Development of plastome-based HRM markers for the authentication of 12 major medicinal herbs in the Apiaceae family

Abstract

Medicinal plants of the Apiaceae family often exhibit high morphological similarity, making accurate species identification challenging and leading to frequent cases of adulteration or substitution. To establish a systematic barcoding framework, we simultaneously assembled the chloroplast genomes (plastomes) and 45 S nuclear ribosomal DNA (nrDNA) sequences of 12 major medicinal herb species and performed comparative genomic analyses to identify interspecific variations in single-nucleotide polymorphisms (SNPs), insertions/deletions, nucleotide diversity (π), and simple sequence repeats (SSRs). Based on these genomic insights, seven high-resolution melting (HRM) markers were developed, which generated species-specific melting curve profiles and enabled unambiguous identification and discrimination of all 12 taxa. Compared with conventional DNA barcoding, HRM demonstrated advantages in cost-effectiveness, speed, and suitability for high-throughput analysis. Collectively, our study provides genomic resources and molecular tools that establish a comprehensive barcoding system for 12 medicinal herbs in the Apiaceae family, facilitating reliable species identification, genetic diversity assessment, and phylogenetic research, while contributing to the conservation and safe utilization of these plants.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-026-47346-4.

Introduction

Medicinal plants play a pivotal role in both traditional medicine and the modern health industry, and their value continues to rise steadily. Globally, the demand for medicinal and aromatic plants has shown consistent growth, with the market recently estimated at around USD 70 billion and an annual compound growth rate of approximately 7%¹. This increasing trend is largely driven by consumer preference for natural products and the expansion of the health and wellness industry, thereby highlighting the growing need for research and utilization of medicinal and aromatic plants.

Among these, the Apiaceae family represents one of the most important groups of medicinal plants, widely used in traditional remedies, folk medicine, and modern pharmaceuticals. Apiaceae is a large angiosperm family that comprises approximately 434 genera and 3,780 species^2–4, with most species distributed across the temperate regions of the Northern Hemisphere^5,6. Members of this family, including roots, fruits, and aerial parts, are known to possess diverse pharmacological activities^7–9. For example, species of Angelica have been traditionally used for women’s health and gastrointestinal disorders and are reported to exhibit cardiovascular, immune-modulating, anti-inflammatory, antimicrobial, and antioxidant effects^10–14. Bupleurum species are well known for their antidepressant, anti-inflammatory, anticancer, antipyretic, antibacterial, and antiviral activities^15,16. Similarly, Cnidium species have analgesic, antipruritic, anti-allergic, antidermatophytic, antibacterial, and antifungal properties^17,18.

However, species within the same genus often exhibit high morphological similarity, particularly in dried medicinal materials or processed products, making accurate identification difficult. As a result, substitution or adulteration occurs frequently, which can lead to discrepancies in pharmacological efficacy or severe safety issues due to unintentional use of toxic species (e.g., Conium maculatum)^19,20. Traditionally, identification of herbal medicinal species has relied on morphological classification, histological examination, or chemical profiling of specific metabolites²¹. Yet, morphological identification is highly dependent on expert knowledge, and both morphology and chemical composition can be influenced by environmental conditions or harvest time, often reducing accuracy.

To overcome these limitations, molecular techniques have been increasingly applied, among which DNA barcoding has emerged as a widely used method for species identification based on nucleotide sequence information²². However, standard DNA barcodes (e.g., ITS, matK, rbcL) often show low resolution in the Apiaceae family due to high sequence similarity among closely related species^23,24. Although considerable research has been conducted on species-level identification within genera^25–32, studies focusing on the development of markers for intergeneric discrimination within Apiaceae remain limited. This is particularly critical in the Apiaceae family, where species from different genera (e.g., Ledebouriella, Peucedanum, and Glehnia) are frequently confused in the herbal market due to their morphological resemblance in processed forms, posing significant quality control challenges. Consequently, a transition from single-locus DNA barcoding to systematic barcoding utilizing whole-plastome data is essential. This genome-wide approach enables the identification of specific hypervariable regions that offer superior resolution compared to universal markers, providing a robust phylogenetic framework for distinguishing closely related species.

