Genetic and metabolite diversity of Sundaland Heptapleurum (Araliaceae) insight into evolutionary and specialized metabolite Authors: Faizal Maulana, Nur Kholilatul Izzah, Hyeji Lee, Jee Young Park, Jong-Soo Kang, Yun Sun Lee, Yeon Jeong Kim, Waras Nurcholis, Tae-Jin Yang Published: 2026 ABSTRACT AbstractBackgroundThe genus Heptapleurum Gaertn (previously treated as Schefflera J.R.Forst. & G.Forst.) within the Araliaceae family is recognized for its significant medicinal value and complex taxonomy. However, an integrated understanding of its evolutionary and metabolite diversity remains unexplored, especially in the Sundaland region (i.e., Java and Sumatra). Here, we integrate genomics and metabolomics to unravel the evolutionary relationships and metabolite diversity of 10 Heptapleurum species from Sundaland.ResultsWe assembled 10 new complete plastid genomes (plastomes) and 45S nuclear ribosomal DNA (nrDNA) sequences, identifying significant variation and potential key molecular markers. Metabolomics identified 152 metabolites, mainly phenolics and terpenoids. Metabolite profiles of H. rhynchocarpum and H. capituliferum were more correlated with phylogeny than with geography; these two species were separate from the main Heptapleurum clade. Four species, H. farinosum, H. longifolium, H. rigidum, and H. fastigiatum, have almost identical plastomes and 45S nrDNA structures, suggesting they may represent closely related species with different phenotypes, as evidenced by distinctive metabolite compositions.ConclusionsCrucially, there is an incongruence between the genetic and chemical phylogenies, underscoring that while chemotaxonomy reflects functional diversity, genetic data remains the definitive standard for evolutionary inference, with the potential for reclassifying H. rhynchocarpum and H. capituliferum. This study provides a foundation for future taxonomic revisions, conservation, and drug discovery of Heptapleurum.Supplementary InformationThe online version contains supplementary material available at https://doi.org/10.1186/s12870-026-09297-w.Keywords: Evolutionary, Genomics, Heptapleurum, Metabolomics, Sundaland Background The Araliaceae, or ginseng family, comprises roughly 45 genera and 1,500 species, primarily distributed in tropical and subtropical regions and ranging in form from herbs to trees and lianas [1, 2]. Genera like Panax, Aralia, Schefflera, and Eleutherococcus are economically and medicinally valuable because of their rich bioactive compounds [3]. The genus Heptapleurum, previously treated as Schefflera, is notable for its wide morphological diversity and complex taxonomy, making it the largest genus in Araliaceae, a family famous for its medicinal uses such as Panax ginseng [4, 5]. The genus Schefflera was long considered the largest and most complex genus in the Araliaceae, encompassing over 600 described species [4, 5]. However, molecular phylogenetic studies have conclusively demonstrated its polyphyletic [4–6]. This finding was later confirmed by several additional studies, which revealed five distinct clades and suggested that Schefflera can be divided into five genera, each originating from a particular region of the world, namely the Neotropics, Oceania, Asia, and Africa-Madagascar [6–10]. Consequently, the genus Schefflera is now restricted to only a few species, including its type species S. digitata, and its closely related species distributed in the Pacific Islands [4, 5]. This re-circumscription necessitated the taxonomic re-evaluation of the remaining clades. For the largest of these, the Asian clade, Lowry & Plunkett formally resurrected the genus Heptapleurum. This taxonomic treatment, supported by extensive phylogenetic work, accommodates the approximately 300 species previously included in Schefflera that belong to this distinct Asian lineage, making Heptapleurum now the largest genus in Araliaceae [5]. Heptapleurum species are found in the Sundaland region of Indonesia and are characterized as evergreen shrubs, trees, or lianas. They have palmately compound leaves and small, greenish-whitish flowers that develop into dark drupes [4, 6, 11]. Ecologically, these plants thrive in the montane environments of Java and Sumatra. Their habitats include humid cloud forests and subalpine shrublands. They act as important structural components and serve as pioneer species in these cool, moist, and often misty areas [5, 11]. Heptapleurum has been used in Asian countries such as Indonesia, China, and Thailand for various medicinal purposes. These uses include wound treatment, anti-inflammatory effects, relief from joint pain, improvement of blood circulation, and treatment of gastrointestinal inflammation [12–14]. One example is the use of H. ellipticum in Kalimantan (Borneo) for wounds [15, 16]. These medicinal properties are attributed to bioactive compounds, including terpenoids, saponins, and phenolics [17]. Despite extensive traditional uses, the metabolite diversity of these species remains largely unexplored, particularly those in Indonesia’s Sundaland (Java and Sumatra). Recent advancements in metabolomics, combined with chemometric and feature-based molecular network (FBMN) analysis, offer powerful approaches for comprehensively characterizing plant metabolites. These methods also help visualize relationships and identify key compounds [18–23]. Recent advances in sequencing have enabled the assembly of plastid genomes (plastomes) and nuclear ribosomal DNA (nrDNA), providing key insights into plant evolutionary relationships and genetic diversity [24, 25]. For example, studies using plastome sequences from H. actinophyllum, H. heptaphyllum, and H. delavayi have clarified phylogenetic relationships within Heptapleurum and the Araliaceae family, with particularly close ties between H. heptaphyllum and H. delavayi [26–28]. While the plastome and 45S nrDNA of H. arboricola have informed phylogenetic studies of the Araliaceae family [29], the genetic and metabolite diversity of many other Heptapleurum species native to Indonesia's Sundaland region (Java and Sumatra) remains largely understudied, making Sundaland-origin species a primary focus for addressing these knowledge gaps. Therefore, this study presents a comprehensive investigation of 10 Heptapleurum species from the Sundaland region to bridge key knowledge gaps. Our objectives are: (1) to sequence the plastome and 45S nrDNA of 10 Sundaland Heptapleurum (from Java and Sumatra) and compare plastomes for evolutionary relationships; (2) to profile its metabolite using UHPLC-Q-Orbitrap HRMS-based untargeted metabolomics; and (3) to integrate genomic and metabolomic data to reveal variation in genetic and metabolite diversity. This research aims to provide a holistic view of Sundaland Heptapleurum, highlight its drug-discovery potential, and inform conservation strategies. Materials and methods Plant materials and data retrieval Fresh leaves of 10 Heptapleurum species, including H. aromaticum (Blume) Boerl., H. capituliferum (Merr.) Lowry & G.M. Plunkett, H. ellipticum (Blume) Seem., H. farinosum (Blume) Lowry & G.M. Plunkett, H. fastigiatum (Miq.) Seem., H. longifolium (Blume) Seem., H. rhynchocarpum (Merr.) G.M. Plunkett & Lowry, H. rigidum (Blume) Hassk., H. rugosum (Blume) Boerl., and H. sibayakense (Merr.) G.M. Plunkett & Lowry were collected from Bogor Botanical Garden and Cibodas Botanical Garden locations in Bogor, Indonesia, as detailed in Supplementary Table S1. The collection process was conducted legally and in compliance with all necessary permits and regulations. A dataset of plant sequences from the wider Araliaceae family was obtained from the NCBI GenBank Database (Supplementary Table S2). Finally, an in-house metabolite database was constructed from previous studies in Heptapleurum metabolite analysis. DNA extraction and sequencing Total genomic DNA was extracted from about 200 mg of frozen, ground leaf tissue using a CTAB method [30]. DNA quality and concentration were evaluated on a NanoDrop 2000 spectrophotometer (Thermo Fisher Scientific, USA) with a cutoff OD 260/280 ratio of 1.7–2.0, and integrity was checked by agarose gel electrophoresis. Following confirmation of DNA quality and integrity, approximately 1 μg of genomic DNA per sample was used for library preparation, which was performed along with sequencing by Phyzen Co. Ltd. (Seongnam, South Korea). Libraries were sequenced on NGS platforms; for Heptapleurum species, the Illumina MiSeq platform was used, generating paired-end reads of 2 × 300 bp. Subsequently, reads with a Phred score below 33 were trimmed using CLC Assembly Cell (CLC Inc., Denmark), ensuring inclusion of only high-quality reads in the downstream bioinformatics analysis. This clarified workflow facilitated accurate genomic characterization and comparative studies of Heptapleurum species. Plastome and 45S nrDNA assembly and annotation Following quality trimming of raw reads, the plastome and 45S nrDNA sequences for 10 Heptapleurum species were reconstructed using a de novo assembly of low-coverage whole-genome sequencing (dnaLCW) [31, 32] within the CLC Genome Assembler platform (ver. 4.6 beta). Assembly parameters utilized overlapping read distances ranging from 150 to 500 bp. Plastome-specific contigs were identified and extracted from the total assembly using MUMmer, with H. arboricola (NC_085755.1) as the reference sequence. These extracted contigs were assembled into draft plastome sequences and manually curated based on mapping validation with the original NGS reads. Concurrently, 45S nrDNA contigs were identified and assembled using the H. arboricola 45S nrDNA sequence (OQ338359.1) to guide the assembly. Annotation of the complete plastome sequences was performed using the GeSeq and CPGAVAS2 web service [33, 34], followed by manual verification and correction in the Artemis program [35]. Circular visualization maps of the plastome were generated using OGDRAW [36]. The constituent regions of the 45S nrDNA locus (18S, 5.8S, 26S rRNA genes, and the ITS1 and ITS2) were annotated via BLAST searches. The new assembled and annotated plastome and 45S nrDNA sequences for the 10 Heptapleurum species were deposited in the NCBI GenBank database under the accession numbers listed in Table 1. Table 1General characteristics of the 10 Heptapleurum plastome and 45S nrDNASymbolSpecies nameOrigin islandGenBankaccessionnumber (plastome/45S nrDNA)Total readsPlastomes45S nrDNAMean coverage (x)Total length(bp)GC content (%)Total genesMean Coverage (x)Total length (bp)GC content (%)HaroH. aromaticumSumatraPX591082/PX5930954,678,038537.87156,70338.271341,688.545,81754.7HcaH.capituliferumSumatraPX591090/PX5931034,876,110317.50156,09238.181342,110.485,82155HelH.ellipticumJavaPX591083/PX5930963,817,238259.99156,71338.18134865.185,81954.7HfarH. farinosumJavaPX591086/PX5930996,110,628279.15156,71038.171341,208.065,81954.8HfasH. fastigiatumJavaPX591087/PX5931004,479,428234.65156,65438.18134355.035,81954.8HloH. longifoliumJavaPX591088/PX5931014,532,624173.03156,63238.17134429.085,81954.8HrhyH. rhynchocarpumJavaPX591091/PX5931044,406,892333.01156,09538.211341,198.785,82155HriH. rigidumSumatraPX591089/PX5931024,016,710301.92156,65438.17134221.965,81854.7HruH. rugosumJavaPX591084/PX5930973,546,086240.70156,62938.21134682.95,82054.6HsiH. sibayakenseSumatraPX591085/PX5930983,800,620153.80156,55138.21134857.255,82054.7Haro = H. aromaticum, Hca = H. capituliferum, Hel = H. ellipticum, Hfar = H. farinosum, Hfas = H. fastigiatum, Hlo = H. longifolium, Hrhy = H. rhynchocarpum, Hri = H. rigidum, Hru = H. rugosum, Hsi = H. sibayakense General characteristics of the 10 Heptapleurum plastome and 45S nrDNA Haro = H. aromaticum, Hca = H. capituliferum, Hel = H. ellipticum, Hfar = H. farinosum, Hfas = H. fastigiatum, Hlo = H. longifolium, Hrhy = H. rhynchocarpum, Hri = H. rigidum, Hru = H. rugosum, Hsi = H. sibayakense Comparative analysis of genetic diversity Plastome and 45S nrDNA sequences were aligned using MAFFT [37]. After the alignment, single-nucleotide polymorphisms (SNPs) and insertion-deletions (InDels) were counted and categorized using in-house Python code. The Krait software identified comparative plastomes in simple sequence repeats (SSRs) or microsatellites within the genomic sequences [38]. The detection parameters were configured to identify repeat motifs meeting minimum length criteria of ≥ 10 units for mononucleotides, ≥ 5 for dinucleotides, ≥ 4 for trinucleotides, and ≥ 3 for tetranucleotide, pentanucleotide, and hexanucleotide repeats. Additionally, larger dispersed repeat structures were characterized using the REPuter online platform [39], specifying a minimum repeat length of 30 base pairs and requiring a sequence identity of 90% or higher (corresponding to a Hamming distance of 3 or less) between repeat copies. The junction sites were comparatively analyzed and visualized across the Heptapleurum plastomes using CPJSdraw [40]. Nucleotide diversity (Pi) was subsequently calculated across the aligned sequences using DnaSP [41] with a 600 bp sliding window and a 200 bp step size. Phylogenetic analysis Phylogenetic relationships were reconstructed using two approaches. First, we compared the 10 newly sequenced Heptapleurum species and outgroup Polyscias fruticosa, using both complete plastomes and 45S nrDNA sequences. Sequences were aligned with MAFFT [37] using the 'auto' strategy and default gap penalties. A Neighbor-joining tree was constructed in MEGA11 [42] using the Kimura 2-parameter model and 1,000 bootstrap replicates to assess consistency among new accessions. Second, to robustly infer genome-scale evolution within the wider Araliaceae family, Maximum Likelihood (ML) phylogenetic analysis was conducted using aligned full-length plastome sequences. These alignments were processed via MAFFT, and the tree was reconstructed using RAxML [43]. We implemented the GTRGAMMA model of nucleotide substitution, which accounts for both across-site rate variation and general time-reversible substitution patterns. Nodal support was rigorously assessed using 1,000 rapid bootstrap replicates. This ML-based topology serves as the primary basis for our evolutionary and taxonomic interpretations. Sample preparation and UHPLC-Q-Orbitrap HRMS analysis Powdered freeze-dried leaf samples, stored at −80ºC in a deep freezer (Biobase, USA), were used for metabolite extraction. For this preliminary integrative study, a single representative biological sample was collected for each of the 10 Heptapleurum species. 