Deciphering the Plastome and Molecular Identity of Six Medicinal Amomum Species

1. Introduction

Species identification is crucial in field of biology and ecology, holding widespread importance [1]. It serves as the foundation for ecological research, enabling the understanding of species richness, diversity, and ecosystem health [2]. Additionally, it aids in identifying endangered, invasive, and keystone species within ecosystems, facilitating effective conservation and management strategies [3]. Moreover, species identification plays a crucial role in predicting and preventing infectious disease outbreaks by identifying potential disease hosts and transmitters among wild animal species [4]. In food production industries, species identification ensures authenticity, quality, and safety, preventing fraud and the circulation of substandard products [5]. Furthermore, it has relevance in criminal and forensic cases, aiding in identifying the origin of wildlife products. [6]. Traditional methods of species identification, relying on morphological characteristics, have limitations in discriminating taxa with minimal morphological differences or complex phylogeny. To overcome these challenges, DNA barcoding technology has emerged as an effective advancement.

DNA barcoding is a molecular biology technique for the identification of biological species by examining distinct DNA segments. It uses variations in short DNA sequences to provide rapid and reliable species identification [5,7,8,9,10,11]. DNA barcoding enables the analysis of specific gene regions, aiding in the identification and differentiation of morphologically similar species [5,10]. The concept of DNA barcoding was first proposed by Paul Hebert, who suggested using a small, highly conserved genetic sequence called the “ribosomal RNA gene region” to identify species [5]. Initially, DNA barcoding was widely used in animal, where the gene encoding cytochrome c oxidase I (COI) in mitochondria has a high species differentiation potential, especially in insects, birds and fish [12,13]. Therefore, the COI gene has become the preferred choice for universal DNA barcoding in animals due to its high level of accuracy in species identification [14]. However, in plant mitochondrial genomes, the COI gene shows a high degree of conservation and is not suitable as a DNA barcode selection [15]. In addition, complex evolutionary events such as hybridization, polyploidization, and lineage selection are more common in plants than in animals, further increasing the difficulty of screening fragments suitable for DNA barcoding [16]. Unlike the universal COI gene fragments in animals, DNA barcoding research in plants has undergone a screening process of a large number of fragments [17]. Currently, the internationally recognized universal plant DNA barcodes include four gene regions including ITS (internal transcribed spacer: internal transcribed spacer 1-5.8S-internal transcribed spacer 2), matK, rbcL and psbA-trnH [18]. The selection of these gene regions takes into account the genetic diversity and evolutionary history of the plant kingdom to improve the identification ability and applicability of plant DNA barcodes. But these fragments also have limitations, so Kane and Cronk proposed ultra-barcoding, which uses the complete plastomes for plant species identification [19]. DNA barcoding has significant success in plant species identification and classification and has provided a common standard for the international botanical community [20]. It has been widely used in diverse biological areas such as unveiling hidden species, identifying invasive ones, and elucidating food networks [21]. Moreover, it serves as a reliable method for verifying herbal medicinal products, detecting instances of product substitution, and contamination [22,23,24,25]. Distressingly, it’s not uncommon to find herbs that appear similar being used as adulterants in the commercial herbal arena. Although discerning closely related species using DNA barcoding can pose challenges, the technique excels in distinguishing between species that are morphologically indistinguishable but genetically distinct [26]. In conclusion, DNA barcoding is a valuable tool in biological research, enabling rapid and reliable species identification and classification in diverse organisms.