Recently, High Resolution Melting (HRM) analysis has gained attention as a rapid and highly sensitive alternative. HRM analysis detects subtle sequence differences by analyzing DNA melting curves after PCR amplification, enabling identification of single-nucleotide polymorphisms (SNPs) or insertion-deletion (InDel) mutations^33,34. This approach is fast, cost-effective, and provides high accuracy without the need for sequencing, and has been successfully applied for species identification across animals, plants, and microorganisms^35–37. Notably, its ability to simultaneously analyze large numbers of samples makes it particularly suitable for applications in herbal medicine, where closely related species are often mixed.

In this study, we aimed to assemble the complete plastid genomes (Plastomes) of the 12 medicinal plant species belonging to the Apiaceae family and to develop HRM-based DNA markers for accurate species identification. The markers developed in this study are expected to help address widespread mixing between genera in the herbal medicine market and provide a practical tool for rapid and unambiguous identification of all 12 medicinal Apiaceae species, establishing a framework for systematic barcoding. Moreover, the assembled plastomes provide valuable resources for quantifying genetic variation and reconstructing phylogenetic relationships, thereby contributing to a deeper understanding of evolutionary patterns and taxonomic resolution within Apiaceae. Overall, this study facilitates reliable identification of medicinal plants and provides useful resources for future research on drug development and conservation.

Results

Assembly and structural features of plastomes and 45 S nrDNA sequences in 12 Apiaceae species

Raw sequencing data ranging from 1.13 to 8.70 Gb were generated using the Illumina MiSeq platform and used for simultaneous assembly of plastome and 45 S nrDNA sequences using the dnaLCW approach (Table 1)³⁸. The plastomes were assembled for each sample, comprising a large single copy (LSC) region, a small single copy (SSC) region, and two inverted repeat (IR) regions (Fig. 1A). Total plastome lengths ranged from 146,811 bp (A. dahurica) to 158,510 bp (C. tenuissimum) (Table 1). LSC lengths varied from 84,818 to 93,908 bp, SSC from 17,313 to 17,662 bp, and IR from 17,818 to 27,987 bp, reflecting species-specific variation. The IR junctions included genes ycf2, rpl22, rpl23, and rps19. The overall GC content ranged between 37.5 and 37.7%. C. tenuissimum exhibited the highest gene content with 131 genes, including 86 protein-coding sequences (CDS). Each sample contained 36 or 37 tRNA genes and 8 rRNA genes. The assembled 45 S nrDNA sequences ranged from 5,805 to 5,815 bp. The 18 S rRNA region was consistently 1,809 bp, while the ITS1 region varied from 213 to 245 bp. The 5.8 S rRNA region ranged from 156 to 164 bp and exhibited the highest nucleotide diversity (π) (Fig. 1B), ITS2 from 224 to 235 bp, and the 26 S rRNA region from 3,361 to 3,400 bp (Table S1). These length variations may reflect interspecific differences.

Table 1.

Information of plastome sequence of 12 Apiaceae species.

Scientific Name	Abb.	Common name	Korean medicine name	Chinese medicine name	Raw data (Gbp)	Length (bp)	IR junction Position	LSC (bp)	SSC (bp)	IRs (bp)	G + C (%)	Number of				GenBank Accession
												Genes	CDS	tRNA	rRNA
Angelica gigas	Ag	Korean angelica	Danggui	Zhaoxian Danggui	4.73	146,933	ycf2	93,152	17,605	18,088	37.55	126	82	36	8	PV995367*
Angelica reflexa	Ar	Notopterygium Rhizome	Ganghwal	Qiang Huo	1.15	147,281	ycf2	93,187	17,662	18,216	37.54	126	82	36	8	PV942563*
Glehnia littoralis	Gl	Beach Silvertop	Haebangpung	Bei Sha Shen	4.72	147,477	ycf2	93,496	17,555	18,213	37.51	126	82	36	8	KU866532.1
Angelica dahurica	Ad	Dahurian Angelica	Baekji	Bai Zhi	7.42	146,811	ycf2	93,546	17,629	17,818	37.52	126	82	36	8	PV995368*
Angelica japonica	Aj	Coastal angelica	Gaetganghwal	Rìben Danggui	8.70	154,302	rpl23	86,603	17,547	25,076	37.55	127	83	36	8	PV917076*
Ledebouriella seseloides	Ls	Siler	Bangpoong	Fang Feng	5.05	147,830	ycf2	93,202	17,324	18,652	37.53	126	82	36	8	KU866529.1
Peucedanum japonicum	Pj	Coastal Hog Fennel	Sikbangpung	Bin Hai Qian Hu	1.35	147,525	ycf2	92,563	17,628	18,667	37.54	126	82	36	8	PQ315778.1
Conioselinum tenuissimum	Ct	Korean gobon	Gobon	Gao Ben	1.13	158,510	rpl22	84,874	17,662	27,987	37.61	131	86	37	8	PV942564*
Cnidium officinale	Co	Sichuan Lovage Rhizome	Cheongung	Chuan Xiong	1.36	148,520	ycf2	93,977	17,607	18,468	37.6	126	82	36	8	PV946923*
Levisticum officinale	Lo	Lovage	Gudanggwi	Ou Danggui	3.93	148,547	ycf2	93,908	17,703	18,468	37.58	126	82	36	8	PV942565*
Anthriscus sylvestris	As	Cow parsley	Jeonho	Qian Hu	6.73	154,951	rps19	84,818	17,313	26,410	37.50	130	85	37	8	PV995369*
Bupleurum komarovianum	Bk	Bupleurum root	Siho	Chai Hu	1.52	155,817	rps19	85,736	17,499	26,291	37.66	130	85	37	8	PV995370*

