Comparative plastome analyses and phylogenetic insights of Blumea DC Authors: Pingxuan Xie, Junli Xie, Changliu Shao, Kailang Mu, Zhigang Ju, Xiaomin Tang, Yunheng Ji, Hanjing Yan, Yuan Yuan, Yuxin Pang Published: 2026 ABSTRACT BackgroundBlumea DC. comprises approximately 100 species with significant morphological diversity. Yet its plastome has rarely been systematically investigated. Previous studies had divided the genus into three main clades: the B. lacera clade, the B. densiflora clade, and the B. balsamifera clade; however, the phylogenetic positions of these clades remain ambiguous. Furthermore, the monophyly of B. formosana, B. sinuata, B. megacephala, B. aromatica, B. axillaris, and B. hieraciifolia is unresolved and requires further investigation. The exact phylogenetic position of Cyathocline Cass., which has been subsumed into Blumea, also remains unclear.MethodsTo investigate plastome features in Blumea, plastomes from 23 species and two varieties (including 23 newly sequenced samples and two publicly available representatives) were compared. Comparative analyses assessed structural and sequence variation, divergence hotspots, simple sequence repeats (SSRs), and codon usage bias. Additionally, phylogenetic inference was performed using a dataset of 47 complete plastomes and 38 nrDNA sequences to reconstruct the backbone phylogeny and address the aforementioned phylogenetic problems.ResultsAll Blumea plastomes exhibited a typical quadripartite structure with sizes ranging from 150,779 bp to 151,281 bp and contained 113 unique genes (79 protein-coding genes, 30 tRNA genes, and four rRNA genes). IRscope and Mauve analyses revealed minimal structural variation among plastomes. Despite the genes flanking each junction being identical, the junction between LSC, SSC, and IRs regions was classified into four types based on the variations in the distances between genes and their respective junctions. Nine divergent hotspot regions were identified as candidate DNA barcodes for future species identification. Between 79 and 95 SSRs were detected per plastome, predominantly in the large single-copy region and mainly comprising mononucleotide repeats. The genus displayed mono-, di-, tri-, tetra-, penta-, and hexa-nucleotide repeats, with specific types and quantities varying among species. Codon usage bias analysis indicated conservation in preferred codon types, numbers, and RSCU values. Phylogenetic analyses consistently supported the division of Blumea into four clades: the B. lacera clade, B. balsamifera clade, B. densiflora clade, and B. stricta clade. Besides, the monophyly of B. formosana, B. sinuata, B. aromatica, and B. hieraciifolia was well supported.ConclusionsThis study conducts the first large-scale comparative analysis of Blumea plastomes to date, systematically revealing the conservation and specificity of plastome features of the genus. Phylogeny based on plastome and nrDNA dataset has provided an enhanced phylogenetic framework, and preliminary clarified the phylogeny of the genus. Moreover, the monophyly of taxa including B. formosana, B. sinuata, B. megacephala, B. aromatica, B. axillaris, and B. hieraciifolia was well examined. In summary, this study provided substantial informative genetic data pertinent to Blumea and offered new insights into the phylogeny of Blumea, laying the foundation for subsequent taxonomic, systematic, and identification studies. Introduction Blumea DC. (Asteraceae) comprises approximately 100 species, primarily distributed in tropical Asia, with a few in Australia and Africa (Anderberg and Eldenäs, 2007). The genus includes annual or perennial subshrubs, shrubs, and herbs, and many taxa (e.g., B. balsamifera (L.) DC., B. riparia DC., B. aromatica DC., etc.) are of great medicinal value (Editorial Committee of Chinese Materia Medica of the National Administration of Traditional Chinese Medicine, 1999). Due to its morphological diversity, the classification of Blumea has been controversial. Randeria (1960) proposed a widely adopted classification system, dividing the genus into seven sections (Semivestitae, Macrophyllae, Sagittatae, Hieraciifoliae, Paniculatae, Oxyodontae, and Dissitiflorae) based morphological characteristics. However, molecular studies have failed to support this arrangement, instead, grouping Blumea into three main clades: the B. lacera clade, B. densifora clade, and B. balsamifera clade (Pornpongrungrueng et al., 2007; Pornpongrungrueng et al., 2009; Chung et al., 2022).,Besides, these phylogenetic analyses based on