The Persicaria species, including Persicaria chinensis, Persicaria hydropiper, Persicaria capitata, Persicaria glabra, Persicaria perfoliata, Persicaria tinctoria, Persicaria longiseta, and Persicaria maackiana, Persicaria hastatosagittata, Persicaria lapathifolia, and Persicaria pubescens were collected from various regions in Guangxi Province, China. Specifically, P. capitata (OR730796) and P. pubescens (OR730798) were sampled from Wuming District in Nanning, Guangxi; P. glabra (OR730797), P. tinctoria (OR730799), P. longiseta (OR730801), P. maackiana (OR730802), P. hastatosagittata (OR730803), P. perfoliata (OR730805) and P. lapathifolia (OR730804) were collected from Shanglin County in Nanning; P. chinensis (OR730800) was collected from Xixiangtang District in Nanning; P. hydropiper (OR570614) and (OR570615) were respectively collected from Xing’an County in Guilin, Guangxi and Sanjiang County in Liuzhou, Guangxi. The morphological identification of the collected specimens was performed by the Zhuang and Yao Medicinal Herb Laboratory (Guangxi University of Chinese Medicine) and verified by comparison with the Plant Photo Bank of China (PPBC). Specimens of these Persicaria species are deposited in the Herbarium of Traditional Chinese Medicine Plants, Guangxi University of Chinese Medicine (GXCM) (Supplementary Table S1). For DNA extraction, healthy plant leaves were selected, frozen in liquid nitrogen, and stored at ultra-low temperatures. Genomic DNA was subsequently extracted from the fresh leaf tissue using the hexadecyl trimethyl ammonium bromide (CTAB) method. DNA concentration and purity were assessed using a NanoDrop 2000 spectrophotometer (Thermo Fisher Scientific, USA). DNA samples that met the standards for library construction and fragment length were numbered and sequenced using the Illumina NovaSeq 6000 platform. Sequencing was conducted on an Illumina HiSeq platform by Genesky Biotechnologies Inc. (Shanghai, China). Following sequencing, the raw data were subjected to quality control and filtering using FastQC 0.11.8 version and R 3.6.1 version. The filtering criteria were as follows: (1) sequences containing more than three N bases were discarded; (2) sequences with less than 60% high-quality bases (Phred score ≥20) were removed; (3) low-quality bases at the 3’ end were excluded; (4) sequences shorter than 60 base pairs were discarded. The resulting high-quality sequences (clean reads) were saved in FASTQ format for subsequent assembly and annotation. The chloroplasts genome of Persicaria species plants were sequenced with an average depth of 21289 × and a coverage of 100.18%.

The high-quality sequencing reads obtained were assembled using the metaSPAdes algorithm (Bankevich et al., 2012) to construct the chloroplast genomes of 11 Persicaria species, with reference to the previously reported chloroplast genome sequences of Persicaria (NCBI accession numbers: NC_050358, NC_058319, NC_067040; Supplementary Table S1). The chloroplast genomes of the 11 Persicaria species were annotated using CpGAVAS. After chloroplast genomes were predicted and annotated, we manually edited by comparison with the published chloroplast genome sequences (NCBI accession numbers: NC_050358, NC_058319, and NC_067040). The annotated genome sequences have been uploaded to GenBank (https://www.ncbi.nlm.nih.gov/genbank/) with the following accession numbers: OR730796, OR730797, OR730798, OR730799, OR730800, OR730801, OR730802, OR730803, OR730804, OR730805, OR570614, OR570615 (Table 1). Finally, the chloroplast genome maps were generated using the OGDRAW 1.1.1 version. The Relative Synonymous Codon Usage (RSCU) analysis was performed using CodonW version 1.4.4 (http://codonw.sourceforge.net/).

Tandem Repeats Finder (Benson, 1999) was used to identify repetitive sequences. Simple Sequence Repeats (SSRs) in the assembled chloroplast genomes were identified using the MISA software. Global alignment analysis of the chloroplast genomes of the 11 Persicaria species was performed using the mVISTA software (https://genome.lbl.gov/vista/index.shtml). The inversion (IR) boundaries of these chloroplast genomes were mapped using the IRscope online tool (), and the contraction and expansion of the IR regions were analyzed based on the gene position differences at the boundaries. In order to identify the hypervariable regions within the chloroplast genome of Persicaria species, the chloroplast genomes of accessions OR730796 through OR730805 and OR570614 through OR570615 were aligned utilizing MEGA7 software. Subsequent manual adjustments to the alignment were performed using Se-Al version 2.04. The nucleotide diversity (Pi) across the chloroplast genome was then calculated employing DnaSP version 5 software (Librado and Rozas, 2009) through sliding window analysis.