100 mg of powder was weighed and transferred into a 10 mL glass vial, followed by the addition of 5 mL 99.9% methanol. The mixture was sonicated for 1 h at room temperature, then filtered through a 0.2 µm filter membrane (Anpel, Shanghai, China). Samples were analyzed on a Vanquish UHPLC system coupled with a Q Exactive Plus Orbitrap High-Resolution Mass Spectrometer (Thermo Scientific, USA) using a 2 µL injection volume. Chromatographic separation was used with an Accucore™ Vanquish C18 + column (1.5 µm particle size, 120 Å pore size, 2.1 × 100 mm). Mobile phases were (A) water with 0.1% formic acid and (B) acetonitrile with 0.1% formic acid. Gradient elution started at 5% B (0–25 min), shifted to 95% B (25–28 min), then returned to 5% B (28–33 min) at 0.2 mL/min. The column was held at 30ºC. The mass spectrometer operated in both ESI + (3800 V) and ESI − (3200 V) modes with electrospray ionization. Mass analysis used a quadrupole-Orbitrap MS (m/z 100–1500), with ion fragmentation energies of 18, 35, and 53 eV. Metabolites identification The resulting raw data were processed using Thermo Scientific Xcalibur software. Data were imported into the Qual Browser, where relevant mass ranges were selected, and filters for both positive and negative ion modes were applied, resulting in the base peak chromatogram. Metabolite identification, the raw data were further analyzed using MzMine software [44]. The data were loaded, and feature detection was performed to identify relevant peaks. Parameters for polarity were set to both positive and negative ion modes, with a retention time range of 0–33.0 min and a noise level threshold of 104. The ADAP chromatogram builder was used, with a minimum intensity threshold of 106 and an m/z tolerance of 0.05 m/z or 5 ppm. Chromatogram deconvolution was applied with baseline cut-off algorithm with baseline level 106 and minimum peak height 3 × 106. Followed by gap filling for the m/z and retention time with m/z tolerance of 5 ppm, and grouping of isotopic peaks with retention time tolerance of 0.25 min. All samples were aligned, and metabolite identification was performed using an in-house metabolite database. Putative metabolite identification was performed (corresponding to Level 2 of the Metabolomics Standards Initiative) by comparing MS and MS–MS spectra against an in-house database and the MetFrag webserver [45]. To avoid redundancy, metabolites detected in both positive and negative modes were merged based on retention time alignment and MS/MS spectral matching, resulting in a final list of 152 non-redundant metabolites. Heatmap, metabolites pathway enrichment analysis, chemotaxonomy, and feature-based molecular networking analysis Chemometrics was performed to analyze metabolic data from Heptapleurum using heatmap, pathway enrichment, and hierarchical clustering analysis (HCA) for chemotaxonomy, all conducted via MetaboAnalyst [46]. Metabolites were annotated with the Kyoto Encyclopedia of Genes and Genomes (KEGG) compound database [47] and matched to the KEGG pathway database. Differentially abundant metabolites were imported into MetaboAnalyst for pathway enrichment analysis. Enrichment results, calculated for significance (-log10(p) > 2) by hypergeometric tests, were presented as bubble plots. To visualize the relative abundances and distribution patterns of known metabolites across the samples, a heatmap was constructed specifically utilizing the identified peaks. To construct the chemotaxonomic dendrograms, the complete dataset of all aligned and filtered metabolic features (all detected peaks) was recovered from the untargeted UHPLC-Q-Orbitrap HRMS. This approach ensures a holistic representation of each species' global chemical fingerprint. HCA was performed using Euclidean distance as the dissimilarity metric to construct the distance matrix. The dendrogram was subsequently generated using Ward’s minimum-variance method. FBMN was performed using the Global Natural Products Social Molecular Networking (GNPS) [48] from MzMine data. A Mantel test evaluated the correlation between interspecific chemical divergence and evolutionary phylogenetic distance. The genetic distance matrix (gen_dist) was built from SNP data to quantify genomic divergence among species. The metabolomic distance matrix (met_dist) was a dissimilarity matrix from holistic peak intensity profiles. Pearson’s correlation assessed the linear relationship between these matrices. Statistical significance was tested using 9,999 random permutations. Matrix transformations and the Mantel test were performed in R with the vegan package's mantel function. Results and discussions Geographical and genomic characteristics of plastomes and 45S nrDNAs Distribution of Heptapleurum species based on Plants of the World Online (POWO), H. aromaticum, H. longifolium, and H. rugosum are endemic to both Java and Sumatra Island, suggesting either a shared ancestral presence on the once-connected Sundaland landmass or effective dispersal capabilities. In contrast, H. fastigiatum and H. rigidum are endemic in Java, while H. Capituliferum, H. rhynchocarpum, and H. sibayakense are endemic to Sumatra. Heptapleurum ellipticum is widespread across Asia, and H. farinosum is found throughout the Sundaland region [5]. However, this study also includes H. rigidum from the Sumatra accession and H. rhynchocarpum from the Java accession, challenging the view from previous studies that it is exclusive to Java or Sumatra. This potential for broader distribution is consistent with the region’s geology, as Sumatra, Java, Borneo, and the Malay Peninsula were periodically connected during Pleistocene glacial periods when lower sea levels exposed the Sunda Shelf, forming a contiguous landmass. This cyclical isolation and connection have been a potent force for plant evolution. When separated, populations on Java and Sumatra experienced restricted gene flow, allowing them to accumulate distinct genetic mutations and adapt to local ecological conditions through processes such as genetic drift and natural selection. Over millennia, such divergence can lead to allopatric speciation, resulting in distinct species endemic to each island, explaining why some Heptapleurum species are found on one island but not the other [4, 5, 49]. Despite their geographic distribution, all 10 Heptapleurum species share palmately compound leaves. Unique species like H. longifolium are immediately identifiable by their exceptionally long, narrow, and drooping leaflets, which create a graceful, weeping appearance. A similar but less extreme form is