Amomum Roxb. is the second-largest genus in the Zingiberaceae family after Alpinia, which includes approximately 111 [27] to 150 [28,29] species distributed in tropical Asia and Australia, particularly in Southeast Asia, such as India, Malaysia and Indonesia [29]. In China, Amomum comprises 39 species (29 endemic, one introduced) [29], mainly distributed across provinces like Fujian, Guangdong, Guangxi, Guizhou, Yunnan and Tibet [28]. Among them, six species have been listed in the Chinese Pharmacopoeia [30]. These species encompass (1) A. compactum Solander ex Maton (synonyms: Wurfbainia compacta (Sol. ex Maton) Škorničk. & A.D.Poulsen [31]), (2) A. kravanh Pierre ex Gagnep. (synonyms: A. krervanh Pierre ex Gagnep [32], W. vera (Blackw.) Škorničk. & A.D.Poulsen and A. verum Blackw. [32,33]), (3) A. longiligulare T. L. Wu (synonyms: W. longiligularis (T.L.Wu) Škorničk. & A.D.Poulsen [34]), (4) A. tsao-ko Crevost et Lemarie(synonym: Lanxangia tsao-ko (Crevost & Lemarié) M.F.Newman & Škorničk [35]), (5) A. villosum Lour. (synonyms: W. villosa (Lour.) Škorničk. & A.D.Poulsen [36]) and (6) A. villosum var. xanthioides (Wall.ex Bak.) T.L.Wu & S.J.Chen (synonyms: W. villosa var. xanthioides (Wall. ex Baker) Škorničk. & A.D.Poulsen [37]). They exhibit a diverse range of characteristics and applications. For instance, A. compactum is a widely used culinary spice, and its fruits, leaves and seeds have a wide range of pharmacological activities in traditional medicine, such as antifungal, antibacterial, antioxidant, gastroprotective, anti-inflammatory, immunomodulatory, anticancer, antiasthmatic and acute renal failure [38]. Fruits of A. kravanh have showed antibacterial activity [39]. The active ingredients in A. longiligulare and A. villosum var. xanthioides have antibacterial activity [40,41]. Besides, powerful antioxidant properties of A. villosum var. xanthioides in the treatment of non-alcoholic fatty liver disease (NAFLD) and non-alcoholic steatohepatitis (NASH) [42]. A. tsao-ko has been found to contain antifungal active substances [43] and antioxidant ingredients [44], indicating its potential medicinal properties; recent research also suggests that it has the ability to relieve constipation and could be a promising candidate for developing laxatives in the future [45]. The total flavonoids extracted from A. villosum have shown promising potential for developing new drugs to treat gastric cancer [46]. Chemical components found in the seeds of A. villosum can enhance cellular antioxidant activity, as reported by [47]. Additionally, Chen et al. (2018) have confirmed the potential beneficial effects of A. villosum in the treatment of inflammatory bowel disease [48]. Moreover, Li et al. (2016) have demonstrated that the fresh stems and leaves of A. villosum can be used as high-quality feed for cattle, sheep, and other grass-eating livestock [49]. However, their morphological similarities make it easy to confuse these species with one another, and they are also prone to being replaced by other species within the same genus. Therefore, employing molecular identification through DNA barcoding is crucial for accurately identifying Amomum six species.

The ITS sequence is approximately 500-700 bp long. It exhibits a high degree of conservation, making them applicable across a wide spectrum of biological species, particularly in the case of plants and fungi. The sequencing and analysis of ITS are often characterized by their rapidity and cost-effectiveness, especially when compared to traditional morphological classification methods. Additionally, an extensive repository of ITS is available in public databases, providing researchers with a wealth of reference resources that facilitate expedited species identification and classification. The GenBank database at the National Center for Biotechnology Information (NCBI) hosts an extensive collection of ITS sequences for Amomum and its synonymous plants. As of April 11, 2024, it includes 572 sequences that represent 159 species. This extensive dataset serves as a valuable resource for our DNA barcoding research, providing comprehensive and diverse information. In the identification of medicinal plants and distinguishing them from counterfeits, Selvaraj et al. (2012) found that ITS and specifically ITS1 (internal transcribed spacer 1) are effective DNA barcodes for Boerhavia diffusa Linnaeus [50]. The ITS2 (internal transcribed spacer 2) region has been utilized for the identification of medicinal plants and their closely related species [51], such as within the Polygonaceae A. L. Jussieu family [52] and the Dendrobium Sw. genus [53]. The ITS2 region has been demonstrated to be the most promising universal DNA barcode for Zingiberaceae Martinov family [54]. The super-barcode complete plastomes, as well as the matK and rbcL genes, can effectively distinguish A. compactum, A. longiligulare, and A. villosum [55,56]. Additionally, the matK gene and the psbA-trnH intergenic spacer exhibit high identification efficiency for A. tsao-ko and other Amomum species [57]. Among them, the barcodes that are more effective for molecular identification of Amomum are ITS [57,58,59], ITS1 [60,61], and ITS2 [61,62,63]. These research findings demonstrate the promising potential application of DNA barcoding technology in species identification and classification within Amomum. By using DNA barcoding, researchers can accurately identify and classify different Amomum species, which helps us understand their diversity and evolutionary relationships, and provides effective tools and methods for the protection, sustainable utilization and medicinal value research of Amomum.