To determine the evolutionary relationships within the Persicaria genus and clarify the phylogenetic relationships and positions of 11 medicinal species of Persicaria, we assembled the chloroplast genomes of the 11 species and included 53 additional chloroplast genome sequences from other Polygonaceae species available in GenBank. These sequences were used to construct a phylogenetic tree for phylogenetic analysis. Orthofinder 2.2.7, with multiple sequence alignment performed using MAFFT, was utilized for the analysis. Phylogenetic trees were constructed using the Maximum Likelihood (ML) method with 1000 bootstrap replicates (Zou et al., 2015), with RAxML-HPC2 7.6.3 (Hu et al., 2022) employed for tree construction.

To estimate the divergence times among lineages of Persicaria, we employed BEAST v1.10.4 (Drummond and Rambaut, 2007) utilizing a concatenated data matrix of complete chloroplast genomes to maximize genus inclusion. The analysis was conducted using the GTR + G substitution model, a relaxed molecular clock model, and the Yule process as the tree prior. In accordance with previous research (Manchester and O’leary, 2010; Yao et al., 2019), two fossil calibrations and two secondary calibrations were incorporated. The crown age of Polygonaceae was constrained to a range of 72.1-66.0 Ma, employing a lognormal calibration prior (Manchester and O’leary, 2010). The crown age of Muehlenbeckia was set within a range of 22.0-19.0 Ma, also using a lognormal calibration prior (Pole, 1992; Schuster et al., 2013). The crown age of Ruprechtia_albida was set within a range of 78.0-76.0 Ma, also using a lognormal calibration prior (Zhang et al., 2022). Both secondary calibration priors were modeled using a normal distribution. The Markov Chain Monte Carlo (MCMC) analysis was executed for 1 x 10⁹ generations, with sampling occurring every 10,000 generations. Convergence of the two independent runs and the stationarity of the chains were assessed using Tracer v1.7 (Rambaut et al., 2018), ensuring that all pertinent parameters exhibited an effective sample size (ESS) greater than 200. The initial 25% of the trees were discarded as burn-in. Subsequently, a maximum clade credibility tree was constructed, reporting mean node heights and 95% highest posterior density intervals (95% HPDs), utilizing TreeAnnotator v1.10.4 (Suchard et al., 2018). The resultant phylogenetic tree was visualized using FigTree v1.3.1.

After assembling the clean reads, the length of the chloroplast genomes of 11 Persicaria species ranged from 159,028-161,023 bp (Table 1; Figure 1). Like most angiosperms, the chloroplast genomes of these 11 Persicaria species exhibit a typical quadripartite structure, consisting of a large single-copy region (LSC), a small single-copy region (SSC), and two inverted repeat regions (IR) (Figure 1). The length of the LSC region for these 11 Persicaria species ranged from 83,572-85,544 bp, the SSC region from 12,894-14,740 bp, and the IR regions from 30,893-31,185 bp. The GC content in the LSC, SSC, and IR regions, as well as in the entire chloroplast genome, ranged from 35.9%-36.6%, 32.7%-33.4%, and 41.3%-41.5%, respectively, which is similar to previously reported GC content for chloroplast genomes in Persicaria species (Table 1).

Through gene annotation, we found that the chloroplast genomes of 11 Persicaria species from Guangxi, contained 4 rRNA genes and 30 tRNA genes. The primary differences in the chloroplast genomes were observed in the protein-coding genes. Specifically, P. capitata and P. chinensis had 74 protein-coding genes, P. tinctoria had 75, P. glabra, P. pubescens, P. longiseta, and P. hydropiper had 76, P. perfoliata had 77, and P. maackiana, P. hastatosagittata, and P. lapathifolia had 78 (Supplementary Table S2). With the exception of P. capitata and P. tinctoria, 18 genes were duplicated across the species, including 7 tRNA genes (trnI-CAU, trnL-CAA, trnV-GAC, trnI-GAU, trnA-UGC, trnR-ACG, trnN-GUU), 4 rRNA genes (rrn16, rrn23, rrn4.5, rrn5), and 7 protein-coding genes (ycf2, ycf1, ndhB, rpl2, rps12, rps19, rps7). Compared to other species, P. capitata and P. chinensis lacks the psbL, ndhD, rpoA, and rps15 genes, whereas P. tinctoria lacks the psbL, ndhD, and ycf1 genes. These missing genes were primarily located in the LSC and SSC regions (Supplementary Table S2, Figure 1). Furthermore, 18 genes contained introns, including 12 protein-coding genes (rpl16, rpl2, rps12, rps16, rpoC1, ndhB, ndhA, petB, petD, atpF, clpP, ycf3) and 6 tRNA genes (trnK-UUU, trnG-UCC, trnL-UAA, trnV-UAC, trnI-GAU, trnA-UGC). Genes containing introns were mainly distributed in the LSC and IR regions.