seen in H. aromaticum, which has slender, lance-shaped leaflets. Heptapleurum ellipticum is a medium-sized shrub with a rounded canopy, with leaves that are ovate to elliptical. Heptapleurum farinosum presents as a slender and upright plant. In contrast, a group is defined by large, broad, and robust foliage, giving the plants a lush and bold look. Within this group, H. rigidum and H. fastigiatum are similar in their large, broad leaflets, while H. rugosum shares this broad-leaf characteristic but is likely distinguished by a wrinkled leaf texture, kinda similar to H. sibayakense. The species with visually very similar characteristics, like H. capituliferum and H. rynchocarpum (Fig. 1a), are difficult to differentiate from the morphology alone. Therefore, a combination of genomics and metabolomics helps identify plant species using genetic and chemical markers, even in the absence of geographical and morphological information. Morphological and genomic characteristics of 10 Heptapleurum species. a Photographs of the studied H. capituliferum (Hca), H. sibayakense (Hsi), H. ellipticum (Hel), H. aromaticum (Haro), H. longifolium (Hlo), H. rigidum (Hri), H. farinosum (Hfar), H. rugosum (Hru), H. fastigiatum (Hfas), and H. rhynchocarpum (Hrhy). b Structural map of the complete plastomes showing gene organization and functional groups. c Schematic of the 45S nrDNA structure The Heptapleurum plastomes exhibit a conserved architecture typical of angiosperms, structured into large single copy (LSC), small single copy (SSC), and two inverted repeat (IR) regions. These plastomes (Fig. 1b) vary in length from approximately 156,092 bp (H. capituliferum) to 156,713 bp (H. ellipticum), with a GC content ranging from 38.1–38.3%. Invariably, they contain 134 genes across the genus, comprising 89 protein-coding genes, 37 tRNA genes, and eight rRNA genes, detailed in Table 1 and Supplementary Table S8. These are quite similar to complete plastome sequences of Heptapleurum species that have been published, such as H. heptaphyllum (156,685 bp) [27], H. delavayi (156.341 bp) [26], H. actinophyllum (156,675 bp) [28], and H. arboricola (156,704 bp) [29]. Simultaneously, the 45S nrDNA repeat unit (Fig. 1c) in Heptapleurum is highly consistent, with each unit transcribed as a single polycistronic precursor RNA. It comprises the 18S rRNA (1,807 bp), 5.8S rRNA (160 bp), and 26S rRNA (3,395–3,399 bp), separated by ITS1 (222–225 bp) and ITS2 (232–234 bp). This results in a total 45S unit length ranging from 5,817 bp (H. aromaticum) to 5,821 bp (H. capituliferum and H. rhynchocarpum), with a stable GC content of 54.6% to 55.0% (Table 1), which is quite similar to the 45S nrDNA of H. arboricola (5,819 bp) [29]. Genetic diversity and phylogenetic relationships The highly conserved plastome of Heptapleurum exhibits variation in SNPs, InDels, and SSRs among closely related species. SNPs are the most abundant form of genetic variation, while SSRs show higher variation than InDels [50]. A comparative analysis of SNPs and Indels in both plastomes and 45S nrDNA (Table 2) provides crucial evolutionary insights. High numbers indicate significant genetic divergence, while low numbers suggest close genetic relationships [9, 51, 52]. For instance, low SNP and Indel counts within the plastome indicate that H. fastigiatum, H. farinosum, H. longifolium, and H. rigidum form a closely related group. Similarly, H. capituliferum and H. rhynchocarpum are closely related. By contrast, H. aromaticum exhibits substantially higher SNP and Indel counts, highlighting its genetic distance and suggesting a more divergent lineage. Analysis of SSRs and long repeats within Heptapleurum plastomes (Supplementary Fig. S1; Supplementary Table S3, S4, S5, S6, and S7) reveals mononucleotide A/T repeats as the most abundant. Dinucleotide AT/AT and tetranucleotide repeats are also common. Long repeats predominantly range from 30 to 39 base pairs and are more abundant in the LSC region than in SSC and IR. These repeats occur especially in intergenic and protein-coding regions, such as accD, rbcL, ycf1, and ycf2 [50, 53]. This supports the assumption of higher mutation rates in non-protein-coding regions [54]. These patterns are consistent with observed phylogenetic relationships (Figs. 2a and 3). Table 2Pairwise comparison of SNP and Indel variations in 10 Heptapleurum species. Values in the lower-left triangle (below the diagonal) represent SNP variations. Values in the upper-right triangle (above the diagonal) represent Indel variations. All values are presented in the format of Plastome/45S nrDNAIndelHaroHelHruHsiHfarHfasHloHriHcapHrhySNPHaro-134/1125/2130/2138/1142/1141/1142/2202/4203/4Hel684/28-61/165/1108/0112/0111/0112/1199/3200/3Hru697/30302/18-21/0104/1108/1107/1108/2181/4182/4Hsi681/31305/1586/5-109/1113/1112/1113/2182/4183/4Hfar610/23466/22486/27487/24-5/04/05/1198/3199/3Hfas612/22468/21488/26489/234/2-3/00/1201/3202/3Hlo611/23467/22487/27488/241/03/2-3/1200/3201/3Hri613/22469/21487/26490/235/21/04/2-201/2202/2Hcap1002/65930/67958/71957/68911/64912/63913/64913/63-2/0Hrhy1011/66939/68967/72964/69917/65918/64919/65919/6416/1-Values in the format 'X/Y' represent variation counts for the plastome and 45S nrDNA, respectively (plastome/45S nrDNA) Pairwise comparison of SNP and Indel variations in 10 Heptapleurum species. Values in the lower-left triangle (below the diagonal) represent SNP variations. Values in the upper-right triangle (above the diagonal) represent Indel variations. All values are presented in the format of Plastome/45S nrDNA Values in the format 'X/Y' represent variation counts for the plastome and 45S nrDNA, respectively (plastome/45S nrDNA) Comparison of (a) phylogenetic trees derived from complete plastid genomes (left) and 45S nrDNA sequences (right) using Neighbor-joining with a bootstrap of 1000 replications. Modeled using the Kimura 2-parameter model, with bootstrap support values indicated at the nodes. b Chemotaxonomic dendrogram based on positive- and negative-ionization-mode metabolite profiles Maximum likelihood phylogram showing the phylogenetic relationships of new Heptapleurum species (*) within the Araliaceae family. The tree was constructed based on complete plastome sequences with 1000 bootstrap replicates. The newly sequenced Heptapleurum species are indicated with asterisks (*). Major clades and subfamilies are highlighted with colored bars. Bootstrap support values are shown at the nodes, and branch lengths are proportional to the number of substitutions per site (see scale bar) Comparative analysis of the IR junction sites, namely JLB (LSC/IRb), JSB (IRb/SSC), JSA (SSC/IRa), and JLA (IRa/LSC) (Fig. 4). Notably, the rps19 gene shows variability, being 285 bp long in H. aromaticum, whereas in the other nine species, it is 279 bp. Similarly, the distance between the JLA and JLB junction and the rpl2 gene varies across species, ranging from 60 to 121 bp. Considerable variation is also observed at the JSB junction, particularly concerning the ndhF gene (2217 and 