156,964	1,024,609,279	95.74	1.94
22	MYH_2	R. ovatum	subgen. Azaleastrum	sect. Azaleastrum	7,987,170	1,143,354,013	94.15	2.16
23	MYH_3	R. ovatum	subgen. Azaleastrum	sect. Azaleastrum	6,943,508	994,044,215	94.69	1.88
24	CMDJ_1	R. championiae	subg. Azaleastrum	sect. Choniastrum	7,001,716	1,002,884,795	91.92	1.90
25	CMDJ_2	R. championiae	subgen. Azaleastrum	sect. Choniastrum	7,083,544	1,014,407,528	91.71	1.92
26	CMDJ_3	R. championiae	subgen. Azaleastrum	sect. Choniastrum	6,998,250	1,002,294,502	90.80	1.90
27	LJDJ_1	R. latoucheae	subgen. Azaleastrum	sect. Choniastrum	7,988,094	1,144,587,841	92.62	2.17
28	LJDJ_2	R. latoucheae	subgen. Azaleastrum	sect. Choniastrum	6,100,640	873,849,781	92.48	1.65
29	LJDJ_3	R. latoucheae	subgen. Azaleastrum	sect. Choniastrum	8,249,558	1,181,825,031	91.94	2.24
30	HTDJ_1	R. simiarum	subgen. Hymenanthes	subt. Argyrophylla	7,336,492	1,050,798,076	90.36	1.99
31	HTDJ_2	R. simiarum	subgen. Hymenanthes	subt. Argyrophylla	8,418,280	1,205,949,808	88.20	2.28
32	YJDJ_1	R. fortunei	subgen. Hymenanthes	subt. Fortunea	6,998,250	1,071,807,669	90.95	2.03
33	YJDJ_2	R. fortunei	subgen. Hymenanthes	subt. Fortunea	7,842,910	980,284,260	92.32	1.85
34	YJDJ_3	R. fortunei	subgen. Hymenanthes	subt. Fortunea	8,447,614	997,989,300	91.97	1.89

The infrageneric taxon system of Rhododendron follows classification outlined in the Flora of China (2005).

). Among these, R. simsii var. putuoense, a variety published in 2010, has not yet been assigned an infrageneric taxon system. Field sampling was conducted in September 2024. Except R. simsii var. putuoense, which was collected from Dinghai District, Zhoushan (122.11° E, 30.08° N), all other specimens were obtained from the Jinhua Pan’an Rhododendron Villa (120.54° E, 28.39° N). This repository is dedicated to the conservation and study of native Rhododendron species, particularly those indigenous to Zhejiang Province.

2.2. DNA Extraction and Library Preparation

For each individual, 0.5 g of fresh leaf tissue was collected, and total genomic DNA was extracted using the CTAB method, followed by quality assessment with 1.2% agarose gel electrophoresis. A combination of restriction endonucleases DpnII and MspI was employed for genomic digestion, selecting DNA insert fragments ranging from 300 to 500 bp for the library. The library was sequenced using paired-end (PE) sequencing with 2 × 500 bp on the Illumina NovaSeq platform, with library construction and sequencing carried out by Shanghai Personalbio Biotechnology Co., Ltd., Shanghai, China.

2.3. SNP Calling

After sequencing data were generated, the fastp (v. 0.20.0) [20] was employed to remove adapter sequences and low-quality reads from the raw data using a sliding window method based on the following criteria: a 5 bp sliding filter was applied, removing windows with a Q value < 20, and trimming any reads with a length < 50 bp or containing ≥5 N bases at either end. The filtered sequences (clean reads) were aligned to the R. simsii reference genome (ASM1428224v1) using the mem program of BWA software (v. 0.7.12-r1039), followed by SNP detection using GATK software (v. 3.4.46) [21]. To ensure the reliability of the identified SNP loci, further filtering of the SNP genotype results was conducted, adhering to the following criteria: Fisher test of strand bias (FS) ≤ 60, Haplotype Score ≤ 13.0, Mapping Quality (MQ) ≥ 40, Quality Depth (QD) ≥ 2, Read Pos Rank Sum ≥ −8.0, and MQ Rank Sum > −12.5.

