Integrating plastome comparison and phylogenetic analyses reveals the phylogenetic relationships of Pachypleurum (Apiaceae) Authors: Bo-Ni Song, Ya-Ni Song, Feng Yong, Jiao-Jiao Deng, Xin-Yu Hong, Chang-Kun Liu Published: 2026 ABSTRACT BackgroundPhylogenomic data rich in informative loci can significantly improve the support and resolution of phylogenetic tree and has been proven to be an effective tool for studying plant taxonomy and phylogeny. The genus Pachypleurum Ledeb., belonging to subfamily Apioideae (Apiaceae), comprises approximately seven species worldwide (six in China). Due to the long-standing taxonomic ambiguity of this genus, it is urgent need for investigation. ResultsIn this current study, we obtained six Pachypleurum plastomes and performed comprehensively comparative analyses. The results showed that six Pachypleurum plastomes were conserved in genome structure, gene number, gene arrangement, codon bias and SSRs, but varied in genome size, GC content and SC/IR boundaries. These findings indicated that Pachypleurum plastomes showed both conservation and diversity. In addition, nine mutation hotspot regions (matK, rps16, atpF, clpP, ndhE, trnH-psbA, psbK-psbI, accD-psaI, ycf4-cemA) were identified that can serve as potential DNA barcodes for species identification within the genus Pachypleurum. Phylogenetic analyses based on 79 commonly shared PCGs and ITS sequences robustly supported that the non-monophyly of Pachypleurum and six Pachypleurum members fell into two clades: Selineae and Acronema Clade. The phylogenetic results also showed that Chinese Pachypleurum taxa may be circumscribed to only one species (P. alpinum) and other five species previously placed in this genus may be transferred out this genus and the systematic position of them were also discussed. Finally, combined with plastome comparison, phylogenetic analyses and morphological characteristics, we discussed the morphological delimitation of Pachypleurum genus.ConclusionsThis study characterized six Pachypleurum plastomes and highlighted the power of plastome data to significantly improve the phylogenetic supports and resolutions. In conclusion, our study not only enriched the data on the plastomes of Pachypleurum, but also provided a framework for Pachypleurum and improved the taxonomic system of the Apiaceae family.Supplementary InformationThe online version contains supplementary material available at 10.1186/s12870-026-09263-6. Introduction The genus Pachypleurum Ledeb., belonging to the subfamily Apioideae (Apiaceae), is widely distributed in China, Nepal, and Russia, spanning from Central Asia to Europe. The plants of the genus grow in alpine meadows, mountain slopes, and gully shrublands at elevations ranging from 2,000 to 4,600 m [1, 2]. The main morphological features of its members are characterized by calyx teeth prominent, petals white or purple and apex notched with small incurved lobule, fruit dorsally compressed, ribs all winged, subequal; vittae 1 (–2) in each furrow, 2 (–4) or absent on commissure [1, 2]. The genus was established by Leded. in 1829 based on only one species, Pachypleurum alpinum Ledeb. Subsequently, new species and combinations were continuously published. According to the International Plant Names Index (https://www.ipni.org) and Plants of the World Online (https://powo.science.kew.org) record, a total of eighteen species were placed in this genus, but these treatments have not been widely recognized. Zhang suggested that the genus contained seven species, with five species distributed in China [1]. Pu and Watson recorded that this genus comprised six species and placed Pachypleurum mucronatum (Schrenk) Schischk. in Ligusticum L. genus [2]. Pimenov further narrowed the scope of the genus and proposed that the genus only included three species (Pachypleurum alpinum Ledeb., Pachypleurum mutellinoides (Crantz) Holub, Pachypleurum altaicum Revuschkin), with only one species (P. alpinum) distributed in China [3, 4]. Pachypleurum muliense R. H. Shan & F. T. Pu was transferred to Ostericum Hoffm. genus [5]; Pachypleurum nyalamense Ho T. Chang & R. H. Shan and Pachypleurum lhasanum Ho T. Chang & R. H. Shan were transferred to Cortia DC.genus [4]; a new genus Kailashia Pimenov & Kljuykova was established to accommodate Pachypleurum xizangense H. T. Chang & Shan [5]. Therefore, it can be seen that although Pachypleurum genus includes several species and is a small genus in Apiaceae