Codon usage bias analysis of the chloroplast genomes of 11 Persicaria species from Guangxi indicated that the overall size and base composition of the protein-coding regions were nearly identical across the genomes. There were minimal differences in codon usage preferences among the species. The most frequently used codons encoded the amino acid leucine (Leu), appearing 2796, 2853, 2850, 2476, 2799, 2805, 2947, 2929, 2940, 2859, 2920, and 2916 times, respectively. In contrast, the least frequently used codons encoded cysteine (Cys), appearing 297, 302, 298, 267, 297, 295, 305, 305, 305, 302, 299, and 299 times, respectively. Tryptophan (Trp) and methionine (Met) each had a single codon, while other amino acids were represented by 2–6 synonymous codons. Relative Synonymous Codon Usage (RSCU) values greater than 1 indicate a codon preference, values less than 1 indicate lower usage, and RSCU values equal to 1 suggest no preference (Zhong et al., 2023). Most codons ending in A or T exhibited RSCU values greater than 1 (Supplementary Table S3), indicating a preference for codons ending in A or T in the chloroplast genomes of these 11 Persicaria species.

Repeated sequences are believed to play an important role in genomic recombination and rearrangement. The repeat sequences in the chloroplast genome include palindromic repeats (P), forward repeats (F), tandem repeats (T), and reverse repeats (R). Our analysis of repeated sequences in the chloroplast genomes of 11 Persicaria species from Guangxi revealed that the total number of repeats ranged from 68 to 78. Among these, palindromic repeats ranged from 22 to 27, forward repeats ranged from 17 to 22, tandem repeats ranged from 17 to 28, and reverse repeats ranged from 4 to 8 (Figure 2). Furthermore, among the identified repeats, sequences with lengths between 10 and 30 bp were the most abundant. These results are consistent with previously reported findings in other Persicaria species, showing a high degree of similarity with the results observed in P. perfoliata (Guo et al., 2022; Zeng et al., 2025).

Simple Sequence Repeats (SSRs), or microsatellites, are DNA sequences consisting of short, tandemly repeated nucleotide motifs (1–6 bp in length) that hold significant importance for studies in plant populations (Asaf et al., 2017b). Moreover, chloroplast SSRs exhibiting intraspecific positional variability are commonly utilized as molecular markers for investigating population genetics and evolutionary processes (Yang et al., 2016; Zl et al., 2016). Among the 11 Persicaria species investigated in this study, the identified SSRs (excluding those of the ‘c’ type) primarily consisted of mono-, di-, tri-, and tetranucleotide repeat motifs. Mononucleotide repeats were the most abundant, with A/T base compositions being predominant. Dinucleotide repeats were the second most common type, among which AT/TA repeats constituted the majority (Figure 3). These findings are consistent with the characteristic pattern observed in angiosperm chloroplast genomes, where SSRs are predominantly composed of polyA and polyT repeats rather than tandem repeats of G/C (Du et al., 2022).

Analysis of the expansion and contraction at the inverted repeat (IR) boundaries across the 11 chloroplast genomes of Persicaria species from Guangxi revealed highly conserved flanking genes. Specifically, the junctions JLB (LSC-IRb), JSB (SSC-IRb), JSA (SSC-IRa), and JLA (LSC-IRa) were consistent among most species, indicating structural conservation within the family. At the JLB boundary, the flanking genes were rpl22 and rps19 in P. tinctoria, P. glabra, P. hydropiper, P. perfoliata, P. lapathifolia, P. hastatosagittata, and P. maackiana, with the distance from rpl22 to the JLB ranging from 19 to 25 bp. In contrast, the JLB boundary was located within the rpl2 gene in P. longiseta, P. chinensis, and P. pubescens. The expansion into LSC and IRb was 955 bp and 535 bp in P. longiseta, 970 bp and 520 bp in P. chinensis, and 979 bp and 511 bp in P. pubescens, respectively. In P. capitata, the JLB was located within rps19, with an expansion of 136 bp into LSC and 143 bp into IRb. The JSB boundary was located within the ndhF gene in all 11 species. At the JSA boundary, P. tinctoria, P. glabra, and P. capitata had only one flanking gene, either rps15 or ycf1, while the remaining species were flanked by both rps15 and ycf1. At the JLA boundary, the junction was located within rpl2 in P. longiseta, P. chinensis, and P. pubescens, with expansion into IRa and LSC of 535 bp and 954 bp, 520 bp and 969 bp, and 511 bp and 978 bp, respectively. In the other eight species, the JLA boundary was located within the trnH gene, with a variation in distance from trnH to the boundary ranging from 1 to 5 bp (Figure 4).

