Assembly and comparison of organelle genomes for investigating the maternal inheritance in Peganum (Nitrariaceae) Authors: Hao Xu, Shuang Han, Yu Niu, Yun Han, Jingya Yu, Qingqing Zheng, Guangyuan Zhang, Shilong Chen, Xiaofeng Chi, Faqi Zhang Published: 2026 ABSTRACT AbstractBackgroundPeganum are arid-adapted perennial diploid sandy plants distributed in Arid Central Asia (ACA). This genus features a variety of intermittent and hybrid entities (IHEs), making it an excellent model for studying hybrid speciation. Few organellar genomes have been published for Peganum, and the identification of the hybrid parental species in IHEs remains unresolved. Therefore, we conducted a systematic comparative organellar genome analysis and maternal inheritance study for Peganum.ResultsHere, we assembled the plastomes and mitogenomes of Peganum using sequencing data from Illumina, HiFi, and Oxford Nanopore platforms. The assembled mitogenomes ranged in total length from 426,866 bp to 447,055 bp and were recovered as single circular DNA configurations supported by long-read sequencing data, containing 33–35 protein-coding genes (PCGs), 19 tRNAs, and 3 rRNAs. The plastomes ranged from 159,967 bp to 160,739 bp and contained 86 PCGs, 37 tRNAs, and 8 rRNAs. Genome collinearity analysis revealed more structural inconsistencies in the mitogenomes compared to the plastomes. The mitogenomes and plastomes contained 824 and 619 repeated sequences, respectively. Codon usage analysis revealed a preference for codons ending in A/T. The genes nad2, rpoC1, and atpF, associated with mitochondrial respiration and plastid gene expression and ATP synthesis, showed evidence of purifying selection, indicating strong functional constraints during Peganum evolution. DNA transfer analysis revealed significant genetic interactions between the two organellar genomes and between the organellar and nuclear genomes. Sequence similarity and phylogeny are broadly consistent with predominantly maternal origins of IHEs formed through hybridization, although more complex inheritance signals cannot be excluded.ConclusionsWe assembled and annotated the organellar genomes of Peganum and conducted a comparative analysis. Our findings contribute to the organellar genome database for Peganum and reveal both the stability and heterogeneity of organellar genomes. The results provide a representative case for investigating maternal inheritance in IHEs.Supplementary InformationThe online version contains supplementary material available at 10.1186/s12864-026-13034-w.Keywords: Peganum, Organelle genomes, Maternal inheritance, Comparative analysis, Hybrid speciation Background The genus Peganum L. comprises perennial, herbaceous dicotyledons with deeply divided or finely dissected leaves, ranging from erect to spreading growth forms [1]. Historically, Peganum was classified into Zygophyllaceae based on the Engler system [2, 3]. Recent studies in palynology, embryology, phytochemistry, and molecular phylogeny have demonstrated a closer relationship between Peganum and Nitrariaceae rather than Zygophyllaceae [4, 5]. Consequently, the Angiosperm Phylogeny Group placed Peganum within the subfamily Peganoideae of Nitrariaceae [6]. Peganum species are commonly found in arid and saline environments, serving as key components of desert vegetation [7, 8]. These plants are widely distributed across the arid regions of Asia, northern and southern Africa, Europe, and North America, including P. harmala L., P. mexicanum A. Gray, P. nigellastrum Bunge, and P. multisectum (Maxim.) Bobrov [9]. As a medicinal and poisonous plant, Peganum has a long history in traditional medicine, where its seeds and entire plant are used to treat various ailments [10]. In regions with strong religious traditions, Peganum and its derivatives are believed to ward off evil spirits and bring health and prosperity [11]. The primary pharmacologically active compounds in Peganum are alkaloids, with seeds containing higher concentrations (primarily harmine and harmaline) than other plant organs [12]. Numerous experimental studies have demonstrated that Peganum alkaloids exhibit strong anticancer properties. For example, harmine has been shown to induce pro-death autophagy and apoptosis [13], while harmaline inhibits esophageal squamous cell carcinoma growth by targeting mTOR [14]. Harmaline, harmine, and deoxyvasicine exhibit neuroprotective effects by inhibiting acetylcholinesterase [15, 16]. Beyond medical uses, aqueous extracts of Peganum have demonstrated chemosensitizing effects on other plants, making them useful as herbicides in crop protection [17]. Accurate differentiation and identification of Peganum species are essential to ensuring the safe and effective use of these plants in medicine. Mitochondria and plastids, derived from alphaproteobacterial and cyanobacterial endosymbionts, are vital organelles in angiosperms, driving respiration and photosynthesis [18, 19]. Over evolutionary time, both genomes have undergone significant reduction due to gene loss and transfer to the nuclear genome [20]. However, the mitogenome exhibits far greater structural and size variability than the plastome [21]. Angiosperm mitogenomes vary greatly in size, ranging from 65.7 Kb in Viscum scurruloideum Barlow to 11.3 Mb in Silene conica L [22, 23]. They exhibit dynamic architectures, including circular, linear, and branched forms, and these structural features are influenced by extensive repetitive sequences that mediate homologous recombination [24–26]. These features enhance genomic diversity and adaptability but complicate sequencing and assembly, often due to plastid DNA contamination [27]. In contrast, plastomes are generally smaller and more stable. They typically display a conserved circular quadripartite structure, in which inverted repeats separate the large and small single-copy regions [28, 29]. Although plastomes range from 11.3 Kb to 242.5 Kb and exhibit occasional structural rearrangements, their overall organization is less complex than mitogenomes [30]. Advancements in high-throughput sequencing have deepened our understanding of mitogenome diversity. Despite these advancements, many plant mitogenomes and plastomes remain underexplored, particularly in certain lineages such as Peganum. As a reliable and convenient molecular marker, the use of organellar genomes for phylogenetic reconstruction also enables the accurate selection of the correct species, which is crucial for avoiding medicinal errors and maximizing therapeutic efficacy. Investigating the evolutionary dynamics and inheritance patterns of plastomes and mitogenomes is particularly relevant for Peganum. This genus contains closely related species and frequent hybrid lineages. Organelle genomes can thus provide important insights into species divergence and hybridization processes [31–33]. Although members of Peganum are morphologically distinct, interspecific hybridization can occur among geographically overlapping species (except for P. mexicanum, which is uniquely distributed in North America). The potential for hybridization likely reflects the strong geographic continuity of most species across Arid Central Asia (ACA) [34]. The resulting hybrids display distinctive mixed phenotypes, classified into three intermittent and hybrid entities (IHEs) [35]. A previous study similarly identified IHEs in Peganum (referred to as Peganum sp.), but did not attribute the speciation of these IHEs to interspecific gene flow events [36]. The hybridization relationships among Peganum species provide a valuable model for investigating the evolution of organelle genome. However, the interspecific hybridization network complicates the evolutionary dynamics of organelle genomes in Peganum. Despite this significance, studies on Peganum organelle genomes remain limited, hindering comprehensive large-scale comparative and evolutionary analyses. Expanding the Peganum organelle genome database with new data is thus urgently needed. In this study, we successfully assembled and annotated the mitogenomes of P. multisectum and P. nigellastrum for the first time, while also supplementing the plastome data for P. harmala, P. multisectum, and P. nigellastrum. Our analysis included genome characteristics, repetitive sequences, codon usage preferences, selection signatures, and sequence transfers from organelle and nuclear genomes. Additionally, we extracted organelle genes from IHEs for phylogenetic analysis to examine the contributions