*Newly assembled in this study.

Fig. 1.

Plastome Map and 45 S nrDNA Structure with Nucleotide Diversity (π) Across 12 Apiaceae Samples. (A) The map was drawn using OGDRAW, based on the Angelica gigas plastome. The outer and inner colored boxes indicate genes transcribed in the clockwise and counterclockwise directions, respectively. Colors represent gene identity or functional categories. The red bars at the center of the circle denote the nucleotide diversity (π) values across plastomes of 12 Apiaceae species. (B) Based on the A. gigas nrDNA sequence, the upper panel illustrates the linear arrangement of 45 S nrDNA genes, while the lower panel represents the overall 45 S nrDNA structure. Red bars indicate the positions and values of genetic diversity (π), determined by comparing DNA sequences among the 12 Apiaceae samples.

Sequence variation, nucleotide diversity, and repeat analysis

We aligned the complete plastome and 45 S nrDNA sequences of 12 Apiaceae species representing nine genera to evaluate sequence variation and nucleotide diversity. Based on the multiple sequence alignment, (A) reflexa and G. littoralis exhibited the lowest number of SNPs (398), while (B) komarovianum showed the highest (7,331) (Table 2). Notably, although A. japonica belongs to the same genus as (A) reflexa, it showed more than twice the number of SNPs compared to other genera such as G. littoralis and L. seseloides. The number of InDels ranged from 133 to 973 across species. In the 45 S nrDNA sequences, a similar trend was observed, with (B) komarovianum again showing the highest number of both SNPs and InDels (Table 3).

Table 2.

SNP and InDel counts among 12 Apiaceae plastomes.

Indel SNP	Ag	Ar	Gl	Ad	Aj	Ls	Pj	Ct	Co	Lo	As	Bk
Ag		104	160	185	231	346	338	339	334	324	682	940
Ar	411		135	169	197	350	307	312	306	289	663	926
Gl	598	394		153	190	345	298	316	308	290	657	911
Ad	679	472	494		221	365	321	338	331	316	669	923
Aj	755	544	571	655		315	291	311	322	310	675	941
Ls	994	1254	1288	1342	1476		246	391	390	387	708	946
Pj	1201	981	1001	1069	1053	915		364	357	361	710	937
Ct	1257	1017	1085	1132	1611	2059	1604		188	210	637	899
Co	1407	1204	1350	1399	1063	1465	1217	539		170	635	891
Lo	1356	1115	1291	1340	996	1411	1157	563	716		627	887
As	4060	3852	3921	3958	3928	4283	3976	3766	3761	3685		937
Bk	7527	7348	7414	7424	7444	7786	7439	7380	7302	7215	7971

Table 3.

SNP and InDel counts among 45 S nrDNA sequences of 12 Apiaceae samples.

Indel SNP	Ag	Ar	Gl	Ad	Aj	Ls	Pj	Ct	Co	Lo	As	Bk
Ag	-	1	0	0	1	3	3	5	7	5	7	11
Ar	37	-	1	1	2	4	4	6	8	6	8	12
Gl	38	49	-	0	1	3	3	5	7	5	7	11
Ad	50	55	62	-	1	3	3	5	7	5	7	11
Aj	34	44	46	60	-	2	2	4	6	4	6	10
Ls	44	58	54	69	55	-	0	4	6	4	8	12
Pj	52	64	59	74	61	26	-	4	6	4	8	12
Ct	70	76	86	94	83	76	80	-	2	0	8	14
Co	70	81	83	102	81	81	85	41	-	2	10	15
Lo	108	115	114	129	124	117	122	77	84	-	8	14
As	175	182	181	185	187	170	174	164	176	192	-	11
Bk	277	288	289	298	288	277	278	273	275	302	314	-