2.4. Genetic Diversity and Genetic Differentiation

Genetic diversity and differentiation analyses among species were conducted using the population command in the Stacks package, with parameters including: the average number of individuals per SNP (Num Indv), observed heterozygosity (H_o), expected heterozygosity (H_e), nucleotide diversity (π), inbreeding coefficient (F_is), and pairwise fixation index (F_st).

2.5. PCA and Phylogenomic Tree Reconstruction

Principal Component Analysis (PCA) was performed using GCTA software (v. 1.94.1) [22] on the genome-wide SNP dataset (14,048,702 loci) after filtering for minor allele frequency (MAF ≥ 0.05). Only the first two principal components (PC1 and PC2) were retained for visualization, as they cumulatively explained 25.97% of the genetic variance (see Results)”, retaining only the first two axes were retained for presentation. A phylogenomic tree was constructed using the Maximum Likelihood (ML) in fastTree software (v. 2.1.3) [23] and the reliability of the clades was validated using bootstrap resampling (1000 repetitions).

3. Results

3.1. Sequencing Data Quality Control and Filtering

ddRAD sequencing was conducted on 34 samples of Rhododendron species, resulting in a total of 39.40 Gb of data, with an average data volume of 1.16 Gb per sample. The number of generated raw reads ranged from 6,254,358 to 18,559,976, with an average GC content of 39.63%. The average Q20 and Q30 scores were 98.81% and 96.65%, respectively, indicating high sequencing quality. After filtering, 38.20 Gb of high-quality data was obtained, with an average data volume of 1.12 Gb per sample. The number of clean reads generated ranged from 6,100,640 to 18,210,666. The clean reads were aligned to the Rhododendron reference genome, with the number of aligned reads ranging from 5,642,084 to 17,107,014 and an alignment rate of 83.70% to 96.91%; the average alignment rate was 92.79%. The sequencing depth varied between 1.65× and 4.93×, and the alignment results met the criteria for subsequent data analysis (Table 1).

3.2. SNP Distribution and Identification

After filtering the SNP loci data, a total of 14,048,702 high-quality SNP loci were identified. The number of SNP were counted using a 100 kb sliding window across the populations. Based on the distribution across the 13 chromosomes, the SNP loci showed a relatively even distributed on the chromosomes (Figure 1).

Genomic SNP mutations can be categorized into six types, with the largest proportion being C:G>T:A, followed by T:A>C:G; these two types together account for 64.42%. The remaining four types collectively account for 35.58% (Figure 2).

3.3. Genetic Diversity Analysis

A statistical analysis of the genetic diversity parameters for ten species of Rhododendron was conducted (

Table 2. Genetic diversity statistics for nine species and one variety of Rhododendron species.

Species	Num Indv	Ho	He	π	Fis
R. simsii var. putuoense	4.8102	0.0549	0.0574	0.0642	0.0199
R. simsii	4.7006	0.0474	0.0611	0.0685	0.0466
R. molle	2.8409	0.0293	0.0199	0.0241	−0.0087
R. fortunei	2.7522	0.0362	0.0432	0.0528	0.0306
R. huadingense	3.7156	0.0345	0.0280	0.0323	−0.0035
R. ovatum	2.8092	0.0355	0.0301	0.0365	0.0019
R. championiae	2.8421	0.0379	0.0301	0.0364	−0.0023
R. latoucheae	2.8224	0.0372	0.0318	0.0386	0.0028
R. farrerae	2.7556	0.0318	0.0325	0.0395	0.0140
R. simiarum	1.9170	0.0374	0.0245	0.0330	−0.0066
Mean value	3.1966	0.0382	0.0359	0.0426	0.0193

). The results indicate that the number of individuals (Num Indv) ranged from 1.9170 (R. simiarum) to 4.8102 (R. simsii var. putuoense), with a mean of 3.1966. Observed heterozygosity (H_o) ranged from 0.0239 (R. molle) to 0.0549 (R. molle), with a mean of 0.0382. Expected heterozygosity (H_e) ranged from 0.0199 (R. molle) to 0.0611 (R. simsii), with a mean of 0.0359; Nucleotide diversity (π) ranged from 0.0241 (R. molle) to 0.0685 (R. simsii), with a mean of 0.0426. Inbreeding coefficient (F_is) ranged from −0.0066 (R. simiarum) to 0.0466 (R. simsii), with a mean of 0.0193. Four species (R. molle, R. huadingense, R. championae, and R. simiarum) exhibited negative F_is values, while the others had positive values.