In this study, we employed a combination of newly sequenced data and sequences obtained from the NCBI database, including (1) ITS, (2) ITS1, (3) ITS2, (4) complete plastomes, (5) matK, (6) rbcL, and (7) psbA-trnH, to facilitate the calibration and precise identification of six medicinal plants within the Amomum genus. By utilizing DNA barcode technology, we were able to identify different Amomum medicinal species at the molecular level, thereby reducing the potential errors associated with traditional morphological methods. Our findings have the potential to enhance the sustainable utilization and conservation of Amomum resources, facilitate industry development and quality control, and ultimately provide significant scientific and societal benefits.

2. Results

2.1. Plastome Structural Variation, Sequence Divergences, and Hypervariable Regions

All 41 individuals from the six examined Amomum species exhibited a quadripartite structure (Figure 1) and showed limited intraspecific variation in plastome size (Table S1). The complete plastomes of these species ranged in size from 162,678 to 164,332 bp. The lengths of the Large Single Copy (LSC), Small Single Copy (SSC), and Inverted Repeat (IR) regions in the six Amomum species ranged from 87,632 to 89,067 bp, 14,895 to 15,754 bp, and 29,642 to 29,971 bp, respectively (Table S1). There was only slight variation in the total GC content, which ranged from 36.0% to 36.4% (Table S1). However, the GC content was higher in the IR regions (41.0–41.2%) compared to the LSC (33.7–34.1%) and SSC (29.6–30.3%) regions (Table S1). The Amomum plastomes are highly conserved and encode between 121 and 133 genes, including 82 to 87 protein-coding genes, eight rRNA genes, and 30 to 38 tRNA genes (Table S1).

We compared the contraction and expansion of IRs regions at four junctions between the two IRs (IRa and IRb) and the two single–copy regions (LSC and SSC) among six species of Amomum genus (Figure. 2). Figure 2 showed the boundary shifts in the plastomes of the studied Amomum species. Specifically, the LSC/IRb boundary is embedded in the rpl22-rps19 region (except for A. compactum YWB91902-1 and A. kravanh YWB91901-1, which are directly at the rpl22 gene); the IRb/SSC and SSC/IRa boundary is within the ycf1 gene; the IRa/LSC boundary is in the rps19–psbA region. These boundary shifts exhibit independent variations both between and within species.

The nucleotide diversity (Pi) values were calculated with DnaSP to test divergence level within different regions among the six Amomum complete plastomes. The average value of nucleotide diversity (Pi) was 0.00469. The nucleotide diversity (π) value ranged from 0 to 0.02354 across the plastomes, and the most hypervariable region was ycf1 (Figure 3).

2.2. Sequence Characteristics

The matrices characteristics of ITS, ITS1, ITS2, complete plastomes, matK, rbcL and psbA-trnH of six medicinal plants of Amomum were listed in Table 1. ITS2 had the highest percentage of variable sites, but complete plastomes had the most variable sites. The same was true for singleton sites (Table 1). ITS1 had the highest percentage of parsimony informative sites (Table 1).

2.3. Distance based Species Discrimination

Analyses of intra- and interspecific Kimura 2-parameter (K2P) distances identified varying barcoding gaps within six Amomum plants across different datasets. In barcoding gap analysis, the ITS1 and complete plastomes barcodes exhibited the highest discriminatory power, successfully identifying 50% of the species (3 out of 6 species; Table S2; Figure 4; Figure 7). The ITS barcode was the next most effective, identifying 33% of the species (2 out of 6 species; Table S2; Figure 4; Figure 7). The ITS2 and psbA-trnH barcodes could only identify one species each, accounting for 17% of the species (1 out of 6 species; Table S2; Figure 4; Figure 7). The matK and rbcL barcodes were unable to identify any species (Table S2; Figure 4; Figure 7). In ABGD analysis, ITS and ITS1 performed best (3/6; 50%; Table S3) whereas other five performed the least (1/6; 17%; Table S3). The number of generated OTUs varied across ABGD analysis with the different prior intraspecific divergence both in initial and recursive partitions (Table S3).

2.4. Tree based Species Discrimination

In the ITS dataset, due to the abundance of sequences for A. villosum (W. villosa), Maximum Likelihood (ML) and Bayesian Inference (BI) trees were initially constructed for all individuals (Figures S1-S2). Subsequently, three individuals were selected from the A. villosum (W. villosa) branch of the ML tree to participate in the construction of subsequent ITS, ITS1, and ITS2 trees. Similarly, in the matK (Figures S3-S4) and rbcL (Figures S5-S6) datasets, three individuals were selected from the same branch in the ML tree for the construction of subsequent matK and rbcL trees. In both cases, these individuals were chosen from the top, middle, and bottom of the branch to represent the full range of genetic diversity.