Nucleotide diversity (π), calculated using DnaSP v6, ranged from 0 to 0.13111 among the plastome sequences of the 12 Apiaceae species, with an average of π = 0.01617 (Fig. 2). Based on the π values, we identified five highly variable hotspot regions, including both gene-coding and non-coding regions. The trnE gene showed the highest nucleotide diversity (π = 0.13111), followed by the intergenic regions of rps16–trnQ and ndhF–rpl21 (π = 0.05232), trnK–rps16 (π = 0.04922), and ycf1 (π = 0.04576). These highly variable regions may serve as candidate targets for the development of molecular markers.

Fig. 2.

Nucleotide diversity in the plastomes of 12 Apiaceae species. Sliding window analysis was performed to estimate nucleotide diversity using a window size of 600 bp and a step size of 200 bp. Peaks represent regions of high sequence variability, which are potential hotspots for molecular marker development. The x-axis indicates the nucleotide position along the plastome, while the y-axis shows the nucleotide diversity (π) values.

In addition, we analyzed repetitive motifs using REPuter and Tandem Repeats Finder and identified numerous simple sequence repeats (SSRs) and tandem repeats distributed throughout the plastome (Fig. S1). Repeat sequence analysis revealed substantial variation in the number and types of repeats, including tandem, palindromic, forward, reverse, and complementary motifs (Fig. S1A). Among the taxa, L. seseloides and P. japonicum exhibited the highest number of tandem repeats (52 and 50, respectively), while A. dahurica and (A) sylvestris had the fewest (16 and 19, respectively). Palindromic and forward repeats were widely distributed, with C. tenuissimum showing the greatest number of forward repeats (25). Complementary repeats were rare overall, found in more than one instance only in L. officinale and (B) komarovianum. Mononucleotide repeats were the most abundant, ranging from 34 in A. reflexa to 54 in L. officinale (Fig. S1B). Dinucleotide and trinucleotide repeats were less common, with trinucleotide SSRs observed only in A. dahurica and L. seseloides.

Phylogenetic analysis of 12 Apiaceae plastomes and 45 S nrDNA sequences

Plastome sequences from 12 Apiaceae species, including nine newly assembled and three previously assembled in our laboratory³⁹, were analyzed alongside 45 S nrDNA sequences from the same species to construct an nrDNA phylogenetic tree (Fig. 3). The topological congruence between the plastome-based (Fig. 3A) and nrDNA-based (Fig. 3B) trees was then compared. The resulting phylogenies exhibited highly consistent overall topologies, reflecting well-supported clades corresponding to recognized taxonomic groups within Apiaceae. A notable difference was observed in the relative positions of A. dahurica and G. littoralis in the plastome-based tree: A. dahurica clustered with G. littoralis, whereas in the nrDNA tree, G. littoralis grouped with A. japonica.

Fig. 3.

Maximum-likelihood phylogenetic trees of 12 Apiaceae species inferred from plastome and nrDNA sequences. Phylogenetic trees were constructed using the complete plastome and 45 S nrDNA sequences of 12 Apiaceae species, including nine newly assembled plastomes. Both trees were generated using RAxML⁴⁰. The outgroup is Panax ginseng, a member of the Apiaceae family. Bootstrap values greater than 50% are shown above the branches. (A) plastome-based phylogenetic tree. (B) nrDNA-based phylogenetic tree. Species abbreviations: Ag (A. gigas), Ar (A. reflexa), Gl (G. littoralis), Ad (A. dahurica), Aj (A. japonica), Ls (L. seseloides), Pj (P. japonicum), Ct (C. tenuissimum), Co (C. officinale), Lo (L. officinale), As (A. sylvestris), Bk (B. komarovianum). Species belonging to the same genus are connected with the same color.

To evaluate the quality and accuracy of the plastome assemblies, an additional phylogenetic analysis was conducted incorporating both the assembled plastomes and publicly available reference plastome sequences downloaded from the NCBI database (Fig. S1). The phylogenetic tree showed that the assembled plastomes clustered closely with their corresponding reference sequences. This finding reinforces the reliability of the assembly pipeline and confirms the identity of the sampled species.