3.4. Genetic Differentiation Analysis

The F_st values for the ten species of Rhododendron range from 0.097 to 0.806, with an average of 0.641. Among these species, the greatest genetic differentiation is observed between R. molle and R. huadingense, while the smallest genetic differentiation occurs between R. simsii var. putuoense and R. simsii (F_st = 0.097). R. molle (F_st range: 0.681–0.806) and R. huadingense (F_st range: 0.732–0.757) show strong differentiation with the other species (Figure 3).

To further investigate the genetic differentiation among the nine species and one variety of the genus Rhododendron, a PCA was conducted. The contribution rates of the first principal component (PC1) and the second principal component (PC2) were 13.87% (eigenvalue = 4.52) and 12.10% (eigenvalue = 3.94), respectively, resulting in a cumulative contribution rate of 25.97%. Based on the positions and mutual distances of the species in the two-dimensional plot, they can be broadly categorized into six groups (Figure 4).

3.5. Phylogenetic Analysis

Phylogenetic trees based on 14,048,702 SNP loci were constructed using the ML method, revealing that the 34 samples clustered into six well-supported genetic clades. The first clade comprised R. simsii var. putuoense, R. simsii, and R. farrerae, which belong to subgen. Tsutsusi. The second clade included only R. ovatum, which is classified under subgen. Azaleastrum. The third clade contained solely R. huadingense. The fourth clade encompassed R. championiae and R. latoucheae, which are classified under subgen. Azaleastrum. The fifth clade consisted solely of R. molle, belonging to subgen. Pentanthera and the sixth calde comprised R. simiarum and R. fortunei. These clustering results are consistent with the findings from the PCA (Figure 5).

4. Discussion

4.1. Genetic Diversity and Genetic Differentiation of Rhododendron Species Produced in Zhejiang Province

Genetic diversity is the result of long-term evolution in biological populations and acts as a prerequisite for their survival, adaptation, and development. Higher genetic diversity and richer variations enhance the ability to adapt to environmental changes. This study demonstrates that at the species level, R. simsii and R. simsii var. putuoense exhibit high genetic diversity, whereas R. molle and R. simiarum show low genetic diversity. The Rhododendron species R. simsii, R. simsii var. putuoense, and R. molle have been described as widely distributed across Zhejiang Province [24]. However, R. molle is confined to low hills and mountains below 1000 m in elevation, while R. simiarum is classified as an endemic species, restricted to high-altitude regions. For widely distributed species, larger population sizes and the potential for gene flow among populations across different regions further promote increased genetic diversity. In contrast, endemic species with restricted distribution ranges and smaller population sizes are more susceptible to genetic drift, which reduces genetic variation. Additionally, their adaptation to stable environments with lower selective pressures may further limit the generation and maintenance of novel genetic diversity, resulting in overall lower genetic diversity compared to widespread species.

As a classical metric for quantifying population divergence, pairwise F_st provide critical insights into genetic differentiation. Following Wright’s categorization (F_st < 0.05: negligible; 0.05–0.15: moderate; 0.15–0.25: considerable; >0.25: substantial) [25], the moderate differentiation (F_st = 0.097) values for R. simsii var. putuoense and R. simsii likely reflect a combination of ancestral gene flow and incipient divergence. Phylogenomic evidence supports their recent divergence from a common ancestor (Figure 5), with ongoing but limited gene flow, potentially maintained through historical range overlaps or occasional hybridization. However, their current parapatric distributions—R. simsii in mountainous habitats versus R. simsii var. putuoense in coastal island ecosystems—create ecological barriers that promote adaptive differentiation. In contrast, strong genetic boundaries (F_st > 0.25) among other congeneric species imply reproductive isolation, potentially driven by long-term geographic separation, ecological specialization, or intrinsic postzygotic barriers.