single or multiple DNA loci (including ITS, psbA-trnH, and trnL-trnF) have produced inconsistent evolutionary trees, making the relationships among these clades unclear. Moreover, poor resolution and statistical support in these studies have left phylogenetic relationships within Blumea unresolved. Previous research has also failed to confirm the monophyly of B. formosana Kitam., B. sinuata (Lour.) Merr., B. megacephala (Randeria) C. C. Chang & Y. Q. Tseng, B. aromatica, B. lacera (N. L. Burman) DC., B. axillaris (Lam.) DC., B. saxatilis Zoll. & Moritzi, and B. hieraciifolia (Spreng.) DC (Pornpongrungrueng et al., 2007; Pornpongrungrueng et al., 2009; Chung et al., 2022). The genus Cyathocline Cass. was incorporated into Blumea based on molecular and karyotypic data (Li et al., 2014), with C. purpurea (Buch.-Ham. ex D. Don) Kuntze described as the synonym of B. stricta (DC.) Anderb. & Bengtson (Rajbhandari et al., 2024). Yet, the exact phylogenetic position of Cyathocline within Blumea is still unresolved. Therefore, to address these issues, a more comprehensive phylogenetic analysis with extensive sampling and genomic data is necessary. To date, plastid genome (plastome) and nuclear ribosomal DNA (nrDNA) derived from genome skimming data are widely employed in phylogenetic studies, frequently exhibiting high analytical performance (Park et al., 2022; Xia et al., 2022; Fu et al., 2024). Plastomes of Blumea have also been reported, displaying a typical quadripartite structure comprising one large single-copy (LSC) region, one small single-copy (SSC) region, and two inverted repeat (IR) regions (Abdullah et al., 2021; Zhao et al., 2021; Cao et al., 2026). However, only a few plastomes of this genus have been deposited in GenBank, and systematic comparative or phylogenetic analyses based on plastomes remain lacking. Moreover, the number of nrDNA sequences of Blumea is also limited. In this study, we newly sequenced 35 accessions from 21 species and two varieties, as well as three outgroup species to obtain complete plastomes and nrDNA sequences. Combined with nine Blumea plastomes downloaded from GenBank, a final dataset comprising 47 complete plastomes and 38 nrDNA sequences was used. Our primary objectives are as follows: (1) investigate the plastome features of Blumea; (2) reconstruct an enhanced backbone phylogeny and preliminarily elucidate the phylogenetic relationships within the genus; and (3) examine the monophyly of B. formosana, B. sinuata, B. megacephala, B. aromatica, B. axillaris, and B. hieraciifolia. Materials and methods Plant sampling, DNA extraction, and sequencing Fresh leaves from 38 samples were collected and dried with silica gel, 35 of which represent 21 species and two varieties of Blumea. The identifications of the samples were conducted by Professor Yuxing Pang, following the Flora Reipublicae Popularis Sinicae (FRPS) (Editorial Committee of the Flora of China, 1979), as well as relevant literature on new Blumea species (Li et al., 2025). Voucher specimens were deposited in the Herbarium of Guizhou University of Traditional Chinese Medicine, with collection details provided in Supplementary Table 1. Genomic DNA was extracted from 100 mg of silica gel-dried plant tissue using the CTAB method (Doyle and Doyle, 1987). Total DNA was fragmented into 150 bp for pair-end library preparation utilizing a NEBNext® Ultra™ Illumina DNA Library Prep Kit (NEB, USA), in accordance with the manufacturer’s instructions. All genomic data were sequenced on the Illumina NovaSeq 6000 platform at Shenzhen Huitong Biotechnology Co., Ltd. (Shenzhen,China). Assembling and annotation After filtering the raw data using fastP v0.15.0 (Chen et al., 2018) with parameters -n 10 and -q 15, high-quality reads were obtained. Subsequently, these high-quality reads were employed to assemble plastome and nrDNA through assembled using GetOrganelle v1.7.7.1 (Jin et al., 2020) with the default settings, employing the plastome (MZ424304) and nrDNA (KF443294) sequences of B. balsamifera as reference and seed, respectively. The assembled plastomes were edited and corrected using Bandage v0.9.0 (Wick et al., 2015) then annotated using the CPGAVAS2 (Shi et al., 2019). Additionally, Geneious v10.2.3 (Kearse et al., 2012). was used to manually adjust the annotated plastome. Circular maps of the plastomes were generated using OrganellarGenomeDRAW (OGDRAW) (Greiner et al., 2019). The nr DNA sequences of Blumea were annotated in Geneious v10.2.3 (Kearse et al., 2012), with the nrDNA sequence of B. balsamifera (KF443294), Cynara cardunculus var. scolymus (L.) Fiori (XR_003067690, XR_003067462) as the references. GenBank accession numbers for the 38 newly generated plastomes and nrDNA sequences are listed in Supplementary Table 1. Combined with nine additional plastomes retrieved from GenBank (Supplementary Table 2), a total of 47 plastomes and 38 nrDNA sequences were used in follow-up analyses. Genomic comparative analysis Plastomes representing 23 species and two varieties were selected for subsequent comparative analyses (Supplementary Table 3). The junction regions of the LSC (large single copy region), IRs (inverted repeat regions), and SSC (small single copy region) among plastomes were compared and plotted using CPJSdraw (Li et al., 2023). After removing one copy of the IR region, the general structural characteristics of Blumea plastomes were further analyzed using Mauve v2.3.1 (Darling et al., 2004). Furthermore, the sequence divergence of the Blumea plastomes was assessed using mVISTA (Frazer et al., 2004) under the Shuffle-LAGAN alignment mode, with the plastome of B. balsamifera set as the reference. Nucleotide variability values (pi) for the aligned Blumea plastomes were calculated using DnaSP v5.0 (Librado and Rozas, 2009), with the window length and step size set to 400 bp and 200 bp, respectively. Coding and non-coding regions with pi value ≥ 0.018 were extracted using Geneious v10.2.3 (Kearse et al., 2012), and the number of variable sites and parsimony-informative (parsim-info) sites was assessed using MEGA X (Kumar et al., 2018). Simple sequence repeats (SSRs) of Blumea plastomes were identified via MISA (http://pgrc.ipk-gatersleben.de/misa/). The minimum repeat thresholds were set to 10, 5, 4, 3, 3, and 3 for mono-, di-, tri-, tetra-, penta-, and hexa-nucleotide repeats, respectively. Additionally, the codon usage patterns of the core protein-coding genes (PCGs) in the plastomes were analyzed using the Sequence Manipulation Suite Version 2 (SMS v2; https://www.detaibio.com/sms2/index.html), and the corresponding relative synonymous codon usage (RSCU) values were calculated. Phylogenetic analyses To infer the phylogeny of Blumea species, phylogenetic trees were reconstructed based on plastomes and nrDNA with Laggera crispata (Vahl) Hepper & J. R. I. Wood, Elephantopus scaber L., and Duhaldea cappa (Buch.-Ham. ex DC.) Anderb. were selected as outgroups according to Zhang et al. (2019) and Li et al. (2014) (Supplementary Tables 1, 2). All plastomes and nrDNA sequences were aligned using MAFFT v7.450 (Katoh and Standley, 2013) implemented in Geneious v10.2.3 (Kearse et al., 2012) with the default settings, and then manually adjusted. The characteristics of each alignment were analyzed using MEGA X (Kumar et al., 2018). Phylogenetic trees were reconstructed using both maximum likelihood (ML) and Bayesian inference (BI) methods, with the optimal nucleotide substitution model (GTR+I+G4) determined by ModelTest-NG v1.2.2 (Darriba et al., 2020). Maximum likelihood trees were then constructed with RAxML-NG (Kozlov et al., 2019), with 1000 bootstrap replicates. Bayesian inference phylogenetic trees were performed using MrBayes v3.2.7 (Ronquist et al., 2012) running 1,000,000 Markov chain Monte Carlo (MCMC) generations, with sampling performed every 100 generations and the first 25% of generations disregarded as burn-in for each matrix. Stationarity was attained when the average standard deviation of split frequencies was less than 0.01. Phylogenetic trees were visualized and edited using FigTree v1.4.2 (Rambaut, 2014). Results Recovery of complete Blumea plastomes The summary of the low-coverage genome sequencing and assembly of plastomes was presented in Supplementary Table 4. The plastomes of Blumea ranged in length from 150,779 bp to 151,281 bp and exhibited a typical quadripartite structure (Figure 1; Table 1), consisting of one large single copy (LSC) region (82,478–82,897 bp), one small single copy (SSC) region (18,353–18,510 bp), and two inverted repeat (IR) regions (24,907–24,987 bp). The total GC content of the plastomes was 37.5%–37.6%. All Blumea plastomes encoded 113 unique genes, including 79 protein-coding genes, 30 transfer RNA (tRNA) genes, and 4 ribosomal RNA (rRNA) genes (Table 1; Supplementary Table 5). The features of nrDNA were listed in Supplementary Table 6. Circular maps of Blumea plastomes. The inside and outside of the circle are genes transcribed clockwise and counterclockwise, respectively. The genes belonging to different functional groups are