of different organelle genomes to maternal inheritance. These findings provide critical insights into the functional evolution of organelle genomes, laying a solid foundation for advancing studies in evolutionary mechanisms and population genetics. Materials and methods Plant sample collection, DNA extraction, and genome sequencing P. nigellastrum seeds were collected from the ACA region, located at Alxa Right County, Alxa League, Inner Mongolia Autonomous Region, China (101.71°E, 39.19°N). The seeds were germinated and grown in a controlled growth chamber under a 16-hour light (26℃)/8-hour dark (20℃) with 60% constant humidity. Fresh leaves from the seedlings were harvested for genome sequencing. High-quality genomic DNA was extracted using a DNeasy Plant Mini Kit (Qiagen, USA) following the manufacturer’s protocol to construct 15-Kb REV-HiFi libraries for circular consensus sequencing (CCS). A total of 10.94 Gb of HiFi reads were generated for the P. nigellastrum using a PacBio Revio platform. Assembly and annotation of organelle genome The ONT sequencing data used to assemble the P. multisectum mitogenome were obtained from the National Genomics Data Center (CRA016417). Third-generation sequencing reads were aligned to the P. harmala mitogenome data (GenBank: MK431826) using minimap2 v2.26 [37], and mitogenome-specific reads were extracted for subsequent assembly. De novo assemblies of long reads for both P. multisectum and P. nigellastrum were performed using Flye v2.9.5 [38]. The final genome sequence was obtained by short-read polishing using Pilon v1.24 [39]. Bandage v0.8.1 [40] was used to visualize the mitogenome, and extended fragments from the plastome and nuclear genomes were manually removed. Contigs were excluded if they corresponded to abnormally long nuclear-derived sequences or quadripartite plastome-derived structures, or if they exhibited anomalously high or low sequencing depth compared with the majority of mitochondrial contigs. To evaluate the completeness of the mitochondrial genome assembly, we aligned the long-read data to the assembled mitochondrial genome using minimap2 v2.26 [37] and calculated the coverage depth with SAMtools v1.15.1 [41]. Short-read data sequenced on the Illumina NovaSeq 6000 platform for Peganum plastome assembly were obtained from NGDC (CRA016417, P. multisectum; CRA016483, samples: PH1_1 and PN7_10). We then used Bowtie2 [42] to map clean reads to the seed database (i.e., embryophyte plastomes database, default: embplant_pt). Target-associated reads were assembled into contigs with SPAdes v3.14.0 [43] via GetOrganelle v1.7.5 [44] using the following parameters: -R 15 -k 21,45,65,85,105 -w 102 -F embplant_pt. Organelle genome annotation was conducted through a multi-step process. Complete organelle genomes were annotated using Geseq [45] from CHLOROBOX (https://chlorobox.mpimp-golm.mpg.de/geseq.html). Additionally, protein coding genes (PCGs) of the plastome were further annotated using PGA (Plastid Genome Annotator) [46], and manual adjustments were made to start/stop codons and exon/intron boundaries using Geneious v4.8.3 [47]. To validate tRNA gene predictions, we used the online program tRNAscan-SE v2.0 [48] (http://lowelab.ucsc.edu/tRNAscan-SE/index.html). The final organelle genome structures were visualized using OGDRAW (v1.3.1, https://chlorobox.mpimpgolm.mpg.de/OGDraw.html). Analysis of repeat fragments, prediction of RNA editing sites, and codon usage preferences We used the MIcroSAtellite (MISA) identification tool to identify short tandem repeats [49]. The minimum threshold for the number of repeat units was set as follows: 10 for mononucleotide repeats, 5 for dinucleotide repeats, 4 for trinucleotide repeats, and 3 for both tetranucleotide, pentanucleotide, and hexanucleotide repeats. Tandem and dispersed repeats were analyzed using Tandem repeats finder v4.09, (TRF, https://tandem.bu.edu/trf/trf.submit.options.html) and the REPuter network server, respectively [50, 51]. Predicted RNA editing sites were identified using the online tool PREPACT3 (http://www.prepact.de/) [52] with explicitly specified parameters: BLASTX mode, forward editing option enabled, amino acid result extension set to 2, and filter threshold set to 30%. In