4.2. Phylogenetic Relationships and Taxonomy of Rhododendron Species Produced in Zhejiang Province

Despite numerous phylogenetic analyses in the genus Rhododendron, persistent controversies remain regarding its infrageneric classification. Early molecular studies relying on limited genomic markers [9,10], and restricted taxon sampling failed to resolve deep phylogenetic relationships. Our phylogenomic analysis confirms the monophyly of subgen. Tsutsusi, encompassing sect. Tsutsusi and sect. Brachycalyx (represented by R. farrerae). Contrasting with classical taxonomy, we reveal the polyphyly of subgen. Azaleastrum: sect. Azaleastrum (R. ovatum) clusters sister to subgen. Tsutsusi, while sect. Choniastrum (R. championiae and R. latoucheae) forms an independent lineage. This finding is corroborated by Shen et al. [15], whose RAD-seq analysis of 144 global species similarly rejected the monophyly of subgen. Azaleastrum. However, while their study proposed a weakly supported sister relationship between subgen. Tsutsusi and the combined Azaleastrum-Choniastrum clade, our ddRAD-seq data with substantially higher SNP density and lower missing rates provides decisive evidence for their phylogenetic separation. These results necessitate a taxonomic revision to split subgen. Azaleastrum, and reclassify sect. Azaleastrum under subgen. Tsutsusi.

R. huadingense is a deciduous shrub that was jointly described in 1990 by botanist Bingyang Ding and his instructor [17]. It is sparsely distributed in high-altitude mountainous regions and is a key protected wild plant endemic to Zhejiang Province. Due to difficulties in natural regeneration, population decline, and human disturbance leading to a reduced population size, this species was designated as a national Level II key protected wild plant in 2021 [26]. The systematic position of R. huadingense has been a topic of long-standing debate due to numerous unique traits. For instance, based on characteristics such as its deciduous nature, soft hairs on young branches, and 3–5 leaf whorls, R. huadingense was initially classified in sect. Brachycalyx of subgen. Tsutsusi when it was first discovered [27]. Cluster analysis using 12 pairs of EST-SSR molecular markers indicated that R. huadingense shares the closest phylogenetic relationship with Rhododendron farrerae, supporting this classification [28]. However, based on the morphology of flower buds, leaf buds, pollen, and seeds, R. huadingense has also been suggested to belong to subgen. Pentanthera [29]. This perspective is supported by chloroplast genome data [30], although the results were affected by significantly insufficient sampling within the genus. According to the phylogenomic reconstruction results of this study, R. huadingense forms a sister group relationship with subgen. Tsutsusi, while it is more distantly related to subgen. Pentanthera represented by R. molle. Given that the systematic position of R. huadingense is now largely clarified at the species level, it is essential to establish this species as a new subgenus.

R. simsii var. putuoense is a semi-evergreen shrub that is exclusively distributed in the Zhoushan Islands and the eastern coastal areas in Zhejiang. The primary distinguishing feature from the original variety R. simsii is the purple corollas of the former, while other morphological characteristics are similar [19]. Consequently, the author previously hypothesized that the phylogenetic relationship between the two should be close. The phylogenetic topology reconstructed in this study supports this hypothesis, indicating that R. simsii var. putuoense belongs to subgen. Tsutsusi and forms a sister group relationship with R. simsii; these two taxa have diverged from their most recent common ancestor—R. simsii—thereby reinforcing the taxonomic status of the variety.

5. Conclusions

This study applied ddRAD-seq to resolve genetic diversity and phylogenetic relationships among Rhododendron species in Zhejiang, China. Key findings include: (1) genomic validation of the sister relationship between R. simsii and its coastal variety R. simsii var. putuoense (F_st = 0.097), highlighting their ecological divergence; (2) recognition of the endemic R. huadingense as a distinct subgenus, resolving long-standing taxonomic ambiguities; (3) contrasting genetic diversity patterns, where widespread species (R. simsii, π = 0.064–0.068) exhibited higher diversity than endemic/low-altitude species (R. simiarum, R. molle; π = 0.024–0.033), driven by genetic drift and limited selective pressures; (4) phylogenomic evidence rejecting the monophyly of subgenus Azaleastrum, necessitating taxonomic reclassification. The identification of 14 million genome-wide SNPs underscores the efficacy of ddRAD-seq for resolving complex evolutionary histories in non-model plants. These results provide critical insights for biodiversity conservation and emphasize the integration of genomic tools in addressing hybridization and adaptive radiation challenges within Rhododendron.