family, the genus boundary is very vague and species frequently transfers between this genus and its closely related genera. In addition, species delimitation of this genus was also blurred, which was also mainly caused by the high variability of morphological characteristics. For example, Kljuykov et al. treated P. altaicum as a synonym of P. alpinum based on morphological characteristics [6]. Zhou et al. carried out the phylogenetic studies of genus Ligusticum L. and its allied genera, yet the Latin names of relevant species still adopted the genus name Pachypleurum [7]. From above, traditional taxonomic studies on this genus mainly focused on morphological features, whereas many species of the genus have always shown extensive morphological variation that caused high difficulties in species classification and delimitation. Hence, other evidence, such as molecular data, should be used to re-evaluate the taxonomic system of the genus. A high support and resolution phylogenetic framework can provide useful information for the taxonomy and phylogeny of the genus Pachypleurum, but no specific molecular systematic study focusing on this genus. And previous studies on the molecular phylogenetic reconstruction of the subfamily Apioideae and its main clade only involved limited Pachypleurum members [8–10]. For example, Valiejo-Roman et al. performed phylogenetic analysis of Ligusticum L. genus, with only one Pachypleurum species (Pachypleurum mutellinoides (Crantz) Holub) involved, and the results showed that Pachypleurum genus was closely related to both Phlojodicarpus Turcz. ex Ledeb. and Seseli L. genus [8]. Downie et al. performed phylogenetic study on the main clade of the subfamily Apioideae based on cpDNA fragments and ITS sequences, which indicated that Pachypleurum genus was non-monophyly, with Pachypleurum species scattered in two clades: Selineae and Acronema Clade [9]. Therefore, additional molecular resources are urgently needed to reconstruct strongly robust phylogeny of genus Pachypleurum and re-evaluate the taxonomic system and interspecific relationships of the genus. The plastid genome (plastome) is highly conserved in gene structure, sequence, content, order and has numerous informative characters, thus it is widely used for phylogenetic analyses, especially for those taxonomically controversial taxa within the family Apiaceae in recent years [11–21]. Although five plastomes of Pachypleurum species were previously submitted in GenBank by us, the plastid phylogenomic analysis of the genus has not been conducted. Therefore, we performed comprehensive plastome analyses for this taxonomically complex genus in this study. Our major aims were to: (1) characterize the plastomes of Pachypleurum plants; (2) select mutation hotspot regions from Pachypleurum plastome as potential DNA barcodes for Pachypleurum species authentication; (3) evaluate the potential of plastome for resolving the phylogeny of genus Pachypleurum and re-evaluate the classification system and interspecific relationships. Our study will well lay the foundation for the phylogenetic and taxonomic studies of Pachypleurum genus. Materials and methods Data collection We searched and downloaded six Pachypleurum plastomes, including P. alpinum (OQ685947), P. lhasanum (OQ685948), P. mucronatum (OP672449), P. muliense (OM281949), P. nyalamense (OQ685950) and P. xizangense (OQ685951) from NCBI databases (https://www.ncbi.nlm.nih.gov). Among them, except for P. mucronatum sequence, other five sequences were previously released by us. We re-annotated each plastome using CPGAVAS2 (http://www.herbalgenomics.org/cpgavas2) [22]. Then, each annotated plastome was further examined and adjusted manually in Geneious v11 (Biomatters, Ltd, Auckland, New Zealand) [23] by comparing with the plastome of congeneric species. Plastome comparison analyses First, to observe IR contraction and expansion of Pachypleurum plastomes, the boundaries between inverted repeat and single-copy (IR/SC) were detected in Geneious R11 [23]. Then, DNA rearrangements among six Pachypleurum plastomes were detected by using Mauve Alignment [24] implemented in Geneious R11 [23], setting other parameters as default values. Finally, mVISTA tool [25] was used to investigate the degree of variation sequences of these six Pachypleurum plastomes, setting LAGAN model and P. alpinum as the reference. Simple sequence repeats (SSRs) and codon bias analyses We used the MISA perl