HRM-based marker development for species identification

To enable rapid and accurate identification of medicinal species within the Apiaceae family, we developed seven HRM markers based on polymorphic regions identified through comparative plastome analysis (Table 4). Although some hypervariable loci exhibited high nucleotide diversity (π), many of them were too variable to allow for the design of stable primers. Therefore, to ensure both amplification efficiency and discriminatory power, we targeted conserved regions that contained one to four species-specific SNPs.

Table 4.

Information of seven developed HRM markers of 12 Apiaceae species.

Primer name	Region	Position (bp) in Angelica gigas	Primer sequences (5’- 3’)	Product size (bp)
Api_1	ycf15-trnV	140,318 − 140,466	F: TTGCTATCCTCCTGCCTAGC R: TGTACCGTGAGAGAGGCAA	149
Api_2	rpl23	86,950 − 87,061	F: ACCTTGACACCAAAGAAGAGT R: CAGACAAAAGTATTCGGTTATTGGG	112
Api_3	psbJ	65,127 − 65,261	F: TCCAAAGAGGAATCCTTCCAG R: TCTCTCTTAGCCCATTCGGA	135
Api_4	psaB	38,861 − 38,961	F: ATTCTTGCCAATAGCCACGC R: GAATAGTTTATCGGTCTGGGCG	101
Api_5	trnA-rrn23S	135,642 − 135,747	F: CACGAGTTGGAGATAAGCGGA R: TGTCTATCGTTGGCCTCTATGG	106
Api_6	petL-petG	66,989 − 67,133	F: TTCTTGGTCATTGAGATTCATGGT R: TCCAGCCAAAGTAACAGGAATTAGA	145
Api_7	atpI	15,459 − 15,567	F: GGTAAAAGGATTGGGGTTGG R: CTGTTGCTTTGGCTTTACTCACG	109

HRM primers were designed to target both coding and non-coding regions, including atpI, ycf15–trnV, psaB, psbJ, rpl23, trnA–rrn23S, and petL–petG (Table 4). These loci exhibited sufficient variability for species discrimination while maintaining sequence conservation across primer binding sites. All seven HRM primer sets successfully amplified their respective target regions across the 12 species tested (Fig. 4). Subsequent HRM analysis generated distinct melting curve profiles for each species. The seven HRM markers produced a total of 12 unique composite genotype profiles (e.g., A. gigas: AAABBBD; (A) reflexa: BAABBBC), each reflecting the combined HRM genotypes across the seven loci and enabling clear discrimination of all 12 Apiaceae species (Table 5). Despite the close phylogenetic relationships and morphological similarities among Angelica species, each was unambiguously distinguished based on its unique combination of melting curve profiles across the seven loci (Fig. 4A–G). Notably, (B) komarovianum, which exhibited the highest number of SNPs in both plastome and nrDNA sequences, displayed distinct HRM patterns at multiple loci, including the Api_1, 3, 4, 6, 7 markers (Fig. 4A, C-D, F-G). In contrast, species such as G. littoralis and A. dahurica displayed similar profiles across several loci, but were still clearly differentiated when all seven markers were combined.

Fig. 4.

HRM analysis of 12 Apiaceae species using seven markers. (A) HRM results with the Api_1 marker, sequence alignment of the variant region, and genotype information. (B-G) HRM results using the Api_2 to Api_7 markers. The apostrophe (‘) indicates that different sequences produced the same HRM result and were classified as the same type. The position information is based on the Angelica gigas plastome sequence.

Table 5.

Genotype information using HRM markers in 12 Apiaceae species.

Marker Species	Api_1	Api_2	Api_3	Api_4	Api_5	Api_6	Api_7	Marker genotype
Ag	A	A	A	B	B	B	D	AAABBBD
Ar	B	A	A	B	B	B	C	BAABBBC
Gl	A	A	A	A	B	B	C	AAAABBC
Ad	A	A	A	B	B	B	C	AAABBBC
Aj	A	A	B	B	B	B	C	AABBBBC
Ls	A	B	B	B	C	B	D	ABBBCBD
Pj	A	B	B	B	B	B	D	ABBBBBD
Ct	A	A	B	B	C	B	C	AABBCBC
Co	A	A	B	B	C	A	C	AABBCAC
Lo	A	A	B	B	C	B	B	AABBCBB
As	A	B	B	A	A	B	C	ABBAABC
Bk	C	A	B	C	B	C	A	CABCBCA

Discussion

This study establishes a systematic molecular barcoding framework that resolves the long-standing taxonomic complexities and identification challenges within the Apiaceae family. By integrating whole-plastome assembly with high-resolution melting (HRM) analysis, we provide the first genomic evidence distinguishing morphologically similar species, such as A. reflexa and A. japonica, and offer a rapid, high-throughput tool for authenticating herbal medicines.