color-coded. The dashed area in the inner circle shows GC content. Table 1Comparison of plastome features among Blumea taxa in this study.Samples/Accession numberSpeciesSize (bp)Total GC content (%)Gene number (unique)GenomeLSCSSCIRTotalPCGsrRNAtRNAcp001 B. balsamifera 151,17182,72918,46824,98737.6%11379430cp002 B. aromatica 151,18682,81418,48024,94637.5%11379430cp003 B. aromatica 151,18682,81418,48024,94637.5%11379430cp007 B. stricta 151,08782,66818,49924,96037.5%11379430cp008 B. martiniana 151,25182,87818,50124,93637.5%11379430cp009 B. lanceolaria 151,17082,81918,47724,93737.5%11379430cp010 B. sagittata 151,14482,78618,46424,94737.5%11379430cp014 B. sinuata 150,78882,53018,35424,95237.6%11379430cp015 B. sinuata 150,79082,53218,35424,95237.6%11379430cp018 B. aromatica 151,18682,81418,48024,94637.5%11379430cp019 B. henryi 151,26982,88318,50824,93937.5%11379430cp020 B. axillaris 151,05482,76218,44024,92637.5%11379430cp022 B. calcicola 151,16982,79318,48424,94637.5%11379430cp024 B. megacephala 151,07782,75218,43524,94537.5%11379430cp026 B. aromatica 151,18682,81418,48024,94637.5%11379430cp028 B. oblongifolia 151,06782,75718,42024,94537.6%11379430cp029 B. clarkei 151,01782,70518,41224,95037.5%11379430cp030 B. sinuata 150,77982,52218,35324,95237.6%11379430cp034 B. napifolia 151,23282,89718,43124,95237.5%11379430cp038 B. megacephala 151,07182,75718,42424,94537.6%11379430cp039 B. axillaris 151,04382,75518,43824,92537.5%11379430cp041 B. megacephala 151,07982,76518,42424,94537.5%11379430cp044 B. sessiliflora 150,92582,65618,35524,95737.6%11379430cp047 B. axillaris 151,04982,76118,43824,92537.5%11379430cp051 B. riparia 151,04682,74018,41824,94437.6%11379430cp069 B. eberhardtii 151,11882,78918,41724,95637.5%11379430cp073 B. hieraciifolia 150,87482,55818,39024,96337.5%11379430cp074 B. hieraciifolia 150,87182,55718,39024,96237.5%11379430cp079 B. fistulosa 150,99282,68518,40724,95037.5%11379430cp104 B. sericans 151,21182,87618,43124,95237.5%11379430cp113 B. hieraciifolia 150,86582,55118,39024,96237.5%11379430cp118B. densiflora var. hookeri151,27282,88618,51024,93837.5%11379430cp122B. densiflora var. densiflora151,26482,88018,51024,93737.5%11379430cp127 B. formosana 151,26082,88518,50324,93637.5%11379430cp128 B. formosana 151,25682,87518,50524,93837.5%11379430PX394520 B. axillaris 151,04382,75518,43824,92537.5%11379430PX404823 B. axillaris 151,06382,77718,43824,92437.5%11379430NC_077558 B. balsamifera 151,17082,74018,46624,98237.6%11379430PX404818 B. formosana 151,28182,89418,50924,93937.5%11379430PX404820 B. megacephala 151,04282,73618,41824,94437.6%11379430PX404822 B. sinuata 150,78382,52418,35524,95237.6%11379430BK013127 B. balsamifera 151,17682,74618,46624,98237.6%11379430BK013128 B. oxyodonta 150,99782,74518,43824,90737.5%11379430BK013129 B. tenella 150,82982,47818,43924,95637.6%11379430 Comparison of plastome features among Blumea taxa in this study. Structural features across plastomes The junction between LSC and IRb (JLB) was located between the rps19 and rpl2 genes across all taxa; the junction between IRb and SSC (JSB) was downstream of the ndhF gene; the junction between SSC and IRa (JSA) was located within the ycf1 gene; and the junction between IRa and LSC (JLA) lay between the rpl2 and trnH genes (Figure 2A). (A) Junctions of the LSC/SSC/IR regions of the Blumea plastomes; (B) Four types of junctions of the LSC/SSC/IR regions. JLB, junction line between LSC and IRb; JSB, junction line between IRb and SSC; JSA, junction line between SSC and IRa; JLA, junction line between IRa and LSC. Furthermore, based on the distances of the rps19, ndhF, ycf1, and trnH genes to each junction (Supplementary Table 7), the junction patterns of Blumea plastomes could be broadly classified into four types (Figure 2B). Type I was characterized by a 10 bp distance between rps19 and JLB, a two bp distance between ndhF and JSB (one bp in a few taxa), a 573 bp (574 bp in a few taxa) overlapping region between ycf1 and IRs, and a four bp distance between trnH and JLA (three bp in a few taxa). This type mainly included B. axillaris, B. clarkei Hook. f., B. eberhardtii Gagnep., B. fistulosa (Roxb.) Kurz, B. hieraciifolia, B. megacephala, B. napifolia DC., B. oblongifolia Kitam., B. oxyodonta DC., B. riparia, B. sericans (Kurz) Hook. f., B. sessiliflora Decne., B. sinuata, and B. tenella DC.; Type II featured a 13 bp distance between rps19 and JLB (25 bp in a few taxa), mostly 40 bp or 29 bp between ndhF and JSB (37 bp and 41 bp in a few taxa), a 564 bp ycf1-IR overlapping region, and a two bp distance between