The ML and BI topologies derived from six of the seven datasets for the six species were congruent in showing which species were monophyletic (Figure 5; Figure 7 and Figure 8; Figures S7-S16), with the exception of the ITS1 dataset, which differed from the others (Figure 6; Figure 8; Figure S17). Across all datasets, including ITS, ITS1, ITS2, complete plastomes, matK, rbcL and psbA-trnH, A. tsao-ko and all its synonymous individuals formed a monophyletic group, demonstrating the successful identification of A. tsao-ko (Figure 5, Figure 6 and Figure 7; Figures S7-S17). Similarly, A. compactum, along with all synonymous individuals, formed a monophyletic group in both the ITS, ITS1 and complete plastomes datasets (Figure 5, Figure 6 and Figure 7; Figure S7; Figure S10). In the ITS, ITS1 and complete plastomes datasets, individuals of A. kravanh, along with all synonymous individuals, exhibited a monophyletic group (Figure 5, Figure 6 and Figure 7; Figure S7; Figure S10; Figure S17). Overall, the ITS, ITS1 and complete plastomes datasets can successfully identify A. compactum, A. kravanh and A. tsao-ko (3/6; Figure 5, Figure 6 and Figure 7; Figure S7; Figure S10); the ITS2, matK, psbA-trnH datasets can successfully identify A. tsao-ko (1/6; Figures S8-S9; Figures S11-S14); the rbcL dataset cannot identify any species (Figure 8; Figures S15-S16). However, A. longiligulare, A. villosum, A. villosum var. xanthioides and their synonymous individuals didn’t form monophyly in the four datasets (Figure 5, Figure 6 and Figure 7; Figures S7-S17).

3. Discussion

3.1. Plastome Characteristics and DNA Barcodes Performance

The plastomes of Amomum are highly conserved and exhibited a typical quadripartite structure, a characteristic shared with nine species within the subfamily Zingiberoideae [64], Zingiber Boehm. [65], various species of Curcuma L. [66], and other photosynthetic angiosperms [67,68,69]. In six medicinal Amomum plants, the maximum possible species discrimination was 3/6 because A. longiligulare, A. villosum, and A. villosum var. xanthioides were non-monophyletic for both ITS, ITS1 and plastome (Figure 5, Figure 6 and Figure 7; Figure S7; Figure S10).

Taxon-specific markers present a feasible alternative that balances the costs associated with comprehensive super-barcodes, such as whole plastomes, against the limited genetic variability often found in standard barcodes. For the genus Amomum, we pinpointed the most significant mutational hotspots in the ycf1 gene with a π value of 0.02354 (Figure 3), a pattern also observed in other members of the Zingiberaceae family [65,66]. While the four conventional barcodes (ITS2, matK, rbcL, and psbA-trnH) were each only able to reliably identify a single species at most, so the ycf1 gene region could serve as a viable alternative when these standard barcodes are inadequate. Given the financial and temporal demands of full plastome sequencing, this gene region offers a cost-effective and efficient method for future population genetic research on Amomum. This approach aids in the development of a growing database for taxon-specific barcodes.

3.2. Performance Comparison of Species Delimitation Methods

Consistent with previous research [70,71,72], species delimitation outcomes vary with the data and methodologies applied. Among the methods evaluated—ABGD, BG, BI, and ML—ML stands out as the most effective for species identification, closely followed by BI, as illustrated in Figure 8. Additionally, the topological structures produced by ML and BI are largely similar, suggesting that these methods consistently achieve the highest identification rates (Figure 8). Consequently, ML is recommended as the primary choice, with BI as a secondary option. While the identification rates for ABGD and BG differ, ABGD generally outperforms BG (Figure 8), leading to a method ranking of ML > BI > ABGD > BG. Given the demonstrated robustness and efficiency of ML and BI in this study, these methods are recommended as the preferred approaches for species delimitation in DNA barcode-based identification, particularly when employing super-barcodes.