script [26] to investigate the simple sequence repeats (SSRs) in each Pachypleurum plastomes, with the thresholds of 10, 5, 4, 3, 3, and 3 repeat units were set for mono-, di-, tri-, tetra-, penta-, and hexa-nucleotides according to previous study [12]. In addition, we extracted protein coding genes (PCGs) from six Pachypleurum plastomes for codon analysis. All overlapping genes were removed and finally gained 79 PCGs for each species. Relative synonymous codon usage (RSCU) is the value of the observed and expected frequency of a codon encoding a specific amino acid [12, 27]. RSCU value greater than 1.00 indicates positive codon usage bias, less than 1.00 indicates negative codon usage bias, and equal to 1.00 indicates no bias. Then, CodonW v1.4.2 program [28] was used to calculate the RSCU value and the RSCU heatmap of six Pachypleurum plastomes was visualized in TBtools [29]. Identification of divergence hotspots To identify mutation hotspot regions, the protein coding regions, non-coding regions and intron regions of these six Pachypleurum plastomes were respectively extracted in Geneious R11 [23]. Then, we aligned the sequences using MAFFT v7.221 [30] and manually corrected with Geneious R11 [23]. Due to the variation caused by shorter sequences, so we discarded alignments less than 200 bp [12]. Therefore, the alignments with longer than 200 bp in length were filtered to calculate nucleotide diversity (Pi) using DnaSP v5.0 [31]. Phylogenetic analysis Due to the maternal inheritance of plastome, it cannot fully reveal the evolutionary history of the taxon, while ITS sequences are biparental inherited inheritance, representing another different genetic pattern [32]. Therefore, to obtain a comprehensive understanding of phylogeny of Pachypleurum, we used these two datasets to construct the phylogenetic trees. Dataset Ⅰ included 79 commonly shared PCGs of forty-eight complete plastomes and dataset Ⅱ included forty-eight ITS sequences, and these forty-eight species represented the main clade of subfamily Apioideae involving Pachypleurum species. Among them, two Chamaesium species (Chamaesium viridiflorum (Franch.) H. Wolff ex R. H. Shan and Chamaesium mallaeanum Farille & S. B. Malla) were selected as the outgroup according to the result of Wen et al. [11] (Table S1). For plastome data, we extracted 79 commonly shared PCGs of forty-eight complete plastomes, aligned with MAFFT v7.221 [30] and then concatenated the alignments as a super matrix using PhyloSuite v1.2.2 [33]. For ITS sequences, we straightway aligned them with MAFFT v7.221 [30] to obtain the matrix. Two methods [maximum likelihood analysis (ML) and Bayesian inference (BI)] were used to construct the phylogenetic trees. In detail, ML analysis was performed in RAxML v8.2.8 [34] with GTRGAMMA model, and 1000 replicates was used to estimate the bootstrap support (BS) value for each node referred to the RAxML manual. BI analysis was performed in MrBayes v3.2.7 [35] under the Akaike Information Criteria (AIC) [36], and the best-fit nucleotide substitution model (GTR + I+G) for matrix of PCGs dataset and ITS dataset were detected by Modeltest v3.7 [37]. Markov chain Monte Carlo (MCMC) algorithm was run for 5,000,000 generations (sampling every 100 generations) with two runs and four chains (one cold chain and three heated chains). Until the average standard deviation of split frequencies was below 0.01, the running finished and discarded the initial 25% of sampled trees. Finally, the remaining trees were used to generate the consensus trees and calculated posterior probabilities (PP). The software FigTree v1.4.2 [38] was used to visualize the results of ML and BI phylogenetic analyses. Results Plastome features among Pachypleurum plants In this study, we comprehensively compared the whole plastomes of six Pachypleurum taxa. The results showed that significant variation was found in overall size of six plastomes, with P. muliense had the largest length (156,054 bp) and other five plastomes having the length varied from 147,655 bp (P. alpinum) to 148,258 (P. lhasanum) (Table 1). All Pachypleurum plastomes displayed a typical quadripartite structure, including one large single copy region (LSC: 85,484 bp–93,433 bp), one small single copy region (SSC: 17,502 bp–17,684 bp) divided by a pair of inverted repeat regions (IRs: 18,307–26,443 bp) (Fig. 1; Table 1). The gene composition was highly consistent among six plastomes, which possessed 114 