trnH and JLA (13 bp in a few taxa), and was observed in B. aromatica, B. calcicola XiongLi & K.W. Luo, B. densiflora var. densiflora DC., B. densiflora var. Hookeri (C. B. Clarke ex Hook. f.) C. C. Chang & Y. Q. Tseng, B. formosana, B. henryi Dunn, B. lanceolaria (Roxb.) Druce, B. martiniana Vaniot, and B. sagittata Gagnep.; Type III had an 11 bp distance between rps19 and JLB, 16 bp between ndhF and JSB, a 581 bp ycf1-IR overlapping region, and a three bp distance between trnH and JLA (two bp in a few taxa). This type was restricted to the B. balsamifera; Type IV was only present in the B. stricta, with a 13 bp distance between rps19 and JLB, a 3 bp distance between ndhF and JSB, a 574 bp ycf1-IR overlapping region, and a one bp distance between trnH and JLA. Mauve analysis revealed a highly conserved gene order in Blumea plastomes, with no gene inversion or rearrangement detected (Supplementary Figure 1). Comparison of sequence variation mVISTA analysis showed overall high sequence similarity across the genus (Figure 3). Sequence similarity was significantly lower in non-coding regions than in coding regions, most notably in intergenic spacers (IGSs). Among them, four coding regions (matK, ndhF, ycf1, and rpoC2) and 15 non-coding regions (trnH-psbA, trnK-rps16, rps16-trnQ, trnS-trnC, trnC-petN, petN-psbM, trnY-rpoB, ndhA intron, trnT-psbD, accD-psaI, psbE-petL, ndhF-rpl32, rpl32-trnL, ccsA-ndhD, and rps15-ycf1) displayed relatively high differences. mVISTA visualization of the alignment for the 25 Blumea plastomes. The result sliding-window analysis revealed that the LSC and SSC regions harbored most high-Pi peaks, whereas the IRs region showed low nucleotide diversity (Figure 4A). Using a Pi threshold of 0.018, nine highly variable regions (ycf1, trnH-psbA, trnC-petN, trnY-rpoB, ycf2-trnL, ndhF-rpl32, rpl32-trnL, ccsA-ndhD, and ndhG-ndhI) were identified as divergence hotspots for Blumea. Among these, ycf1 had the highest number of variable sites (313) and parsimony-informative sites (168). Although trnC-petN (108) and rpl32-trnL (89) ranked second and third in variable sites, their parsimony-informative sites (35, 36) were fewer than those of ndhF-rpl32 (47) and trnY-rpoB (40) (Figure 4B). (A) Sliding window analysis of the 25 Blumea plastomes to show high variable regions. X-axis, position of the midpoint of a window; Y-axis, nucleotide variability values (p) of each window. The dashed line indicates that the p value is equal to 0.018; (B) Number of variable sites and parsim-info sites in the alignment of divergence hotspots. Characterization of SSRs and codon usage bias A total of 53–68 simple sequence repeats (SSRs) were detected across the 25 Blumea plastomes (Supplementary Table 8). B. megacephala harbored the most SSRs (68), while B. sagittata had the fewest (53). These SSRs were predominantly distributed in the LSC region (39–54), with fewer in the SSC (4–8) and IR regions (6–10) (Figure 5A). All Blumea plastomes contained mono-, di-, tri-, and tetra-nucleotide repeats. Pentanucleotide repeats were only found in B. axillaris, B. oxyodonta, B. tenella, B. henryi, B. martiniana, B. densiflora var. densiflora, B. densiflora var. hookeri, and B. formosana, whereas hexanucleotide repeats were restricted to B. henryi (Figure 5B). Analyses of simple sequence repeats (SSRs) in 25 Blumeas plastomes: (A) distribution of SSRs in LSC, IRS, and SSC regions; (B) number of different SSR types; (C) distribution of non-shared SSR motifs across different species. Among all SSRs, mononucleotide repeats were the most abundant (40–51), predominantly A/T motifs. Tetranucleotide repeats were the second most abundant (6–10), with dominant motifs AAAT, AATA, ATTG, TATT, TTAT, and TTTC. Dinucleotide (3–5) and trinucleotide (2–6) repeats were relatively rare (Supplementary Table 8; Figures 5A, B). Additionally, 17 non-shared SSR motifs were identified across taxa, and several motifs exhibited species-specific distributions (Supplementary Table 8; Figure 5C). Specifically, the G motif was detected in B. densiflora var. hookeri, B. densiflora var. densiflora, B. formosana, B. henryi, B. martiniana, B. calcicola, B. aromatica, B. lanceolaria, and B. sagittata; the TG motif occurred in B. oxyodonta, B. axillaris, B. hieraciifolia, B. tenella, B. sinuata, B. sessiliflora, B. clarkei, B. fistulosa, B. napifolia, B. sericans, B. megacephala, B. riparia, B. oblongifolia, and B. eberhardtii; the ATTT motif was found in B. clarkei, B. fistulosa, B. napifolia, B. sericans, B. megacephala, B. riparia, B. oblongifolia, and