3.3. DNA Barcoding in Six Medicinal Plants within Amomum

Previous studies have indicated that the single plastome fragment is not suitable for the identification of several medicinal plants within the genus Amomum [55,61,63,73,74,75,76]. The complete plastomes have demonstrated a strong capability to differentiate species of Amomum [55]. The results of this study have further validated these findings. The ability to distinguish species of Amomum is enhanced by the length of the complete plastomes sequence, which is approximately 160,000 bp long, and its inclusion of a multitude of informative sites. However, it is important to note that the sequencing and analysis of the complete plastomes are considerably more expensive and resource-intensive compared to ITS sequences. ITS sequencing is more cost-effective and demands fewer computational resources for analysis. Despite its relatively short length of approximately 600 bp, the informative sites within the ITS region can accurately distinguish between A. compactum, A. kravanh, and A. tsao-ko, similar to the capabilities of ITS1. ITS2 can only successfully identify A. tsao-ko. Although ITS2 contains the highest proportion of variable sites, the complete plastomes holds the greatest total number of variable sites (Table 1).

Previous studies have shown that the identification rate of ITS/ITS1/ITS2 is higher compared to the plastome fragments [59,61,77]. This may be because the plastome only contains maternal genetic information [78], while ITS/ITS1/ITS2 contains richer biparental genetic information [79]. ITS sequences typically have multiple copies, which may increase its variability and improve the accuracy of species identification. Conversely, plastome fragments may only have a single copy, limiting its identification capabilities in certain taxa. Compared to this, ITS sequences may have higher variability in these taxa. These taxa may contain hybrid species or show hybridization phenomena, leading to difficulties in species identification using plastome fragments. In this case, ITS sequences may better reflect the genetic differences between the six species, thereby improving the identification rate.

In the seven datasets, some individuals were placed within monophyletic groups (Figure 5, Figure 6 and Figure 7; Figures S1-S17), which may be due to misidentification. The inclusion of some non-target species individuals in the monophyletic branches might be due to errors in species identification, given that the NCBI database has a very wide range of sources. Previously, a number of studies solely relied on distance for species identification. However, subsequent research has indicated that the barcode gap may be a result of human error in under-sampled taxonomic groups [80]. Therefore, when carrying out species identification, we should incorporate other analyses. The relationship between the minimum interspecific distance and the maximum intraspecific distance among the six species, along with the consistency between the ABGD grouping results and the tree results, provides strong evidence to support the species identification and classification of these species.

3.4. Application of NCBI Database

NCBI database has provided comprehensive biological and biomedical information [81], offering a vast collection of genetic sequences, gene expression data, protein structures, and scientific literature. Its user-friendly interface and open access policy promote global scientific collaboration. However, challenges include navigating through the extensive data and ensuring the quality and accuracy of the information due to varying submissions from researchers and institutions. This research was conducted based on a large amount of NCBI data and obtained reliable results. The NCBI database provides great convenience for research.

ITS1 sequence is approximately 100-200 bp in length. In most cases, it does not require sequencing but can be obtained directly through Polymerase Chain Reaction (PCR) amplification. This process is cost-effective and can swiftly obtain the necessary data. Furthermore, abundant ITS sequences of the genus Amomum can be directly extracted from the NCBI database. Through multiple analyses in this study, it has been mutually verified that ITS1 has the highest identification rate. This suggests that in the future identification of these six medicinal plants within the genus Amomum, ITS1 should be considered first.

3.5. ITS vs. ITS1

ITS1 is a part of ITS, both of which can be directly obtained through PCR amplification. The ITS region and/or its subregion (ITS2) have been proposed as a standard DNA barcode marker in fungi [82] and plants [83]. In our study, the identification rate of ITS1 was higher than that of ITS2, which aligned with the view that ITS1 is a better barcode than ITS2 in eukaryotes [84]. Despite the evaluation of ITS1 and ITS2 as meta-barcode markers for fungi [85], their identification efficacy as DNA barcode markers varies across different taxa. In this study, the individuals used across the ITS, ITS1, and ITS2 datasets are consistent. Therefore, this research serves as a reference, suggesting that the ITS1 dataset might be considered first in practical applications when the experimental individuals are identical.

Although ITS is significantly longer than ITS1, there is no noticeable difference in the difficulty of amplification between the two. Even though the ITS dataset had more variable sites than ITS1, it didn’t necessarily mean that it surpasses ITS1 in identification rate. Importantly, the percentage of variable sites within the ITS1 sequence is higher than that of variable sites within the entire ITS sequence. There are many factors that can affect the identification rate, including: the presence of key variable sites that can significantly distinguish species, the amount of recognizable feature information within the dataset, and the size of the differences in the species being identified. In these aspects, the ITS1 dataset may be superior to ITS, and thus, have a higher identification rate.