unique genes, comprising 80 protein coding genes (PCGs), 30 tRNA genes and 4 rRNA genes. And the total GC content of six Pachypleurum plastomes ranged from 37.4% to 37.6% (Table 1). Table 1Comparison of plastome features among Pachypleurum plantsTaxonTotal length(bp)LSC(bp)SSC(bp)IR(bp)Total GCcontent(%)Total genes(unique)Protein codinggenes(unique)rRNA genes (unique)tRNA genes (unique) P. alpinum 147,65593,43317,60818,30737.611379430 P. lhasanum 148,25892,30017,50219,22837.411379430 P. mucronatum 147,76893,09917,56118,55437.611379430 P. muliense 156,05485,48417,68426,44337.411379430 P. nyalamense 147,88592,29717,54419,02237.511379430 P. xizangense 148,17992,33517,62819,10837.511379430 Table 1Comparison of plastome features among Pachypleurum plantsTaxonTotal length(bp)LSC(bp)SSC(bp)IR(bp)Total GCcontent(%)Total genes(unique)Protein codinggenes(unique)rRNA genes (unique)tRNA genes (unique) P. alpinum 147,65593,43317,60818,30737.611379430 P. lhasanum 148,25892,30017,50219,22837.411379430 P. mucronatum 147,76893,09917,56118,55437.611379430 P. muliense 156,05485,48417,68426,44337.411379430 P. nyalamense 147,88592,29717,54419,02237.511379430 P. xizangense 148,17992,33517,62819,10837.511379430 Comparison of plastome features among Pachypleurum plants Fig. 1Plastome maps of six Pachypleurum species. Genes shown outside of the outer layer circle are transcribed clockwise, while those insides are transcribed counterclockwise. The genes belonging to different functional groups are color-coded. The dark gray area of the inner circle denotes the GC content of plastome Plastome maps of six Pachypleurum species. Genes shown outside of the outer layer circle are transcribed clockwise, while those insides are transcribed counterclockwise. The genes belonging to different functional groups are color-coded. The dark gray area of the inner circle denotes the GC content of plastome IR/SC borders analysis The borders of IRa/SSC and IRb/SSC among the six Pachypleurum plastomes were slightly conserved: the IRa/SSC junctions were located between ycf1 gene and ndhF gene, and the ndhF gene had 10–71 bp away from the IRa/SSC borders. The ycf1 gene crossed the SSC/IRb borders with 1,921–2,052 bp in IRb region. However, the junctions of IRa/LSC and IRb/LSC in Pachypleurum plastomes were divergent: the ycf2 gene crossed the IRa/LSC borders in five Pachypleurum plastomes (P. alpinum, P. lhasanum, P. mucronatum, P. nyalamense and P. xizangense), but in P. muliense plastome, the ycf2 gene located in IRa region and the rps19 gene located at IRa/LSC junction with 81 bp in IRa region. The IRb/LSC borders located between ycf2 gene and trnH gene in five Pachypleurum plastomes (P. alpinum, P. lhasanum, P. mucronatum, P. nyalamense and P. xizangense) and trnH gene had 3–916 bp away from the IRb/LSC borders, and in P. muliense plastome, rps19 gene and trnH gene located in the IRb/LSC borders (Fig. 2). Fig. 2Comparison of SC/IR borders among the six Pachypleurum plastomes. Different colors represent different regions or different genes Comparison of SC/IR borders among the six Pachypleurum plastomes. Different colors represent different regions or different genes The characteristics of simple sequence repeats We investigated the simple sequence repeats (SSRs) and identified 452 SSRs in six Pachypleurum plastomes (Table S2). And the total number of SSRs varied from 64 (P. muliense) to 84 (P. xizangense) among these six Pachypleurum plastomes. Of these six type repeats, the most abundant type was mono-repeats (248, 54.87%), followed by di-repeats (110, 24.34%), tetra-repeats (58,12.83), tri-repeats (18, 3.98%), penta-repeatswere (13, 2.88%) and hexa-repeats (5, 1.11%) (Fig. 3, Table S2). The mono-repeats were present in all Pachypleurum plastomes, and penta-repeats were missing in P. mucronatum plastome, hexa-repeats were missing in P. muliense plastome and the remaining five plastomes had hexa-repeats: AAATAG/ATTTCT was present in P. alpinum, AATATT/AATATT was present in P. mucronatum, AGATAT/ATATCT was present in P. lhasanum, P. nyalamense and P. xizangense. In addition, bases A and T were the dominant elements for all identified SSRs in the six plastomes (Fig. 4). Fig. 3Analyses of simple sequence repeats (SSRs) in six Pachypleurum plastomes Analyses of simple sequence repeats (SSRs) in six Pachypleurum plastomes Fig. 4Numbers of SSR motif identified in six Pachypleurum plastomes Numbers of SSR motif identified in six Pachypleurum