Comparative analysis of chloroplast genomes serves as a crucial component in genomics, enabling the assessment of conservation in gene order and the identification of structural variations among different species. Accordingly, we performed a global alignment of the 11 Persicaria chloroplast genomes using mVISTA (Figure 5). The results demonstrated that the LSC and SSC regions exhibited greater sequence divergence than the IR (a/b) regions, and non-coding regions were more variable than coding regions. Furthermore, highly divergent regions in these 11 Persicaria species were primarily distributed in intergenic spacers.

To identify highly variable regions, we conducted an analysis of variable sites within the 11 Persicaria chloroplast genomes using a sliding window approach implemented in the software DnaSP. The divergent loci, specifically trnQ-UUG-psbK, psbM-trnD-GUC, psbD, psbD-psbC, trnL-UAA, trnF-GAA-ndhK, atpB-atpE, clpP, ndhI-ndhA and ycf1, exhibited a nucleotide diversity (pi) value of 0.06 or higher (Figure 6). The trnQ-UUG-psbK, psbM-trnD-GUC, psbD, psbD-psbC, trnL-UAA, trnF-GAA-ndhK, atpB-atpE and clpP loci were located in the LSC region, whereas the ndhI-ndhA and ycf1 were situated in the SSC region; no divergent loci were identified in the IR region. These findings corroborate the observation that the LSC and SSC regions are less conserved than the IR regions. The ten genes trnQ-UUG-psbK, psbM-trnD-GUC, psbD, psbD-psbC, trnL-UAA, trnF-GAA-ndhK, atpB-atpE, clpP, ndhI-ndhA and ycf1 were identified as highly divergent and may serve as potential molecular markers for phylogenetic analyzes.

To further elucidate the phylogenetic relationships among these 11 Persicaria species and clarify their taxonomic positions at the genomic level, a phylogenetic tree was reconstructed using the chloroplast genome sequences of these 11 species along with 53 other Polygonaceae species (Figure 7). The results demonstrated that all 11 species from Guangxi cluster within the genus Persicaria, forming distinct clades separate from the genera Calligonum and Fallopia. The Persicaria species in Guangxi can be further classified into three groups. P. chinensis and P. capitata (clade I) are perennials grown for their stingless foliage distinct from other species perhaps representing synapomorphies. P. perfoliata, P. maackiana and P. hastatosagittata are annual plants and their leaves are armed with spines. These characteristics are different from those of other species. Evolutionary relationships show that it belongs to clade II. Regarding their relationships, P. hydropiper, P. pubescens, P. tinctoria, P. longiseta, P. glabra, and P. lapathifolia exhibited closer phylogenetic affinities. This group of plants is classified into clade III groups. P. lapathifolia and P. glabra have more similarities in morphology, and their genetic evolutionary trees show that they are also more closely related. They belong to the clade IIIa. Similarly, P. longiseta, and P. tinctoria formed a closely related subgroup IIIb, while P. pubescens and P. hydropiper showed a sister relationship. This confirms that chloroplast genomic data can serve as a reliable tool for species identification and helps to address the gap in species discrimination within the Persicaria species using chloroplast genomes.

The results of the dating analyzes indicate that the Persicaria genus diverged from other members of the Polygonaceae family during the late Cretaceous epoch, approximately 81.89 million years ago (Ma), with a 95% highest posterior density (HPD) interval of 66.71 to 96.72 Ma. The two main clades within the Persicaria genus, designated as Clade I and Clade II, diverged approximately 64.6 Ma (95% HPD: 43.98-83.74 Ma). Clade I, which includes Persicaria hydropiper and Persicaria filiformis, likely diverged around 51.88 Ma (95% HPD: 32.67-72.77 Ma). The stem node of Persicaria filiformis and Persicaria amphibia indicates that it diverged from its sister clades during the Eocene epoch, with a minimum age of 43.71 Ma (26.4-61.08 Ma) (Figure 8). In Guangxi, Persicaria species are predominantly found within Clade I. The stem node of Clade Ib, which includes Persicaria hastatosagittata and Persicaria maackiana, diversified approximately 32.82 Ma (18.01-49.45 Ma). Within Clade II, which includes the Persicaria chinensis and Persicaria capitata diverged during the Eocene epoch, approximately 39.72 Ma (20.79-57.7 Ma). The stem node of the Clade II is a little earlier time frame to the clade Ib diversified around 32.82 (18.01-49.45) Ma (Figure 8).