Comparative analysis revealed distinct differences in the distribution of SNPs and InDels across species. In particular, B. komarovianum exhibited the highest number of SNPs and InDels, whereas A. reflexa and G. littoralis showed the lowest levels of variation (Table 3). Even within the same genus, A. japonica had 980 SNPs compared with A. reflexa. This represents approximately 2.5 times greater sequence divergence than that observed with G. littoralis, a species of a different genus. This is consistent with previous reports that the genus Angelica exhibits discordance between morphological similarity and molecular phylogeny^41,42. Such patterns may reflect incomplete lineage sorting, introgression, or taxonomic inconsistencies at the genus level. Similar conflicts between plastome-based and morphology-based classifications have been reported in other plant groups, sometimes revealing that species cluster into clades with members of different genera rather than within their own genus^42,43. Therefore, our findings emphasize the necessity of genome-wide approaches, rather than analyses based on single or a few genes, to ensure accurate species delimitation.

Nucleotide diversity (π) analysis of plastomes identified relatively high variation in regions such as trnE, rps16–trnQ, ndhF–rpl21, trnK–rps16, and ycf1. These hypervariable regions may serve as potential DNA barcoding candidates for detecting interspecific genetic differences. Previous studies have also proposed ycf1, trnK, and rps16–trnQ as useful barcoding markers in the Apiaceae family^24,26,44,45. Our results support these findings and suggest that these regions have broad utility within Apiaceae. Particularly, ycf1 deserves attention as it provides higher resolution than commonly used markers such as matK or rbcL^46,47. Repeats and SSRs analyses also provided interesting results. SSR were abundantly distributed across plastomes, with species-specific differences in their type and frequency consistent with previous studies in Apiaceae^48–51. The SSRs identified in this study may serve as valuable molecular markers for assessing genetic diversity and distinguishing species in future research.

Phylogenetic analyses using both plastome and 45S nrDNA sequences produced largely congruent topologies. Importantly, clustering of assembled plastomes with publicly available reference sequences further validated the accuracy of our assembly pipeline and confirmed species identity (Fig. 3). These findings indicate that plastome and nrDNA data complement each other in reconstructing the phylogeny of Apiaceae and provide valuable references for taxonomy and evolutionary research. Minor incongruences, such as the differing placement of A. dahurica and G. littoralis, may reflect incomplete lineage sorting, introgression, or clade-specific evolutionary histories. Similar phenomena have been reported in various plant lineages^52–54, suggesting that reliance on a single genomic dataset may limit phylogenetic inference. In particular, conflicting signals between maternally inherited plastid DNA and biparentally inherited nuclear DNA are well-documented; for instance, the rice varieties ‘Tongil’ and ‘Milyang 23’ exhibit distinct maternal and nuclear lineages due to their complex hybridization and breeding histories³⁸. Comparison of the newly assembled plastomes with previously reported sequences revealed small sequence differences, confirming both the overall accuracy of the assemblies and minor inter-study variation. Specifically, A. gigas differed by 17 bp⁵⁵, A. dahurica by 107 bp⁵⁶, C. tenuissimum differed by 10 bp compared with the NCBI reference sequence (KT963039.1), C. officinale by 2 bp⁵⁷, L. officinale by 1 bp⁵⁸, (A) sylvestris by 86 bp⁵⁸, and (B) komarovianum by 172 bp⁵⁹. These small differences did not affect overall gene content or genome structure, further confirming the reliability of our assemblies. These sequence differences likely contribute to the minor incongruences observed in phylogenetic placement.