addition, a cutoff value of 0.001 was applied to remove low-confidence predictions and reduce false positives arising from spurious sequence similarity. Organelle genomes were extracted to obtain PCGs using Python scripts from PhyloSuite v1.2.3 [53]. We used CodonW v1.4.2 [54] to analyze codon usage bias and calculate the relative synonymous codon usage (RSCU) and the effective number of codons (ENC) with default settings. To determine the GC content of PCGs, we used the online CUSP software (https://www.bioinformatics.nl/cgi-bin/emboss/cusp) to analyze the GC content of the first, second, and third codon positions (GC1, GC2, and GC3). GC content visualization was implemented using the ggplot2 package [55] in Rscript v4.3.1 [56]. To evaluate the relative contributions of mutation pressure and natural selection to codon usage bias, neutrality plot analysis, ENC-GC3 plot analysis, and PR2-plot analysis were performed. In the neutrality plot analysis, a scatter plot was constructed with GC3 as the x-axis and GC12 (the average of GC1 and GC2) as the y-axis, and the regression slope between GC12 and GC3 was used to assess the relative influence of mutation pressure and natural selection [57]. ENC-GC3 plot analysis was conducted by plotting ENC values against GC3. Deviations of observed ENC values from the standard curve were used to infer the strength of selective constraints on codon usage [58, 59]. In addition, PR2-plot analysis was performed by plotting G3/(G3 + C3) against A3/(A3 + T3) to reveal asymmetric nucleotide usage at the third codon position and the combined effects of mutation bias and natural selection [60]. Selection analysis of Peganum To investigate natural selection pressures during the evolution of Peganum, we performed adaptive evolution analyses based on protein-coding genes (PCGs). All PCGs were aligned using MAFFT v7.313 [61] with default parameters, and the resulting alignments were converted into PHYLIP format using the Convert Sequence Format module implemented in PhyloSuite v1.2.3 [53]. Maximum-likelihood (ML) phylogenetic trees were reconstructed using IQ-TREE v2.0.3 [62], with the best-fit substitution models selected by ModelFinder v.1.6.12 [63], and node support was assessed with 10,000 bootstrap replicates. Selective pressure was evaluated by estimating the ratio of nonsynonymous to synonymous substitution rates using the codeml program implemented in PAML v4.9j [64]. Branch models were applied to test for variation in selective pressure among lineages, including a null model assuming a single ω ratio across all branches (model = 0, NSsites = 0) and an alternative branch model allowing foreground branches comprising the three species of Peganum to have distinct ω values (model = 2, NSsites = 0). Likelihood ratio tests (LRTs) [65] were conducted to compare nested models and to assess the statistical significance of differences in ω values, with significance evaluated based on a chi-square distribution (P < 0.05). To account for multiple testing, P-values obtained from LRTs were further corrected using the false discovery rate (FDR) approach, and the corresponding Q-values were calculated. Collinear analysis and identification of homologous sequences We used LINKVIEW2 (https://yangjianshun.github.io/LINKVIEW2/) to analyze plastome and mitogenome collinearity across three Peganum species to identify large-scale structural variations in organelle genomes (identity ≥ 85%, block size ≥ 100 bp). To investigate potential DNA migration events, we assessed sequence homology between mitogenome and plastome using blastn v2.13.0+ [66] with the parameters “-evalue 1e-10, -perc_identity 85, -outfmt 6”. Plastome and mitogenome DNA transfer maps were generated using Circos [67] in TBtools-Ⅱ [68]. Additionally, homology between the nuclear and organelle genomes was detected using blastn v2.13.0 + with the parameters “-evalue 1e-10, -perc_identity 95, -outfmt 6”, based on the whole-genome data of P. multisectum. The selection of identity thresholds was based on the following considerations: (1) plastid-derived sequences in plant mitogenomes are generally ancient and have accumulated substantial substitutions after transfer; (2) because nuclear genomes are large and prone to