2226 bp), which terminates 18, 20, 21, 22 bp from the boundary. Fragment of the ycf1 gene, whose length and position relative to JSB differ among species. The JSA junction consistently falls within the ycf1 gene, resulting in varying lengths, whereas the JLA junction shows minor shifts relative to the LSC region. Despite these fluctuations, the lengths of rpl2, trnN, and trnH genes remain conserved across all species. Ten Heptapleurum species could be distinguished into five groups of A-E that are related to their phylogenetic (Figs. 1a and 2). Comparative analysis of the IR boundaries of 10 Heptapleurum plastomes. The diagram shows the junction sites between the LSC, SSC, and IR regions across the 10 species. It highlights variations in gene content and lengths at the boundaries Nucleotide diversity (Pi) analysis is essential for understanding genetic variation and quantifying the average nucleotide differences per site between sequences. It also helps identify regions under selective constraints, diversifying selection, or rapid adaptation. This information is crucial for molecular marker development and phylogenetic resolution [50, 53]. To further explore this, the Pi landscape across aligned Heptapleurum plastomes was analyzed using a sliding window (600 bp window, 200 bp step). This analysis revealed that single-copy regions (LSC and SSC) generally have higher and more variable Pi values, whereas the highly conserved IR regions show markedly lower diversity (Fig. 5). Hotspots of nucleotide diversity (Pi > 0.015) were identified between petB and petD within the LSC. Additional hotspots were found in a region encompassing fragments of ycf1 and ndhF near the IRb/SSC boundary, and within the ycf1 gene in the SSC/IRa region. Together, these three regions, with their elevated sequence variation, hold high potential for developing informative molecular markers for species discrimination among closely related Heptapleurum. Nucleotide diversity (Pi) of 10 Heptapleurum plastomes. The plot shows the sliding window analysis of Pi across the aligned plastomes. The x-axis represents the position in the aligned genome, while the y-axis indicates the Pi value. Regions with high Pi values indicated by the red line are potential hotspots for developing molecular markers A comparison of phylogenetic analyses for a new assembly of 10 Heptapleurum, using P. fruticosa as outgroups and two distinct genomic datasets (plastome and 45S nrDNA), showed identical groupings (Fig. 2a). Both trees place H. rhynchocarpum and H. capituliferum as sisters. Next, H. aromaticum appears as a sister to the remaining Heptapleurum, with 100% bootstrap support. Both datasets recover a well-supported group of H. ellipticum, H. sibayakense, and H. rugosum. Additionally, the relationships among H. farinosum, H. longifolium, H. rigidum, and H. fastigiatum are resolved (100%) and are close. This correlates with the low SNP and Indel variation among these species (Table 2), suggesting they may be a single species with different phenotypes [55]. A broader phylogram of complete plastome sequences (Fig. 3) within Araliaceae supports the previous analysis (Fig. 2a). This phylogram shows a clear topology, with branch lengths reflecting evolutionary divergence and genetic distance among lineages. The family is divided into two subfamilies with substantial evolutionary separation: Hydrocotyloideae (Hydrocotyle group) and Aralioideae (including the Greater Raukaua, Aralia-Panax, Asian Palmate, and Polyscias-Pseudopanax groups). This classification remains robust and consistent with prior studies [1–3, 29, 56–58]. A key contribution is the inclusion of new plastome assemblies from 10 Heptapleurum species (marked by asterisks). These data, shown by their branch lengths, reveal complex evolutionary relationships and varied sequence divergence within the Asian Palmate group. Crucially, the phylogenetic separation of H. rhynchocarpum and H. capituliferum from the core Heptapleurum clade, resolved here as sister to Kalopanax, strongly suggests that these two species should be evaluated further in studies. These two species revise the previous study, which still shows the Heptapleurum as monophyletic [59, 60]. This placement is separate from the other Heptapleurum species, including others like H. arboricola, H. heptaphyllum, H. ellipticum, H. rugosum, H. sibayakense, H. rigidum, H. fastigiatum, H. longifolium, H. farinosum, H. actinophyllum, H. aromaticum, and H. delavayi. The main Heptapleurum lineage, comprising H. delavayi, H. aromaticum, and H. actinophyllum, is not closely related to itself. For instance, the tree strongly supported a cluster containing H. arboricola, H. ellipticum, H. sibayakense, and H. rugosum, suggesting a shared recent ancestor. The other close relationships cluster around H. farinosum, H. longifolium, H. rigidum, and H. fastigiatum, which have almost identical plastome structures. The tentative identification of specialized metabolites The chromatograms (Fig. 6a, b) show metabolic profiles of various Heptapleurum species from the Sundaland region, analyzed in positive and negative ionization modes. The positive-ion chromatogram (a) shows more well-defined peaks than the negative-ion chromatogram (b), suggesting easier detection or higher abundance of some metabolites in positive-ion mode. In total, 102 metabolites were identified in the positive mode and 88 in the negative mode, with no overlap, making 152 specialized metabolites identified overall. Supplementary Tables S9 and S10 provide the lists of metabolites, with more found in positive ionization mode. Most identified metabolites were phenolics and terpenoids. Chromatograms of 10 Heptapleurum species in (a) positive ionization and (b) negative ionization modes. Each chromatogram shows the intensity of the detected ions as a function of retention time, revealing distinct peaks corresponding to metabolites within the samples. Heatmaps illustrating the distribution of metabolites across the 10 Heptapleurum species in (c) positive ionization and (d) negative ionization modes. The heatmaps display individual metabolites in the rows, while the columns represent the different species analyzed. Deep red indicates a higher relative signal intensity, whereas blue indicates a lower one Example of fragmentation in terpenoid based on the identified metabolites in positive ionization on scheffleside L which is a triterpenoid saponin with the chemical formula C42H66O14. The analysis begins with the protonated molecule [M + H]⁺ observed at m/z 795.4512 (Supplementary Table S9). A major initial fragmentation pathway involves the cleavage of the glycosidic bond linking the triterpenoid aglycone to the sugar chain. This cleavage generates the prominent fragment ion at m/z 439.3564, corresponding to the protonated aglycone [C30H47O2]+. The sugar chain, likely after dehydration, gives rise to the fragment ion at m/z 321.0832 [C12H17O10]+, representing the disaccharide portion. Further fragmentation of the aglycone ion with m/z 