The HRM markers developed in this study successfully distinguished all 12 species across seven loci, thereby enabling systematic barcoding and unambiguous identification. Although conventional DNA barcoding has been applied in Apiaceae^23,60, its reliance on sequencing imposes practical limitations. In contrast, HRM analysis provides systematic, cost-effective, and high-throughput species discrimination⁶¹, making it particularly suitable for addressing issues of adulteration and substitution in the herbal medicine market. The practical utility of these markers extends to regulatory enforcement, quality control, and conservation of valuable medicinal resources. Given that adverse events resulting from misused herbal materials have been frequently reported, our findings suggest that such problems can be mitigated through the application of reliable molecular markers. The seven HRM marker sets developed in this study produced accurate and consistent results for 12 Apiaceae samples analyzed. However, several limitations should be acknowledged. First, as the specimens analyzed in this study were represented by a single individual per species (n = 1), a certain degree of intraspecific plastome variation may exist that was not fully captured in the dataset. Similar intraspecific chloroplast genome diversity has been reported in Apiaceae species such as P. japonicum, as well as in other wild plants^62–66. Therefore, we cannot definitively rule out the possibility that some of the identified SNPs represent variations specific to individuals rather than to species. Second, while the markers successfully discriminated the 12 selected species, their resolving power at the intrageneric level—particularly among other closely related species within the same genera—requires further validation. In diverse wild populations, potential occurrences of false positives, false negatives, or novel genotypic patterns should be anticipated. Future studies incorporating larger sample sizes per species and a broader range of congeneric taxa are necessary to robustly confirm the species-specificity and universal applicability of these HRM markers.

Conclusions

In conclusion, this study provides the comprehensive plastome and nrDNA assemblies of 12 medicinal Apiaceae species, including the first assemblies of A. reflexa and A. japonica, and identifies hypervariable regions, SSRs, and HRM markers with validated utility. These genomic resources and molecular tools contribute to reliable species identification, genetic diversity assessment, phylogenetic research, and the assurance of safety and quality in the herbal medicine market. Furthermore, this study enhances our understanding of genetic diversity, phylogeny, and taxonomy within Apiaceae, and highlights the vital role of molecular markers in systematic species identification, conservation, and the sustainable use of medicinal plants.

Methods and materials

Plant materials and genome sequencing

Fresh leaf materials of 12 Apiaceae species were obtained from the Medicinal Plant Garden and the Suwon Farm of Seoul National University (Seoul, Republic of Korea). The botanical identification of all samples was performed by Prof. Tae-Jin Yang at Seoul National University.

Voucher specimens and DNA samples have been deposited in the Plant Genomics & Breeding Institute, Seoul National University, under the following accession numbers: Ag (IM151201-7), Ar (IM180813-28), Gl (IM120207-3), Ad (IM151201-3), Aj (IM151201-2), Ls (IM120207-1), Pj (IM120207-2), Ct (IM180813-29), Co (IM180813-27), Lo (IM180813-41), As (IM151201-8), and Bk (IM160307-5). All plant materials were collected with permission and utilized in strict accordance with relevant institutional, national, and international guidelines and legislation. Genomic DNA was extracted from leaf tissues (n = 1) using the GeneAll PLANT SV MINI kit (GeneAll Biotechnology Ltd., Seoul, South Korea). The genomic DNA concentrations were examined using agarose gel electrophoresis and quantified using NanoDrop ND-1000 (Nanodrop Technologies, Inc., Wilmington, DE, USA). Genomic DNA from 9 Apiaceae species was used for Next-Generation Sequencing (NGS) on the Illumina MiSeq platform (Illumina, San Diego, CA, USA). The plastome and 45 S nrDNA sequences of G. littoralis (KU866532.1), L. seseloides (KU866529.1) and P. japonicum (PQ315778.1) were used based on the sequences reported in our previous study⁶².

Assembly and annotation of the plastomes and 45 S nrDNA

Plastomes and 45 S nrDNA sequences were assembled through de novo assembly of low-coverage whole-genome sequencing (dnaLCW)³⁸ using CLC Assembly Cell (version 4.6 beta; CLC Inc., Aarhus, Denmark) and validated by manual curation of raw read mapping. First, raw paired-end reads were trimmed using CLC quality trim software with a quality score offset of 33. The trimmed reads were then assembled with an overlap distance ranging from 150 to 500 bp. Initial contigs were extracted using MUMmer⁶⁷ and arranged by comparison to the reference chloroplast sequence from A. gigas (GenBank accession number: MT921985.1) and the reference 45 S nrDNA sequence from Ledebouriella seseloides (KX757775.1) in the National Center for Biotechnology Information (NCBI). The assembled plastome sequences from the 12 Apiaceae species were aligned and compared using the MAFFT 7.0 program to identify polymorphic sites⁶⁸.

Gene annotation was performed using GeSeq⁶⁹, circular maps of the plastome were generated using the OGDRAW program⁷⁰, and manually curated using the ARTEMIS annotation tool⁷¹.