plastomes Codon usage bias and relative synonymous codon usage (RSCU) analysis To investigate the codon usage of Pachypleurum plastomes, 79 PCGs of six Pachypleurum plastomes were extracted and connected for each plastome. The results showed that these sequences encoded 22,590–22,781 codons and the codon usage bias analysis also revealed notable differences detected in the encoding frequency of various amino acids. Specifically, Leucine (Leu) was encoded by the highest number of codons (2,378–2,421), whereas Cysteine (Cys) was the least (232–240) (Table S3). Additionally, the RSCU values were highly consistent and it varied from 0.33 to 2.01 among these six Pachypleurum plastomes (Table S3). 30 codons (UUU, UUA, UUG, CUU, AUU, GUU, GUA, UCU, UCA, CCU, CCA, ACU, ACA, GCU, GCA, UAU, UAA, CAU, CAA, AAU, AAA, GAU, GAA, UGU, CGU, CGA, AGU, AGA, GGU, GGA) had the RSCU value greater than 1.00, two codons (AUG, UGG) had RSCU value were equal to 1.00 and the remaining 32 codons had RSCU value less than 1.00 in six plastomes. All 30 codons with RSCU > 1.00 were ended with A/U, except UUG (Fig. 5). Fig. 5The RSCU values of 79 protein coding regions for six Pachypleurum plastomes. The red represents higher RSCU values, while the blue indicates lower RSCU values The RSCU values of 79 protein coding regions for six Pachypleurum plastomes. The red represents higher RSCU values, while the blue indicates lower RSCU values Sequence diversity analysis and hotspots identification The Mauve result illustrated that gene arrangement of six Pachypleurum plastomes were highly conserved (Fig. 6). In addition, the result performed by the mVISTA program illustrated that six plastomes had differences in LSC and SSC regions, while the IR regions were relatively conserved. Most of the high variation regions were detected in the non-coding regions (Fig. 7). Based on the sequence divergences, we calculated Pi value of four regions among the six Pachypleurum plastomes and selected nine mutation hotspot regions as potential DNA barcodes, including five protein coding genes [matK (0.01958), rps16 (0.01614), atpF (0.01576), clpP (0.01581), ndhE (0.01579)] and four non-coding regions (trnH-psbA (0.09310), psbK-psbI (0.04591), accD-psaI (0.05187), ycf4-cemA (0.05117)]. These findings also illustrated that the Pi value of non-coding region was remarkably higher than that of the coding region (Fig. 8). Fig. 6Mauve alignment of six Pachypleurum plastomes. Local collinear blocks within each alignment are represented by blocks of the same color connected with lines Mauve alignment of six Pachypleurum plastomes. Local collinear blocks within each alignment are represented by blocks of the same color connected with lines Fig. 7mVISTA alignment for six Pachypleurum plastomes using P. alpinum as the reference mVISTA alignment for six Pachypleurum plastomes using P. alpinum as the reference Fig. 8The nucleotide diversity (Pi) values among six Pachypleurum plastomes. (A) protein coding regions; (B) non-coding and intron regions The nucleotide diversity (Pi) values among six Pachypleurum plastomes. (A) protein coding regions; (B) non-coding and intron regions Phylogenetic analyses Phylogenetic analyses performed by PCGs data and ITS sequences showed that several incongruences existed in topologies between PCGs-based (Fig. 9A) and ITS-based (Fig. 9B) phylogenetic trees. However, both topologies robustly supported six Pachypleurum members were not clustered as a monophyletic group but fell into two clades: Selineae and Acronema Clade. Among them, P. muliense located in Acronema Clade and other five Pachypleurum species, including [P. alpinum (type species), P. lhasanum, P. mucronatum, P. nyalamense and P. xizangense)] were placed in Selineae (Fig. 9). Fig. 9Phylogenetic trees constructed by maximum likelihood (ML) and Bayesian inference (BI). The bootstrap values (BS) of ML and posterior probabilities (PP) of BI are listed at each node. # means the values < 50. Red words indicates Pachypleurum species. (A): PCGs tree; (B): ITS tree Phylogenetic trees constructed by maximum likelihood (ML) and Bayesian inference (BI). The bootstrap values (BS) of ML and posterior probabilities (PP) of BI are listed at each node. # means the values < 50. Red words indicates Pachypleurum species. (A): PCGs tree; (B): ITS tree In the PCGs-based trees, P. muliense was sister to Ostericum maximowiczii (F. Schmidt ex Maxim.) Kitag. (PP = 1.00, BS = 98); P. alpinum