439.3564 occurs, likely via a Retro-Diels–Alder (RDA) type reaction common in triterpenoids, leading to the ion at m/z 249.1848 [C16H25O2]+, which represents the rings of the aglycone structure. Additionally, fragmentation within the sugar produces smaller characteristic ions, such as the one observed at m/z 85.0290 [C5H5O2]+, likely derived from the terminal sugar unit. These specific fragment ions and their corresponding structural assignments strongly support the identification of scheffleside L [61–65]. The metabolites identified in Sundaland Heptapleurum have also been detected in other Heptapleurum species. Phaseollin is identified in H. rugosum, H. capituliferum, H. aromaticum, and H. farinosum, and it has also been identified in H. heptaphyllum [66]. Scheffleside L was first isolated from H. leucanthum [67], and in this study, it was identified in H. ellipticum, H. rhynchocarpum, H. sibayakense, H. capituliferum, and H. aromaticum. Regarding ursolic acid, it has been identified in H. fastigiatum, H. rigidum, H. longifolium, and H. aromaticum, where it has been isolated in H. heptaphyllum [68], S. mannii [69, 70], and H. arboricola [71]. Asiaticoside, a metabolite of Centella asiatica, was also first isolated from H. heptaphyllum [72]; in this study, it was identified in H. ellipticum. Schekwanglupasides B identified in H. leucanthum [73], and mannioside metabolites from S. mannii [69] have been identified in Sundaland Heptapleurum species. Sinapic acid, (-)-falcarinol, and ginsenoyne K are metabolites identified in Panax ginseng [74–76]. It is also identified in Sundaland Heptapleurum, as shown in Supplementary Tables S9 and S10. This study provides insight into the metabolite diversity of 10 Heptapleurum species without isolating individual metabolites. Specialized metabolite diversity among Heptapleurum species A heatmap is a data visualization technique that uses color gradients to group data by the amount or density of data [77]. The heatmap (Fig. 6c, d) with 102 metabolites (110 with duplication) was generated in positive ionization mode, while 88 (98 with duplication) were visualized in negative mode. The heatmaps displayed distinct clustering patterns based on relative metabolite signal intensities. Specific clusters of highly abundant metabolites were observed for certain species, such as H. ellipticum, suggesting potential species-specific characteristic metabolites (Fig. 6c, d). For instance, the deep red box color of heptaellipside A (M60 at ESI + and M57 at ESI-), a rare 2,28-bidesmosidic lupane-type saponin that was previously isolated in the same species, was prominently detected in H. ellipticum while remaining undetected in the other nine species (see different species in Supplementary Tables S9 and S10) at a Rt of 11.36 min in positive ionization mode [78]. It was also identified in the negative-ion analysis at 11.37 min. Scheffleside L was detected in five specific species, and it showed higher relative intensities in H. capituliferum (M78 at ESI +, M77 at ESI-), H. ellipticum (M59 at ESI-), and H. rhynchocarpum (M78 ESI +, M77 ESI-) compared to H. sibayakense (M77 at ESI-) and H. aromaticum (M77 ESI-). This suggests that the sampled individuals of the three species had higher relative abundances of this compound than those of the other two species. Similarly, asiaticoside (M58 at ESI +) displayed a uniquely strong signal in H. ellipticum. The heatmap of ursolic acid (M100 at ESI +) across four species showed a higher relative abundance in the H. longifolium sample and comparatively lower levels in the other three species. Furthermore, the comparative analysis identified metabolites uniquely detected in each mode; examples exclusive to positive ionization included asiaticoside (M58) and ginsenosyne K (M83), whereas sinapic acid (M2), schekwanglupaside B (M98), and schweinfurthin E (M54) were among those solely identified in negative ionization. Analyzing both positive and negative ionization can help identify a broader range of metabolites. However, to better analyze metabolomic relationships and classify species by all detected metabolites (known and unknown), chemotaxonomic dendrograms (Fig. 2b) were constructed. These dendrograms use metabolite profiles from positive (ESI +) and negative (ESI −) ionization modes. The trees revealed significant incongruence, which is normal [29, 79]. This discrepancy is because each ionization mode is selective for different metabolite classes. For example, asiaticoside (M58) appeared only in ESI +, while sinapic acid (M2) was unique to ESI − (Fig. 6c, d). Thus, each dendrogram represents a classification based on a different subset of the metabolome. This is similar to using methanol (polar) versus hexane (non-polar) as extraction solvents, which yields different chemical profiles and chemotaxonomic groupings. FBMN analysis of Sundaland Heptapleurum revealed metabolite diversity, constructing networks based on MS/MS spectral similarity and grouping structurally related metabolites into clusters, such as 3α-Hydroxylup-20(29)-ene-23,28-dioic acid with schekwangsienin, β-pinene with limonene, and scheffleside J with scheffleside L (Fig. 7a). Numerous unclustered nodes, including heptaellipside A from H. ellipticum [78], suggest the presence of unique or novel metabolites, especially among the many unknown unclustered compounds. The formation of distinct clusters for major secondary metabolites, such as phenolics and terpenoids, aligns with the genus's phytochemical profile. The complementary use of both ionization modes maximizes coverage of the metabolome. Feature-based molecular networking of 10 Heptapleurum (a) Positive ionization (b) Negative ionization. The node colors and pie charts within the networks visually represented this metabolite's distribution and relative concentration across the different Heptapleurum species. The cluster form is based on similarity in its fragmentation pattern Metabolic pathway enrichment analysis of Heptapleurum species from Java and Sumatra (Supplementary Fig. S2) found strong enrichment of flavonoids (mainly flavone and flavonol groups), phenylpropanoids, sesquiterpenoids, and triterpenoids. This points to their high relative abundance and potential role in characterizing the chemical profiles of these species. This result aligns with the known phytochemistry of the Araliaceae family, which contains many terpenoids and phenylpropanoid metabolites [80]. The prominence of flavonoids suggests they are crucial to the global metabolic fingerprint and should be studied further for their diversity and biological activities. These four pathways likely underpin the unique metabolite diversity of Sundaland Heptapleurum. The prominence of flavonoids and terpenoids in these Sundaland Heptapleurum species likely reflects the evolutionary legacy of adaptive responses to the environmental pressures in their native montane habitats. In wild high-altitude environments, plants are exposed to intense UV-B radiation. Flavonoids, especially flavones and