B. eberhardtii; and the TATAT motif was only present in B. densiflora var. hookeri, B. densiflora var. densiflora, B. formosana, B. henryi, and B. martiniana. Notably, some unique SSR motifs were taxon-specific, such as TTTAAA in B. henryi, ATAAA/TATTT in B. tenella, and TTCT/ATCT in B. fistulosa. Codon usage bias analysis was performed based on 79 shared protein-coding genes (PCGs) from 25 Blumea plastomes. Total length of the protein-coding sequences ranged from 67,977 to 68,031 bp, encoding 22,575–22,602 codons (Supplementary Table 9). Leucine (Leu) had the most codons (2,397–2,411), while cysteine (Cys) had the fewest (244–250). Methionine (Met) and tryptophan (Trp) were exclusively encoded by AUG and UGG, respectively. A total of 64 types of codons were identified across the 25 Blumea plastomes, with RSCU values ranging from 0.34 to 1.94 (Figure 6; Supplementary Table 9). Thirty codons (UUA, AGA, UCU, GCU, UAA, ACU, UAU, GAU, AAU, CCU, GGA, GUA, CAA, GAA, CAU, AAA, UGU, AUU, GUU, CGU, UUU, CGA, GGU, CUU, ACA, AGU, UUG, UCA, CCA, GCA) had RSCU > 1. Except for the codon UUG, all codons with RSCU > 1 preferred codons were A/U-ending. Furthermore, AUG and UGG exhibited an RSCU of 1, indicating no codon usage bias for these two codons. The RSCU values of codon in Blumea plastome. Color circle, the orange represent higher RSCU values while the blue indicate lower values. Phylogenetic analyses As shown in Supplementary Table 10, the plastome alignment matrix was 157,779 bp in length, containing 9,788 variable sites (5.97%) and 7,908 parsimony-informative sites (4.83%). The nrDNA alignment matrix was 4,171 bp in length, with 488 variable sites (8.31%) and 337 parsimony-informative sites (5.74%). Both ML and BI analyses based on the complete plastome and nrDNA yielded largely consistent phylogeny topologies, respectively, with the monophyly of Blumea strongly supported (BS = 100, PP = 1.00). In the plastome phylogenetic tree (Figures 7A), Blumea was divided into four clades: (1) the B. stricta clade, consisting of B. stricta (BS = 100, PP = 1.00); (2) the B. densiflora clade, sister to the B. stricta clade, comprising B. aromatica, B. calcicola, B. densiflora var. densiflora, B. densiflora var. hookeri, B. formosana, B. henryi, B. lanceolaria, B. martiniana, and B. sagittata (BS = 100, PP = 1.00). (3) the B. balsamifera clade containing B. balsamifera (BS = 100, PP = 1.00); (4) the B. lacera clade including B. axillaris, B. clarkei, B. eberhardtii, B. fistulosa, B. hieraciifolia, B. megacephala, B. napifolia, B. oblongifolia, B. oxyodonta, B. riparia, B. sericans, B. sessiliflora, B. sinuata, and B. tenella (BS = 100, PP = 1.00). Phylogeny inferred from maximum likelihood (ML) and Bayesian (BI) analyses based on the complete plastomes (A) and nrDNA (B) of Blumea. The ML bootstrap values (BS) and the posterior probabilities values (PP) equal to 100/1 is omitted from each node and is not shown. “-/-” indicates that the value is missing. Species with highlighted in red represent the non-monophyletic species, while red circle on the node represent the monophyletic species. Samples with identical voucher numbers connected by curves. Within the B. densiflora clade, B. lanceolaria was sister to the B. calcicola + B. sagittata clade, and this combined clade then clustered with B. aromatica. The remaining species formed another clade, in which B. formosana (cp127), B. formosana (cp128), B. martiniana, and B. densiflora var. densiflora, B. formosana (PX404818) were successively resolved as sister lineages to the B. henryi + B. densiflora var. hookeri clade. In the B. lacera clade, B. oxyodonta was nested within the B. axillaris and clustered with B. hieraciifoliaat the base of the B. lacera clade; B. tenella and B. sessiliflora were successively sister to B. sinuata, forming the next diverging clade, followed by the B. napifolia + B. sericans clade, the B. fistulosa + B. clarkei clade, B. eberhardtii, and B. oblongifolia; Finally, B. riparia was nested in B. megacephala, forming the terminal clade. In the nrDNA-based phylogenetic tree, Blumea was likewise divided into the same four clades with the same taxa, but their phylogenetic positions were obviously different (Figure 7B). Specifically, the B. stricta clade was sister to the B. balsamifera clade (BS = 100, PP = 1.00), together constituting an early-diverging lineage. In contrast, the B. densiflora clade clustered with the B. lacera clade (BS = 100, PP = 1.00). Notably, substantial topological incongruence was observed between