clustered with Libanotis condensata (L.) Crantz (PP = 1.00, BS = 99); P. mucronatum gathered with Saposhnikovia divaricata (Turcz.) Schischk (PP = 1.00, BS = 97); P. nyalamense made a sister with P. xizangense (PP = 1.00, BS = 100); P. lhasanum formed a separate branch (PP = 1.00, BS = 100) (Fig. 9A). In the ITS-based trees, the phylogenetic position of P. muliense was consistent with the PCGs-based phylogenetic trees and clustered with O. maximowiczii (PP = 0.50, BS = 78). P. alpinum was sister to Sinopeucedanum huangshanense (Lu Q.Huang, H.S. Peng & S.S. Chu) C.K. Liu & X.J. He, comb. nov. with robust supports (PP = 1.00, BS = 93); P. mucronatum clustered with Ligusticum thomsonii C. B. Clarke (PP = 1.00, BS = 82); P. xizangense and Cortiella hookeri (C. B. Clarke) C. Norman clustered together and then resolved as sister to P. lhasanum (PP = 0.95, BS = 66); P. nyalamense formed a separate branch (PP = 1.00, BS = 84) (Fig. 9B). Discussion Comparison of the plastomes in Pachypleurum Previous studies have justified the plastome is quite conservative in structure, order, content, and this phenomenon is no exception in the Pachypleurum. Pachypleurum plastomes displayed a typically quadripartite structure, which is common in other Apiaceae plants [39–43]. In addition, no gene rearrangement or loss were detected in these six Pachypleurum plastomes. And codon bias, SSRs profile also conserved, which were found in other genera of Apiaceae [12, 13, 15, 39, 40] that may be related to maintain the stability of plastome function [12–14, 17]. However, genome size, IR expansions and contractions have been confirmed to occur many times during plastome evolution [27]. This study found that large differences occurred in genome size (147,655 bp–156,054 bp) and IR boundaries, especially in P. muliense plastome. Previous studies have inferred that the variation of plastome size was mainly influenced by the contraction and expansion of IR regions [12]. Thus, our study also supported this conclusion, as the extension of the rps19 gene to the IR regions resulted in P. muliense plastome size was longer than other Pachypleurum plastomes. Potential DNA barcodes and molecular markers DNA barcoding is defined as the short DNA sequence with sufficient variation to identify species [44], such as matK, rbcL and trnH-psbA, which have been widely recognized as universal DNA barcodes [16, 17, 45, 46]. Although universal DNA barcodes have successfully distinguished many species, they have failed and encountered weak solutions for some notorious taxonomic groups [47]. As expected, we found that the variation in rbcL gene was relatively low among Pachypleurum plants. Hence, this region may have limited power to discriminate Pachypleurum species. Therefore, it is urgently needed to develop more effective variable loci as candidate DNA barcodes to distinguish Pachypleurum species. High nucleotide diversity (Pi) value means that certain regions have significant variation, which can be used to design the potential DNA barcodes [12]. This study screened out nine mutation hotspot regions: matK, rps16, atpF, clpP, ndhE, trnH-psbA, psbK-psbI, accD-psaI, ycf4-cemA, which were potentially useful for species identification in Pachypleurum genus. Among them, matK and trnH-psbA region were universal DNA barcodes, atpF and ycf4-cemA region were used as potential DNA barcodes for Sancilua L. species [12], psbK-psbI and accD-psaI were used as potential DNA barcodes for Cyclorhiza M. L. Sheh & R. H. Shan species [13], the remaining three high variable regions (rps16, clpP, ndhE) were specific to the genus Pachypleurum, and therefore they may be able to serve as specific potential DNA barcoding to distinguish Pachypleurum species. Later, we will test whether or not these sequences can serve as reliable DNA barcodes for species identification within Pachypleurum genus. SSRs are shorter tandem or microsatellite repeat sequences with repeat units of 1–6 bp, which widely distributed in different regions of the plastome [12, 48]. Due to a large amount of variation within the species, SSRs have been extensively used as valuable molecular markers for species authentication, population evolutionary analyses, phylogenetic analyses, plant taxonomy, geography of species, genetic diversity and population structure studies [12, 49]. In this study, we investigated the simple sequence repeats and found that AAATAG/ATTTCT was present in P. alpinum, AATATT/AATATT