spurious matches, a stringent identity threshold was applied to minimize false-positive matches. Organelle gene assembly and alignment of IHEs The short-read data for assembling the organelle gene of IHEs were sequenced on the Illumina NovaSeq 6000 platform and obtained from the NGDC (CRA016483, samples: 3 NH, 4 MH, and 30 NM individuals). Three plastomes and three mitogenomes of Peganum used in this study served as references to extract PCGs of IHEs from the next-generation sequencing data. A local database was constructed using organelle genes from Peganum by makeblastdb (GWHESQQ00000000). We used GeneMiner v1.0.1 [69] to assemble a total of 48 shared genes for the mitogenome (29 PCGs, 3 rRNAs, and 16 tRNAs) and 111 shared genes for the plastome (77 PCGs, 4 rRNAs, and 30 tRNAs). Homologous sequences between Peganum and IHEs organelle genomes were identified using blastn v2.13.0+ [66]. This task was performed adhering to specific parameters: “-evalue 1e-10, -num_alignments 3, -outfmt 6”. This analysis aimed to determine the maternal inheritance of IHEs based on sequence similarity. Violin plots were generated using the ggplot2 package [55] and the “geom_violin()” function to visualize sequence identity across different species. We applied the Wilcoxon test with FDR correction to evaluate whether sequence homology differed significantly among species. Phylogenetic reconstruction of maternal inheritance We extracted all shared protein-coding genes (PCGs) from the plastome and mitogenome sequences of P. harmala, P. multisectum, P. nigellastrum, NM, NH, MH, and the outgroup N. tangutorum Bobrov (MK431824 and NC_059748) using Python scripts from PhyloSuite v1.2.3 [53]. The homologous sequences were aligned using default parameters in MAFFT v7.313 [61]. Frameshift variations were corrected using MACSE v2.03 [70], poorly aligned regions were excluded using Gblocks v0.91b [71] with default parameters. To ensure the accuracy of the phylogenetic analysis, we constructed single-gene maximum-likelihood (ML) trees based on the shared PCGs across all species using IQ-TREE v2.0.3 (st = CODON11, m = MFP) [62], and then merged the 99 resulting trees to reconstruct the species tree using ASTRAL v5.7.8 [72]. Local posterior probabilities (local PP) and quartet scores were estimated using ASTRAL v5.7.8 (local PP: -t 3; quartet scores: -t 1). Gene concordance factor (gCF) and site concordance factor (sCF) were calculated on the species tree using IQ-TREE v2.0.3 (--gcf, --scf 1000). Based on the 99 shared PCGs, we used Phyparts v0.0.1 [73] to examine discordance between gene trees and the species tree. Cloud tree plots, displaying conflicts between the single-gene trees and the species tree, were generated using Toytree v2.0.1 [74]. Results Organelle genome assembly, annotation, and characteristics in Peganum In this study, we successfully assembled two complete circular mitogenomes and three circular plastomes (Fig. 1, S1). For mitogenome assembly, only one single circular contig of mitochondrial origin was recovered, representing the dominant configuration supported by long-read sequencing data. Long-read coverage depth profiles (Fig. S2) further strongly support the high completeness of the mitogenome and the structural continuity of the circular molecule (Table S1). The newly sequenced and published plastomes and mitogenomes of P. harmala, P. multisectum, and P. nigellastrum were assembled with lengths ranging from 159,967 bp to 160,739 bp for plastomes, and 426,866 bp to 447,055 bp for mitogenomes (Fig. 1; Table 1). The GC content of plastomes ranged from 37.3% to 37.5%, and the GC content of mitogenomes remained constant at 45.5%. Gene maps of the (A) P. multisectum, B P. nigellastrum, and C P. harmala mitogenomes. Draft assembly graphs of D P. multisectum and E P. nigellastrum mitogenomes Table 1General characters of Peganum mitogenomes and plastomes. (mtDNA = mitogenome, cpDNA = plastome)TypeP. multisectumP. nigellastrumP. harmalamtDNAcpDNAmtDNAcpDNAmtDNAcpDNASize439,502160,739426,866160,070447,055159,967GC%45.537.345.537.545.537.5Genes551315513157131tRNAs193719371937rRNAs383838PCGs338633863586AccessionPQ859560PQ859557PQ859559PQ859558MK431826PQ859556 General characters of Peganum mitogenomes and