the cpDNA- and nrDNA-based trees within both the B. densiflora and the B. lacera clades. In the B. densiflora clade, B. aromatica formed the earliest-diverging lineage, followed by B. lanceolaria, B. formosana, and the B. henryi + B. martiniana clade; B. calcicola and B. densiflora var. hookeri were successively sister to the B. sagittata + B. densiflora var. densiflora clade. Within the B. lacera clade, B. sessiliflora was sister to B. sinuata at the base of the clade, followed by B. fistulosa; Among the remaining taxa, B. napifolia and B. clarkei were successively resolved as sister to a clade in which B. eberhardtii and B. riparia were nested in B. megacephala; Unlike in plastome-based phylogeny, B. sericans and B. oblongifolia were nested in B. axillaris, and together formed a clade sister B. hieraciifolia. Discussion Conservation in Blumea plastome All Blumea plastomes exhibited a typical circular quadripartite structure, and were highly conserved in genome size, GC content and gene composition. Comparative junction analysis and Mauve analysis further confirmed the high structural conservation of Blumea plastomes, with no obvious contraction or expansion of IR regions, nor gene rearrangements detected. mVISTA analysis indicated high sequence similarity among genomes. In addition, the plastomes of this genus showed conservation in codon usage bias, characterized by identical preferred codon types, RSCU values, and the overall comparable distribution patterns and nucleotide repeat compositions of SSRs. These conserved features have also been reported in other plastomes of Asteraceae species (Thode et al., 2021; Duan et al., 2022; Yun and Kim, 2022; Xue et al., 2025). Variations across Blumea plastome Nevertheless, the plastomes of Blumea also displayed distinct, even unique, patterns of variation. Although the genes flanking each junction were identical, the junction regions across the genus could be divided into four types. Unexpectedly, these four types corresponded exactly to the four clades: Type I corresponded to the B. lacera clade; Type II corresponded to the B. densiflora clade; and Type III and Type IV to the independent B. balsamifera clade and B. purpurea clade, respectively, consistent with their distinct lineages in the phylogenetic tree. These results indicate that junction specificity provides valuable information for clade delimitation and offers useful supplementary evidence for phylogenetic inference. Similar patterns have been reported in other plant lineages (Thode and Lohmann, 2019; Liu et al., 2022; Xu et al., 2025). Nucleotide diversity (Pi) analysis revealed the degree of variation in the nucleotide sequence. In this study, nine mutation hotspot regions (trnH-psbA, ycf1, trnC-petN, trnY-rpoB, ycf2-trnL, ndhF-rpl32, rpl32-trnL, ccsA-ndhD, and ndhG-ndhI) were identified, which can be employed as candidate DNA barcodes for species identification in Blumea. Among them, trnH-psbA represents a core universal DNA barcode widely used in previous studies (CBOL Plant Working Group, 2009; Hollingsworth et al., 2009; Thakur et al., 2019); the ycf1 gene usually with the highest number of variable and parsimony-informative sites, was considered one of the most promising barcodes for species discrimination (Dong et al., 2015; Gogoi et al., 2020; Li et al., 2021); the trnY-rpoB region was less frequently mentioned, but the trnY-trnE and the trnE-rpoB regions within it had been recognized as mutation hotspots (Shen et al., 2020; Zheng et al., 2024), as had the other regions identified here (Jiao et al., 2018; Gao et al., 2020; Xie et al., 2021; Wei et al., 2022; Yao et al., 2022; Wang et al., 2024). As one of the most widely used molecular markers, SSRs play important roles in species identification, population genetics, biodiversity research, and marker-assisted selection in molecular breeding (He et al., 2020; Li et al., 2022; Pan et al., 2024; Xu et al., 2024). In this study, extensive interspecific polymorphism of plastomic SSRs was detected among Blumea species. Firstly, substantial variation existed in both the total number of SSRs across plastomes and the number of SSR types, particularly mononucleotide repeats (Supplementary Table 8). Secondly, pentanucleotides repeats were scattered in some species of the B. densiflora clade and B. lacera clade, and hexanucleotide repeats were uniquely present in B. henryi. In addition, the SSR motifs were diverse, even specific across Blumea plastomes (Supplementary Table 8). Some unique SSR motifs were present in the plastomes