flavonols, act as UV screens and antioxidants. Diverse terpenoid profiles naturally defend against specialized herbivores or microbial pathogens in moist cloud forests. The specimens in this study were sampled from managed ex-situ botanical garden collections. The strong detection of these pathway metabolites suggests that these adaptive chemical defenses are deeply conserved at the genetic level. Baseline metabolic divergence was observed among these genetically similar species, even when grown in localized cultivation environments. This highlights how intrinsic evolutionary adaptations shape specialized metabolism [4, 5, 49, 81–83]. Comparative analysis of genetic and metabolite diversity A comparison between the genetic phylogenies (Fig. 2a) and the chemotaxonomic dendrograms from the metabolomic data (Fig. 2b) revealed notable incongruence. The different clustering patterns suggest that the evolution of distinct metabolic pathways can proceed at different rates and in different directions, even within closely related species [81]. This effect is seen in the distinct metabolic profiles among H. farinosum, H. longifolium, H. rigidum, and H. fastigiatum, despite their highly similar phylogenetic positions. The sampled individual of H. longifolium exhibited a broader detected metabolite diversity than the other three species. The heatmaps (Fig. 6c, d) clearly visualize this. For example, H. longifolium shows a distinctly richer profile and a strongly elevated relative abundance of ursolic acid (M100 at ESI +). In contrast, its signal is lower in the other three species. Heptapleurum aromaticum is the earliest branching lineage within the Heptapleurum group and forms a sister clade to the other species (Fig. 2a). Its chemical profile in the heatmaps (Fig. 6c, d) is also distinct from that of the others. The sister pair, H. sibayakense and H. rugosum (Fig. 2a), shows the complexities of chemotaxonomy. Genetically, they are each other's closest relative. However, although they share some annotated metabolic traits in the heatmaps (Fig. 6c, d), the holistic chemotaxonomic trees (Fig. 2b) do not group them together. Instead, the dendrogram groups them with other species based on shared, likely convergent, metabolic features. This divergence likely results from the chemotaxonomic tree using the full global profile—including unknown and unannotated metabolites—which drives broader clustering patterns. However, a more thorough integration of the datasets reveals that phylogenetic relationships are also a powerful predictor of chemical profiles, in some cases more so than the historical geographical separation between Java and Sumatra. The most compelling evidence for this is the sister pair H. capituliferum and H. rhynchocarpum. Genetically, they are a distinct, strongly supported pair (Fig. 2a). This genetic link is strikingly mirrored in their metabolites; the heatmaps (Fig. 6c, d) show these two species share unique signal intensity patterns, such as elevated relative abundances of scheffleside L (M78 ESI + and M77 ESI-), distinguishing them from their geographical neighbors. This pattern strongly suggests that their unique metabolite profile is a deeply conserved, inherited trait linked to their shared evolutionary lineage, rather than a simple convergent adaptation. Conversely, the broader discordance seen across the rest of the genus demonstrates that significant chemical diversity can evolve even in the absence of deep genetic divergence. Because these plants were cultivated in the shared environment of a botanical garden, these unique, specialized metabolomes likely represent the evolutionary legacy of historical adaptations to their specific native ecological niches. Furthermore, the presence of shared metabolites among distantly related species may indicate convergent evolution, in which different lineages evolved to produce similar compounds in response to shared ecological pressures in Sundaland. These evolutionary mechanisms explain why chemical diversity can evolve independently of the highly conserved plastid genome [49, 82, 83]. To provide quantitative support for this observed genus-wide discordance, a Mantel test was employed. The results showed a weak, non-significant correlation between genetic and metabolic distances (Mantel r = 0.242, p = 0.0821). We hypothesize that this lack of strict global correlation results from rapid historical adaptation to different wild microhabitats [81, 82], which shifted chemical profiles faster than plastid genomes diverged. However, this remains a hypothesis that requires future population-level genomic testing across natural distributions. The integrated analysis underscores that while genetic data remains the definitive standard for inferring evolutionary history, its combination with metabolomics provides a much richer understanding of functional diversity. Specifically, the chemical divergence among Heptapleurum species is driven by at least two potential forces. The first is recent, rapid historical adaptation, which leads to distinct metabolite profiles among genetically similar species, such as H. farinosum, H. longifolium, H. rigidum, and H. fastigiatum. The second is deep evolutionary history, resulting in conserved, lineage-specific chemical signatures, such as those shared between H. rhynchocarpum and H. capituliferum. This dual analysis is therefore crucial for guiding future taxonomic revisions and for understanding the evolution of specialized metabolites in this genus. Taken together, these findings challenge the current taxonomic framework of the Asian Palmate group and underscore the complex evolutionary history of Araliaceae. Building on the foundational multi-omics baseline from conserved specimens, future research can move into broader genomic and ecological studies. Integrating full nuclear genomes will be key to resolving evolutionary relationships and clarifying the reclassification of distinct lineages, especially for H. rhynchocarpum and H. capituliferum. Future studies should also include broader population-level sampling of wild Heptapleurum in their natural habitats. This expanded sampling and advanced nuclear sequencing will reveal the full phylogenetic breadth of the Sundaland lineage and intra-specific metabolic variation in the wild. These steps will help finalize robust taxonomic revisions for this genus. Conclusions This study provides a comprehensive investigation of the genetics and metabolites of 10 Heptapleurum species from Sundaland, leading to several key conclusions. First, characterizing 10 new plastomes and 45S nrDNA sequences is a valuable genomic resource. Comparative analysis identified substantial genetic variation in SNPs, InDels, SSRs, and nucleotide diversity hotspots. Genetic data support the need for future taxonomic reassessment, placing H. rhynchocarpum and H. capituliferum in a distinct clade sister to Kalopanax, different from the main Heptapleurum clade. Second, the metabolomic analysis revealed a rich metabolite diversity, dominated