4. Materials and Methods

4.1. Taxon Sampling

Based on the phylogenetic relationships of the genus Amomum established by Boer et al. [86], we selected close relatives of six target species for our study. Our data collection and analysis focused on six target species, their close relatives, and the synonyms associated with both the target and related species. We sampled 11 individuals from Amomum (Table 2), as well as numerous individuals represented by ITS, complete plastomes, matK, rbcL, and psbA-trnH sequences from both Amomum and its synonymous taxa available on NCBI (Table S4). To download the second-generation sequencing data encompassing these groups, we utilized the prefetch tool in SRA Toolkit v.3.1.0, accessible at https://github.com/ncbi/sra-tools, from the NCBI database. The cut-off date for downloading data from NCBI was April 11, 2024. Detailed species information that we sequenced is listed in Table 2, and all of them have been uploaded to the NCBI GenBank database. Alpinia nigra (Gaertn.) Burtt (MF076960) and Alpinia galanga (L.) Willd. (AF478715) were chosen as outgroups for constructing the matrices of ITS, ITS1 and ITS2 sequences. For the complete plastomes, matK, rbcL and psbA-trnH matrices, A. nigra (MK940826) and A. galanga (MK940825) were selected as outgroups. This selection of outgroups was informed by the research of Gong et al. [63]. We have downloaded 232, 31, 138, 224 and 53 sequences of ITS/ITS1/ITS2, complete plastomes, matK, rbcL and psbA-trnH respectively from NCBI (Table S4). The ITS, matK and rbcL dataset contains numerous instances of A. villosum (synonym: W. villosa). Initially, we constructed a phylogenetic tree using all available data. Subsequently, we selected three individuals from the A. villosum (W. villosa) clade within the tree of A. villosum (W. villosa) based on genetic distance. These individuals, downloaded from NCBI, were chosen for further analysis.

It is worth noting that A. krervanh was revised to A. kravanh [32]. The species name of the data downloaded by NCBI cannot be changed at will, so A. krervanh on NCBI is still A. krervanh, and our own data is named the revised A. kravanh.

4.2. DNA Extraction, Sequencing, Assembly and Annotation

We extracted total DNA from 0.2 g of the gel-dried leaves and herbarium samples using the modified 4 × CTAB method [87]. The quality of DNA was assessed using 1% agarose gel electrophoresis and a NanoDrop^® ND-1000 spectrophotometer. We constructed a DNA library (300-500 bp) using the NEBNext UItra II DNA library prep kit for Illumina, and performed two-end sequencing (2×150 bp) on the DNBSEQ-T7 high-throughput platform, generating a total data amount of no less than 3 Gb. The length of single-ended sequencing reads was 150 bp (sequencing strategy PE150). To convert SRA files downloaded from NCBI into FASTQ format using fasterq-dump-orig from SRA Toolkit v.3.1.0 (https://github.com/ncbi/sra-tools). Then, compress the ‘fastq’ files into ‘fastq.gz’ format suitable for GetOrganelle assembly using the open-source tool pigz v. 2.2.5 (https://zlib.net/pigz/).

The ITS sequence, spanning approximately 600-700 bp, was first assembled utilizing GetOrganelle v.1.7.5.3 [88]. Following assembly, both the resultant FASTG file and the reference from Amomum sericeum Roxb. (KY438097.1) were aligned using the Map function in Geneious v.9.0.2 [89] to prepare the sequence for annotation. Subsequently, annotation was performed through Geneious v.9.0.2[89] with the reference to acquire the ITS sequence. ITS1/ITS2 sequences were then extracted based on annotation information using Geneious v.9.0.2 [89].

The plastome assembly and annotation methods of sequences were conducted following the protocol described by Li et al. [90]. The clean data obtained from high-throughput sequencing were directly assembled using GetOrganelle v.1.7.5.3 [88], and the complete circular plastid genome was automatically generated. In cases where the circular structure could not be obtained, results were visually inspected using Bandage v.0.8.1 [91]. Subsequently, reliable plastid genome contigs or scaffolds were identified by manually removing non-target contigs from the ‘fastg’ file. The selected sequences were then manually edited and spliced to obtain a complete plastid genome. Annotation of the plastid genome was performed using Geneious v.9.0.2 [89], with the published genome of A. krervanh (NC_036935.1) as the reference, and then combined with ORF (open reading frame) for correction. The matK, rbcL and psbA-trnH were extracted using Geneious v.9.0.2 [89] based on annotation information.

The ITS, ITS1, ITS2,complete plastomes, matK, rbcL and psbA-trnH matrices were constructed by aligning the sequences using the Mafft Multiple Alignment plugins in Geneious v.9.0.2 [89]. All annotated sequences have been uploaded in GenBank and assigned accession numbers (Table 2).