was present in P. mucronatum, AGATAT/ATATCT was present in P. lhasanum, P. nyalamense and P. xizangense. These unique molecular markers may contribute to future selection as molecular markers for distinguishing Pachypleurum species. Phylogenetic position of the Pachypleurum genus In this study, we performed phylogenetic analyses for Pachypleurum using PCGs data and ITS sequences and these two trees generated incongruence, which is a common phenomenon detected in other flower plants [50–53]. This incongruence was potentially attributed to different inherited background and mutation rates of ITS and plastid DNA [32, 54, 55]. The nuclear ITS is biparentally inherited and has a higher mutation rate, while the plastid DNA is maternally inherited and has a lower mutation rate [32, 53, 54]. In addition, most Pachypleurum species located in Selineae. According to Wen et al.’s study, the Selineae was a relatively young evolutionary clade and it originated in the late Miocene (9.67 Mya or 11.30 Mya) [11]. During this period, temperature continued to decline, aridification in Central Asia intensified, and the Qinghai Tibet Plateau expanded, leading to an increase in the number of Apiaceae species. However, during the rapid radiation differentiation process of plants, they had not enough historical and geological time to complete the process of lineage sorting. Therefore, the incongruence of Pachypleurum also may be caused by incomplete lineage sorting (ILS) due to the young evolutionary history. Previous molecular phylogenetic studies supported the non-monophyly of Pachypleurum based on a few molecular markers [7, 9, 10] and our study was no exception. Six Pachypleurum members scattered in Selineae and Acronema Clade. Of them, P. alpinum (type species) located in Selineae and other five Pachypleurum members were far away from P. alpinum, indicating that P. alpinum represented the truly systematic position of Pachypleurum genus and the genus located in Selineae, while other species may be transferred out of the Pachypleurum genus. Therefore, phylogenetic studies showed that Pachypleurum genus may be only included one species (P. alpinum) in China. Combined with Flora of China and previous studies [1–6], morphological characteristics also justified the phylogenetic results and we discussed the morphological boundary of the genus, that is stem well developed, single or 2–3, erect, striate, usually unbranched; bracts and bracteoles has margins white membranous, apex sometimes incised; fruit is sparsely puberulent, the abaxial side is slightly compressed; the main ribs develop into subequal thick wings, vittae are small and numerous, arranged in a nearly annular form, the ventral surface of the endosperm is flat. Systematic status of P. xizangense, P. muliense, P. lhasanum, P. nyalamense, P. mucronatum Pimenov transferred P. xizangense out Pachypleurum and established a new genus (Kailashia Pimenov & Kljuykov) to accommodate this species [4]. Our phylogenetic analyses also supported this taxonomic treatment that P. xizangense was far from P. alpinum. And morphological characteristics of P. xizangense and P. alpinum also supported phylogenetic analyses, such as stem branched vs. unbranched; Rays 20–40 vs. 10–15; Bracts 8–10, margins white membranous vs. 10–15, linear or coexist with pinnate. Therefore, combined with phylogenetic analyses and morphological characteristics, we revealed that the systematic position of P. xizangense that located in Selineae and was closely related to Ligusticopsis Leute and Cortiella C. Norman genus. Pimenov et al. proposed that P. muliense should transfer to Ostericum Hoffm. genus [5]. Our phylogenetic analyses also showed that P. muliense was sister to Ostericum maximowiczii in both PCGs-based tree and ITS-based tree within Acronema clade, which was consistent with the analysis of Jiang et al., who found Ostericum located in Acronema clade [56]. In addition, morphological characteristics also verified the reliability of the molecular data. Therefore, we revealed that the systematic position of P. muliense and supported the taxonomic treatment proposed by Pimenov et al. who classified P. muliense into Ostericum genus [5]. P. lhasanum formed a clade with Cortiella hookeri in ITS-based tree in the study, which was inconsistent with Pimenov et al.’s study that suggested to classify P. lhasanum into Cortia genus [4]. However, this taxonomic treatment may be inappropriate because