of certain species—for instance, the TTTA AA motif in B. henryi, ATAAA/TATTT in B. tenella, and TTCT/ATCT in B. fistulosa—and may serve as species-specific molecular markers. Intriguingly, analogous to the junction types between the LSC, SSC, and IR regions, the distribution patterns of certain SSR motifs across Blumea plastomes appear to align with phylogenetic relationships. For instance, the G motif is exclusively detected in the B. densiflora clade, while the TG motif was restricted to the B. lacera clade. Similarly, the ATTT and TATAT motifs also showed clade-specific distributions. Phylogenetic implications Compared with the previous fragment-based phylogeny of Blumea, the phylogenetic trees constructed from complete plastomes and nrDNA sequences in this study exhibited high node support and topological resolution. Our results also corroborated the taxonomic treatment of the B. lacera, B. balsamifera, and B. densiflora clades within the genus (Pornpongrungrueng et al., 2007; Pornpongrungrueng et al., 2009; Chung et al., 2022). Furthermore, both plastome-based and nrDNA-based trees supported the inclusion of Cyathocline Cass. within Blumea (Li et al., 2014), thus supporting the recognition of the B. stricta clade. Among these four clades, life form patterns were consistent with phylogenetic relationships. The B. lacera clade consisted of herbaceous or vine-like taxa; the B. densiflora clade comprised subshrubby species; and the B. balsamifera and B. stricta clades represented two independent evolutionary lineages corresponding to subshrubs/shrubs (B. balsamifera) and herbs (B. stricta), respectively. Previously, the monophyly of taxa including B. formosana, B. sinuata, B. megacephala, B. aromatica, B. lacera, B. axillaris, B. saxatilis, and B. hieraciifolia was rarely well supported (Pornpongrungrueng et al., 2007; Pornpongrungrueng et al., 2009; Chung et al., 2022). In contrast, our plastome and nrDNA dataset greatly improved the resolution of this phylogenetic uncertainty. As expected, the monophyly of B. sinuata, B. aromatica, and B. hieraciifolia was strongly supported (BS = 100, PP = 1.00) in both trees, although B. axillaris, and B. megacephala still remained non-monophyletic. Notably, the B. formosana, which was non-monophyletic in the plastome tree, was resolved as monophyletic in the nrDNA tree, suggesting that nrDNA sequences may have greater power for resolving phylogenetic relationships in Blumea. Nonetheless, the phenomenon of non-monophyly implied the limitations of using complete plastome or nrDNA in the genus. For recently diverged groups with complex evolutionary histories, such as Panax (Ji et al., 2019) and Fargesia (Lv et al., 2023), plastomes and nrDNA failed to clearly distinguish closely related species or robustly resolve phylogenetic relationships. Indeed, Blumea originated approximately 49.00–18.43 Ma, representing a relatively young genus that underwent rapid radiation during the Miocene (Zhang et al., 2019). Moreover, Chung et al. (2022) proposed that hybridization, introgression, and genetic exchange between widespread and endemic Blumea species may underlie paraphyly (such as B. formosana in plastome-based tree) in the genus. The cytonuclear discordance observed in our study (Figure 7) further supports the possible occurrence of hybridization, introgression, and incomplete lineage sorting (ILS) during the evolutionary history of Blumea. Conclusion This study presents the most comprehensive comparative analysis of plastomes from Blumea to date, systematically revealing the evolutionary characteristics of the plastomes within this genus, including their conservation and divergence. Phylogenetic reconstructions based on plastome and nrDNA datasets established a robust phylogenetic framework for Blumea, encompassing the B. lacera clade, B. balsamifera clade, B. densiflora clade, and B. stricta clade, and preliminarily clarified the phylogenetic relationships among species. Furthermore, the monophyly of taxa including B. formosana, B. sinuata, B. megacephala, B. aromatica, B. axillaris, and B. hieraciifolia was well examined. In summary, this study provides valuable genetic resources for Blumea and yields novel insights into its phylogeny, laying a solid foundation for subsequent taxonomic, systematic, and species identification research. To better elucidate the evolutionary history of Blumea, future studies with more sampling and integration of more genomic data, such as mitochondrial genome (mtDNA) or nuclear single-copy gene data, will be necessary.