4.3. Data Analysis

4.3.1. Plastome Structural Variation, Divergence, and Mutational Hotspot Analyses

We analyzed the characteristics of 41 plastomes of six medicinal Amomum plants, focusing on aspects such as genome size, gene content (including protein-coding genes, tRNAs, and rRNAs), and GC content. We performed comparative analyses on the expansion and contraction of the Inverted Repeats (IR) at the four junctions of the plastomes using Geneious v.9.0.2 [89], and visualized the results with IRscope [92]. To pinpoint hypervariable regions, we carried out a sliding window analysis using DnaSP v.5 [93], with a step size of 200 bp and a window length of 600 bp, identifying the top three sequences as the most variable regions. Finally, we constructed a physical circular map of the plastome with OGDRAW v.1.3.1 [94].

4.3.2. Sequence-Based Analyses

We conducted distance-based analysis using matrices generated from a subset of target and closely related species individuals selected from all individuals of the Amomum genus and its synonyms for tree construction according to Boer et al. [86]. Two primary species delimitation approaches were employed: barcoding gaps (BG) [95] and automatic barcode gap discovery (ABGD) [96]. To investigate the existence of barcoding gaps within each dataset (ITS, ITS1, ITS2, complete plastomes, matK, rbcL and psbA-trnH), we conducted pairwise distance calculations implemented in MEGA-11 [97] using the K2P model. A scatter plot was employed to identify barcoding gaps by visualizing the relationship between the minimum interspecific distance and maximum intraspecific distance for the six species. A species is considered accurately identified when the minimum interspecific distance is larger than its maximum intraspecific distance [98]. The ABGD analysis was conducted using an online platform (https://bioinfo.mnhn.fr/abi/publi c/abgd/), employing three distinct distance models: Jukes-Cantor [JC69], Kimura [K80] TS/TV 2.0 and Simple Distance. The analysis was configured with the following parameters: Pmin = 0.001, Pmax = 0.1, Steps = 10, X = 1.5, Nb bins = 20. The best partition was identified as the one most closely aligning with the delimitation of nominal species among the partitions obtained.

4.3.3. Phylogenetic Tree-Based Analyses

We constructed phylogenetic trees based on ML and BI methods from seven datasets: (1) ITS, (2) ITS1, (3) ITS2, (4) complete plastomes, (5) matK, (6) rbcL, and (7) psbA-trnH sequences. The sequence matrices of each dataset was aligned using MAFFT implemented in Geneious v.9.0.2 [89]. The ML tree was constructed using RAxML v.8.2.11 [99] by the GTRGAMMAI model with 1000 rapid bootstrap replicates. MrBayes v.3.2.7 [100] was utilized for BI analyses runs with 1,000,000 generations, employing the best-fit model specified according to the optimal scheme selected by jModeltest v.2.1.7 [101] using the Akaike Information Criterion (AIC) criteria. Phylogenetic trees were then visualized by tvBOT [102]. When all individuals of the same species and its synonyms cluster into a single clade, we consider it to be successfully identified.

5. Conclusions

We examined plastome structural variations and investigated the efficacy of standard and super DNA barcodes for resolving species boundaries based on within and between species variation within six medicinal Amomum plants. In this study, six medicinal plants of the genus Amomum were molecularly identified using the ITS, ITS1, ITS2, complete plastomes, matK, rbcL, and psbA-trnH sequences. Among these seven sequences, ITS, ITS1 and complete plastomes were effective in identifying A. compactum, A. kravanh, and A. tsao-ko, while ITS2, matK, and psbA-trnH only can successfully identify A. tsao-ko. In contrast, rbcL failed to identify any species. In summary, ITS, ITS1 and complete plastomes demonstrates the highest identification rate, followed by ITS2, matK, and psbA-trnH, with rbcL having the lowest identification rates. In conclusion, considering factors such as cost, for the molecular identification of the six medicinal plants within the Amomum genus, the use of ITS1 is strongly recommended. This study developed reliable molecular identification methods for the genus Amomum, crucial for protecting wild plant resources, rational use of medicinal plants, and preventing resource misuse. In summary, it provided essential molecular tools for species identification and classification, enhancing our understanding of Amomum medicinal plants.