Repeat sequence analysis

Microsatellite repeats were identified using MISA (v2.1; https://webblast.ipk-gatersleben.de/misa/), dispersed repeats (e.g., forward, reverse, palindromic, and complementary repeats) were analyzed using REPuter (https://bibiserv.cebitec.uni-bielefeld.de/reputer/) and tandem repeats were detected using TRF (v4.09; https://tandem.bu.edu/trf/trf.unix.help.html).

Comparative and phylogenetic analyses of plastomes and 45 S nrDNA sequences

The multiple alignment of the assembled plastomes from 12 Apiaceae species and reference sequences downloaded from NCBI GenBank (A. gigas [MT921985.1], A. dahurica [MT921980.1], G. littoralis [KT153022.1], L. seseloides [KT153021.1], P. japonicum [KU866531.1], L. tenuissimum [PQ315778.1], C. officinale [MN832860.1], L. officinale [MT561024.1], (A) sylvestris [MT561042.1], and (B) falcatum [OR508821.1]) was performed using MAFFT 7.0.

From each plastome, coding sequences (CDSs) were extracted using FeatureExtract⁷². Multiple alignments were performed separately for the 80 commonly shared CDSs and the 45 S nrDNA sequences using MAFFT 7.0. Phylogenetic analyses were conducted using randomized accelerated maximum likelihood (RAxML)⁴⁰ with the GTRGAMMA substitution model and 1000 bootstrap replicates based on the aligned plastome and 45 S nrDNA sequences, with Panax ginseng (KM067388.1) used as the outgroup. Nucleotide diversity (π) of the complete cp. genome sequence of the 12 Apiaceae samples was calculated using DnaSP software⁷³. The parameters were set as follows: step size = 400 bp, and window length = 600 bp.

Development and validation of HRM markers

SNP primers for distinguishing 12 Apiaceae species were developed based on the polymorphic sites in the plastome sequences of each sample. The primers were designed using the Primer design tool in NCBI and Primer3 online tool⁷⁴. To ensure the generation of stable and efficient primers, specific design criteria were strictly applied. The target amplicon size was restricted to 100–150 bp to optimize HRM curve resolution. Primer sequences were selected to have a GC content between 40% and 60%, and a melting temperature (Tm) ranging from 58 °C to 62 °C, while avoiding regions prone to secondary structures, hairpin loops, or primer-dimer formation. PCR reactions for HRM analysis against SNP target were performed in 20 µL final volume that is composed of 10 ng of template DNA, 2 µL of 10X inclone Taq buffer, 0.4 µL of 10 mM dNTP, 10 pmol of each primer, and 0.2 µL of Taq DNA polymerase (Inclone biotech Co., Yongin, Republic of Korea), and 1 µL SYTO 9 (ThermoFisher, USA). The PCR protocol consisted of 5 min at 95 °C, followed by 45 cycles of 95 °C for 20 s, annealing at optimal temperatures ranging from 55.3 to 57.6 °C for 20 s, and 72 °C for 20 s. Subsequently, HRM analysis was performed on a LightCycler 96 (Roche Applied Science, Germany) in triplicate. The HRM profile included 95 °C for 1 min, 40 °C for 1 min, and 60 °C for 5 s, followed by a ramp to 95 °C for fluorescence data acquisition. The resulting HRM data were analyzed using the LightCycler 96 Application Software, where genotypes were classified into distinct clusters (A–D) based on variations in melting temperature (Tm) and differences in normalized melting curve shapes.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

JYA conceived and designed the study, contributed to methodology, formal analysis, data curation, visualization, and drafted the original manuscript. JYP and HSP were responsible for investigation and resources. JHL and SKK contributed to formal analysis. YJK and YSL contributed to data curation, methodology, and manuscript review and editing. HSP also contributed to manuscript review and editing. TJY supervised the project, provided conceptualization, funding acquisition, project administration, and contributed to manuscript review and editing. All authors have read and agreed to the published version of the manuscript.

Funding

This study was supported by the Korea Health Technology R&D Project (grant number: HI23C1420) through the Korea Health Industry Development Institute (KHIDI), funded by the Ministry of Health & Welfare, Republic of Korea.

Data availability

The data used in this study are available through NCBI GenBank. The GenBank accession numbers for the plastomes are PV995367, PV942563, PV995368, PV917076, PV942564, PV946923, PV942565, PV995369, and PV995370.

Declarations

Competing interests

The authors declare no competing interests.