by phenolics and terpenoids, and identified characteristic metabolites for each species. Finally, integrating the genetic and metabolite datasets revealed notable incongruence. Chemotaxonomic clustering did not always mirror evolutionary relationships, highlighting that genetic data remains the standard for evolutionary history, while chemical profiles reflect functional diversity. This functional diversity appears driven by rapid historical adaptation within a genetically similar species complex, resulting in different metabolite compositions, and by the deep evolutionary history of the distinct, conserved chemical profiles of the species H. rhynchocarpum and H. capituliferum, which align with their phylogenetic placement. This research delivers new insights that advance our understanding of Heptapleurum evolution and taxonomy. It maps its metabolite diversity and uncovers its promising potential for future drug development and conservation research. Supplementary Information Supplementary Material 1: Supplementary Fig S1. Comparative analysis of SSRs and long repeats of Heptapleurum plastomes (a) Number of SSRs of six types (b) Number of SSRs in different regions (c) Number of SSRs in different repeat units (d) Number of long repeats of four types (e) Number of long repeats of different lengths. Supplementary Fig S2. Metabolites pathway enrichment analysis in Heptapleurum species from (a) Java (b) Sumatra. Analysis revealed several pathways with high statistical significance (-log10(p) > 2) and considerable pathway impact. Supplementary Material 2: Supplementary Table S1. Plants sample information. Supplementary Table S2. Plastome and 45S nrDNA data used for phylogenetic tree. Supplementary Table S3. Number of SSRs of six types. Supplementary Table S4. Number of SSRs in different regions. Supplementary Table S5. Number of SSRs in different repeat units. Supplementary Table S6. Number of long repeats of four types. Supplementary Table S7. Number of long repeats of different length. Supplementary Table S8. List of genes in plastome of Heptapleurum. Supplementary Table S9. Putatively identified metabolites of Sundaland Heptapleurum in positive ionization. Note: Hel = H. ellipticum; Hfar = H. farinosum; Hfas = H. fastigiatum; Hri =H. rigidum; Hlo = H. longifolium; Hru = H. rugosum; Hsi = H. sibayakense; Hca = H. capituliferum; Hrhy = H. rhynchocarpum; Haro = H. aromaticum. Supplementary Table S10. Putatively identified metabolites of Sundaland Heptapleurum in negative ionization. Note: Hel = H. ellipticum; Hfar = H. farinosum; Hfas = H. fastigiatum; Hri = H. rigidum; Hlo = H. longifolium; Hru = H. rugosum; Hsi = H. sibayakense; Hca = H. capituliferum; Hrhy = H. rhynchocarpum; Haro = H. aromaticum. Acknowledgements The authors sincerely thank the National Research Foundation of Korea (NRF), Korea International Cooperation Agency (KOICA), Seoul National University (SNU), IPB University, and National Research and Innovation Agency (BRIN). Authors’ contributions F.M. contributed in conceptualization, methodology, data analysis, writing – original draft. W.N. and N.K.I contributed in methodology and resources. H.L. and J.Y.P. contributed in the methodology and data analysis. J.S.K., Y.S.L., Y.J.K contributes in writing – review and editing. T.J.Y., contributes in supervision, methodology, and writing – review and editing. All authors read and approved the final manuscript. Funding The author(s) declare that financial support was received for the research of this article from the National Research Foundation of Korea (NRF) grant funded by the Korea government (MSIT) (RS-2025–02216304) and KOICA-ICAB (Korea International Cooperation Agency-IPB-SNU-Center for Agriculture and Bioscience) as part of project, “Capacity Building for the Center for Agriculture and Bioscience of IPB University in Indonesia.” Data availability The original contributions presented in the study are included in the article/Supplementary Material. The following newly sequenced data used in this research are accessible from NCBI, with GenBank accession numbers for plastomes: PX591082, PX591083, PX591084, PX591085, PX591086, PX591087, PX591088, PX591089, PX591090, and PX591091. Alongside with 45S nrDNAs data are PX593095, PX593096, PX593097, PX593098, PX593099, PX593100, PX593101, PX593102, PX593103, and PX593104. Further inquiries can be directed to the corresponding author. Declarations Ethics approval and consent ot participate Not applicable. Consent for publication Not applicable. Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Contributor Information Waras Nurcholis, Email: wnurcholis@apps.ipb.ac.id. Tae-Jin Yang, Email: tjyang@snu.ac.kr. References Associated Data Supplementary Materials Supplementary Material 1: Supplementary Fig S1. Comparative analysis of SSRs and long repeats of Heptapleurum plastomes (a) Number of SSRs of six types (b) Number of SSRs in different regions (c) Number of SSRs in different repeat units (d) Number of long repeats of four types (e) Number of long repeats of different lengths. Supplementary Fig S2. Metabolites pathway enrichment analysis in Heptapleurum species from (a) Java (b) Sumatra. Analysis revealed several pathways with high statistical significance (-log10(p) > 2) and considerable pathway impact. Supplementary Material 2: Supplementary Table S1. Plants sample information. Supplementary Table S2. Plastome and 45S nrDNA data used for phylogenetic tree. Supplementary Table S3. Number of SSRs of six types. Supplementary Table S4. Number of SSRs in different regions. Supplementary Table S5. Number of SSRs in different repeat units. Supplementary Table S6. Number of long repeats of four types. Supplementary Table S7. Number of long repeats of different length. Supplementary Table S8. List of genes in plastome of Heptapleurum. Supplementary Table S9. Putatively identified metabolites of Sundaland Heptapleurum in positive ionization. Note: Hel = H. ellipticum; Hfar = H. farinosum; Hfas = H. fastigiatum; Hri =H. rigidum; Hlo = H. longifolium; Hru = H. rugosum; Hsi = H. sibayakense; Hca = H. capituliferum; Hrhy = H. rhynchocarpum; Haro = H. aromaticum. Supplementary Table S10. Putatively identified metabolites of Sundaland Heptapleurum in negative ionization. Note: Hel = H. ellipticum; Hfar = H. farinosum; Hfas = H. fastigiatum; Hri = H. rigidum; Hlo = H. longifolium; Hru = H. rugosum; Hsi = H. sibayakense; Hca = H. capituliferum; Hrhy = H. rhynchocarpum; Haro = H. aromaticum. Data Availability Statement The original contributions presented in the study are included in the article/Supplementary Material. The following newly sequenced data used in this research are accessible from NCBI, with GenBank accession numbers for plastomes: PX591082, PX591083, PX591084, PX591085, PX591086, PX591087, PX591088, PX591089, PX591090, and PX591091. Alongside with 45S nrDNAs data are PX593095, PX593096, PX593097, PX593098, PX593099, PX593100, PX593101, PX593102, PX593103, and PX593104. Further inquiries can be directed to the corresponding author.