Supplementary Material: The following supporting information can be downloaded at the website of this paper posted on Preprints.org. The Supplementary Material for this article as follows: Tables: Supplementary Table 1. Summary of significant characteristics of six medicinal Amomum plants plastomes, including aspects of genome size, G-C content, and gene number. Supplementary Table 2. Net differences between minimum interspecific and maximum intraspecific distances for six medicinal plants in the genus Amomum across seven datasets, derived from barcoding gap analysis. Supplementary Table 3. The number of putative species recognized by automatic barcode gap discovery (ABGD) analyses of seven datasets using three distance metrics. Supplementary Table 4. All samples of Amomum and their synonyms used in this study (those marked with “*” are individuals sequenced by ourselves, others are downloaded from NCBI). Figures: Supplementary Figure 1. Phylogenetic tree was reconstructed based on the Maximum likelihood (ML) method with the ITS set of all individuals of Amomum and its synonyms. The numbers at nodes indicate ML bootstrap values (BS). Supplementary Figure 2. Phylogenetic tree was reconstructed based on the Bayesian Inference (BI) method with the ITS set of all individuals of Amomum and its synonyms. The numbers at nodes indicate BI posterior probabilities (PP). Supplementary Figure 3. Phylogenetic tree was reconstructed based on the Maximum likelihood (ML) method with the matK set of all individuals of Amomum and its synonyms. The numbers at nodes indicate ML bootstrap values (BS). Supplementary Figure 4. Phylogenetic tree was reconstructed based on the Bayesian Inference (BI) method with the matK set of all individuals of Amomum and its synonyms. The numbers at nodes indicate BI posterior probabilities (PP). Supplementary Figure 5. Phylogenetic tree was reconstructed based on the Maximum likelihood (ML) method with the rbcL set of all individuals of Amomum and its synonyms. The numbers at nodes indicate ML bootstrap values (BS). Supplementary Figure 6. Phylogenetic tree was reconstructed based on the Bayesian Inference (BI) method with the rbcL set of all individuals of Amomum and its synonyms. The numbers at nodes indicate BI posterior probabilities (PP). Supplementary Figure 7. Phylogenetic tree was reconstructed based on the Bayesian Inference (BI) method with the ITS set of selected individuals of Amomum and its synonyms. The numbers at nodes indicate BI posterior probabilities (PP). Supplementary Figure 8. Phylogenetic tree was reconstructed based on the Maximum likelihood (ML) method with the ITS2 set of selected individuals of Amomum and its synonyms. The numbers at nodes indicate ML bootstrap values (BS). Supplementary Figure 9. Phylogenetic tree was reconstructed based on the Bayesian Inference (BI) method with the ITS2 set of selected individuals of Amomum and its synonyms. The numbers at nodes indicate BI posterior probabilities (PP). Supplementary Figure 10. Phylogenetic tree was reconstructed based on the Bayesian Inference (BI) method with the complete plastomes set of selected individuals of Amomum and its synonyms. The numbers at nodes indicate BI posterior probabilities (PP). Supplementary Figure 11. Phylogenetic tree was reconstructed based on the Maximum likelihood (ML) method with the matK set of selected individuals of Amomum and its synonyms. The numbers at nodes indicate ML bootstrap values (BS). Supplementary Figure 12. Phylogenetic tree was reconstructed based on the Bayesian Inference (BI) method with the matK set of selected individuals of Amomum and its synonyms. The numbers at nodes indicate BI posterior probabilities (PP). Supplementary Figure 13. Phylogenetic tree was reconstructed based on the Maximum likelihood (ML) method with the psbA-trnH set of selected individuals of Amomum and its synonyms. The numbers at nodes indicate ML bootstrap values (BS). Supplementary Figure 14. Phylogenetic tree was reconstructed based on the Bayesian Inference (BI) method with the psbA-trnH set of selected individuals of Amomum and its synonyms. The numbers at nodes indicate BI posterior probabilities (PP). Supplementary Figure 15. Phylogenetic tree was reconstructed based on the Maximum likelihood (ML) method with the rbcL set of selected individuals of Amomum and its synonyms. The numbers at nodes indicate ML bootstrap values (BS). Supplementary Figure 16. Phylogenetic tree was reconstructed based on the Bayesian Inference (BI) method with the rbcL set of selected individuals of Amomum and its synonyms. The numbers at nodes indicate BI posterior probabilities (PP). Supplementary Figure 17. Phylogenetic tree was reconstructed based on the Bayesian Inference (BI) method with the ITS1 set of selected individuals of Amomum and its synonyms. The numbers at nodes indicate BI posterior probabilities (PP).