the morphological features of P. lhasanum are significantly different from Cortia depressa but similar with Cortiella hookeri, such as the plants of P. lhasanum and Cortiella hookeri are acaulescent; leaves are 1–2-pinnately divided; the carpophores of the fruits are reduced; all the main ribs develop into broad wings; vittae 1 in each furrow; 2 on commissure [1, 2]. Thus, combined with phylogenetic analyse and morphological features, we suggested that transferring P. lhasanum to the Cortiella genus may be more appropriate. However, the systematic position of P. mucronatum and P. nyalamense need to completely explore in future studies. Because P. mucronatum had conflicts between PCGs-based tree and ITS-based tree. It gathered with Saposhnikovia divaricata in PCGs-based tree, whereas it clustered with Ligusticum thomsonii in ITS-based tree. We noticed that P. mucronatum has different morphological features from Saposhnikovia divaricata but is similar with Ligusticum thomsonii in overall morphology. Hence, combined with phylogenetic analyses and morphological features, we found P. mucronatum and L. thomsonii had close relationship, but the phylogenetic status of L. thomsonii had not been resolved to date, and thus, further research is still required to clarify the taxonomic placement of P. mucronatum. In addition, previous studies showed that the morphological features the of P. nyalamense and Cortia depressa are similar, and the type locality of P. nyalamense is Nyalam County, Xizang, China and the type locality of Cortia depressa is in Nepal, where is contiguous Nyalam County. Therefore, Deng’s study supported P. nyalamense should be transferred to Cortia genus and treated as a synonym of Cortia depressa [57]. However, due to the absence of Cortia depressa in this study, the systematic position of P. nyalamense need to completely investigate in future studies. In this study, although plastome data have significantly improved the support and resolution of the phylogenetic trees and provided a framework for the genus, due to the limited sampling involved, it cannot fully reflect the species boundaries in Apiaceae. Therefore, broader population-level sampling and additional nuclear data should be used in future studies. Conclusion In this study, we first investigated the plastome features of Pachypleurum and inferred the phylogeny of Pachypleurum using plastomes. Through a comprehensive comparative analysis, we observed that P. muliense displayed variable differences in genome size and IR/SC border compared with other five Pachypleurum species. The size of P. muliense plastome (156,054 bp) was much longer than that of them (147,655 bp–148,258 bp) and its IRa/LSC and IRb/LSC border also different from them, which rps19 gene expanded to IRa and IRb regions. Therefore, we supported the hypothesis that the genome size variation was caused by IR contraction and expansion. The plastome divergences further implied the non-monophyly of the Pachypleurum. In addition. we performed the phylogenetic analyses for Pachypleurum based on two datasets (PCGs data and ITS sequences) and found that Pachypleurum was not monophyletic group and six Pachypleurum members fell into two clades: Selineae and Acronema Clade. This study also suggested that Pachypleurum genus may only included one species (P. alpinum) in China, and the other five species (P. muliense, P. lhasanum, P. mucronatum, P. nyalamense, P. xizangense) transferred out from the genus may be more appropriate and we also discussed the systematic position of them. Furthermore, the conflicts between PCGs-based and ITS-based phylogenies could largely be attributed to different inherited background and mutation rates of ITS and plastid DNA. Moreover, incomplete lineage sorting (ILS) due to the young evolutionary history was also a main reason of conflicts because the Selineae was a relatively young evolutionary clade and it originated in the late Miocene. The phylogenetic analyses also highlighted the advantage of using plastome with rich information loci in resolving phylogenetic relationships. In conclusion, our study not only enriched the data on the plastomes of Pachypleurum, but also provided a framework for Pachypleurum and improved the taxonomic system of the Apiaceae family. Supplementary Information Supplementary Material 1. Supplementary Material 1. Supplementary Material 2. Supplementary Material 2. Supplementary Material 3. Supplementary Material 3.