plastomes. (mtDNA = mitogenome, cpDNA = plastome) The plastomes of Peganum exhibited a typical quadripartite structure (Fig. S1), comprising a large single-copy region (LSC), a small single-copy region (SSC), and two inverted repeat regions (IRs). The number and types of genes encoded by the plastomes were highly conserved, with each plastome containing 131 genes: 86 PCGs, 37 tRNAs, and 8 rRNAs (Table 2). Table 2Genes content of the organelle genomes in Peganum. (PH = P. harmala; PM = P. multisectum; PN = P. nigellastrum, genes marked with * indicate the presence of an additional copy)Gene groupGene nameMitogenomePlastomeATP synthaseatp1, atp4, atp6, atp8, atp9atpA, atpB, atpE, atpF, atpH, atpINADH dehydrogenasenad1, nad2, nad3, nad4, nad4L, nad5, nad6, nad7, nad9, nad9*(PH, PN)ndhA, ndhB, ndhB*, ndhC, ndhD, ndhE, ndhF, ndhG, ndhH, ndhI, ndhJ, ndhKUbiquinol cytochrome c reductaseccmB, ccmC, ccmC* (only PH), ccmFC, ccmFN(PH, PM)\Cytochrome c oxidasecox1, cox2, cox3\MaturasematRmatKLarge subunit of ribosomerpl5, rpl10, rpl16rpl2, rpl2*, rpl14, rpl16, rpl20, rpl22, rpl23, rpl23*, rpl32, rpl33, rpl36Small subunit of ribosomerps1, rps3, rps4, rps10, rps12rps2, rps3, rps4, rps7, rps7*, rps8, rps11, rps12, rps12*, rps14, rps15, rps16, rps18, rps19, rps19*DNA dependent RNA polymerase\rpoA, rpoB, rpoC1, rpoC2Ribosome RNArrn5, rrn18, rrn26rrn4.5, rrn4.5*, rrn5, rrn5*, rrn16, rrn16*, rrn23, rrn23*Transfer RNAtrnC-GCA, trnC-GCA*, trnD-GUC, trnE-UUC, trnF-GAA, trnF-GAA*, trnG-GCC, trnH-GUG, trnI-CAU, trnK-UUU, trnM-CAU, trnN-GUU, trnP-UGG, trnP-UGG*, trnQ-UUG, trnS-GCU, trnS-GGA, trnS-UGA, trnS-UGA*, trnW-CCA, trnY-GUAtrnA-UGC, trnA-UGC*, trnC-GCA, trnD-GUC, trnE-UUC, trnF-GAA, trnG-GCC, trnG-UCC, trnH-GUG, trnI-CAU, trnI-CAU*, trnI-GAU, trnI-GAU*, trnK-UUU, trnL-CAA, trnL-CAA*, trnL-UAA, trnL-UAG, trnM-CAU, trnfM-CAU, trnN-GUU, trnN-GUU*, trnP-UGG, trnQ-UUG, trnR-ACG, trnR-ACG*, trnR-UCU, trnS-GCU, trnS-GGA, trnS-UGA, trnT-GGU, trnT-UGU, trnV-GAC, trnV-GAC*, trnV-UAC, trnW-CCA, trnY-GUAPhotosystem genes\psaA, psaB, psaC, psaI, psaJ, psbA, psbB, psbC, psbD, psbE, psbF, psbH, psbI, psbJ, psbK, psbL, psbM, psbN, psbT, psbZCytochrome\petA, petB, petD, petG, petL, petNOther genescob, mttB, sdh4accD, ccsA, cemA, clpP, rbcL, ycf1, ycf2, ycf2*, ycf3, ycf4 Genes content of the organelle genomes in Peganum. (PH = P. harmala; PM = P. multisectum; PN = P. nigellastrum, genes marked with * indicate the presence of an additional copy) Regarding gene content, the mitochondrial genomes of Peganum species consistently included 3 rRNAs and 19 tRNAs, with identical gene types (Table 2). However, the number of PCGs varied: P. multisectum contained 33, P. nigellastrum had 33, and P. harmala had 35 PCGs. Specific gene content variations in the mitogenomes included the absence of ccmFN in P. nigellastrum, an additional copy of ccmC in P. harmala, and the absence of the second nad9 in P. multisectum. Repeat elements in Peganum organelle genomes The organelle genomes of Peganum contained six types of repeat sequences: simple sequence repeats (SSRs), tandem repeats (TRs), forward repeats (FRs), palindromic repeats (PRs), reverse repeats (RRs), and complementary repeats (CRs). SSRs, also known as microsatellite markers, are short repetitive sequences that are uniformly distributed across eukaryotic genomes, primarily composed of 1–6 nucleotide tandem repeats. The identified SSRs were categorized into six types based on the number of nucleotides in the repeat unit: mononucleotide, dinucleotide, trinucleotide, tetranucleotide, pentanucleotide, and hexanucleotide (Fig. 2A, B). Our study revealed similar distributions of these six SSR types in Peganum. In both organelle genomes (mitogenome/plastome), mononucleotide repeats were the most abundant (58–62/46–59), accounting for approximately 39%/78% of the total SSRs. Tetranucleotide repeats were the second most common (42–45/5–8), comprising around 30%/10% of the total. Dinucleotide, trinucleotide, pentanucleotide, and hexanucleotide repeats were less frequent, collectively contributing about 31%/12% to the total SSRs in mitogenomes and plastomes, respectively. Analysis of repeat fragments from the Peganum organelle genomes. Type of simple sequence repeats (SSRs) in (A) mitogenomes and (B) plastomes. The Mono, Di, Tri, Tetra, Penta and Hexa represent Monomeric, Dimeric, Trimeric, Tetrameric, Pentameric and Hexameric, respectively. The number of dispersed repeats in mitogenomes (C) and plastomes (D). The count of scattered repeats is denoted by TR (Tandem repeats), PR (Palindromic repeats), FR (Forward repeats), RR (Reverse repeats), and CR (Complementary repeats). (PH = P. harmala; PM = P. multisectum; PN = P. nigellastrum) The types of dispersed repeats were also highly similar among different Peganum species, although notable differences were observed between mitogenomes and plastomes (Fig. 2C, D). In plastomes, three Peganum species exhibited comparable numbers of TRs (26–30), FRs (20–27), and PRs (23–30), with no occurrences of RRs or CRs. In mitogenomes, TRs were the most abundant dispersed repeats (35–40), followed by PRs and FRs (19–21), while RRs and CRs were the least frequent, with average counts not exceeding 10. Except for the conserved inverted repeat (IR) regions in plastomes, all remaining dispersed repeat sequences were short, with lengths shorter than 50 bp (Table S2). In contrast, mitogenomes contained several long repeat pairs with high sequence identity. Specifically, P. harmala, P. multisectum, and P. nigellastrum had 5, 1, and 2 repeat pairs, respectively, with lengths ≥ 500 bp and sequence identity ≥ 90%. These long and highly similar repeats are considered potentially recombination-competent. Prediction of RNA editing events RNA editing refers to the insertion, deletion, or substitution of nucleotides during RNA transcription, occurring in mitochondria, chloroplasts, and the nucleus [75]. In this study, RNA editing sites were predicted in protein-coding genes (PCGs) of Peganum organelle genomes. The number of predicted RNA-editing sites was largely consistent in the same organelle genomes. 442–479 predicted RNA-editing sites were identified across 31 PCGs in mitogenomes, and 249–255 predicted RNA-editing sites were detected across 51 PCGs in plastomes (Fig. 3). In mitogenomes, mttB exhibited the highest average number of predicted RNA-editing sites, with 37 sites identified. Only one predicted RNA-editing event was found in rps3 and atp9. Among all predicted RNA-editing sites, 195 codons were predicted to change to Leucine, accounting for approximately 43% of the total (Table S3, S4). In plastomes, ycf1 had the highest number of predicted RNA-editing sites, with an average of 49 identified. However, 21 PCGs contained only one predicted RNA-editing sites. Predicted RNA editing in plastomes showed a preference for converting codons to Phenylalanine, with 69 codons affected, representing about 27% of the total predicted RNA-editing sites. Prediction of RNA editing sites based on the PCGs of (A) mitogenomes and (B) plastomes. (PH = P. harmala; PM = P. multisectum; PN = P. nigellastrum) We observed that amino acid changes at the first and second nucleotide positions of PCGs were the primary contributors to predicted RNA-editing sites, with the second position being the most frequently altered (Table S3, S4). No changes were observed at the third nucleotide position. Some predicted sites in mitogenomes were associated with codon changes corresponding to stop codons in atp9 and ccmFN. Additionally, predicted RNA-editing sites in the mitogenomes caused approximately 46% of hydrophilic amino acids to be converted to hydrophobic amino acids, whereas only 37% were converted in plastomes. Conversely, around 8% of hydrophobic amino acids in mitogenomes were transformed into hydrophilic amino acids, compared to 13% in plastomes. Analysis of codon usage bias In the complete organelle genomes of Peganum (P. multisectum, P. nigellastrum, P. harmala), we identified 9,696, 9,327, and 10,167 codons in the mitogenomes, and 26,381, 26,378, and 26,367 codons in the plastomes, respectively (Table S5, S6). Both organelle genomes encoded all 20 amino acids, utilizing 61 different codon types. However, the most frequently used codons differed between the mitogenomes and plastomes, with UUU and AUU being the most prevalent in the mitogenomes and plastomes, respectively. Statistical analysis of codon usage for amino acids revealed a similar trend in both organelles (mitogenome/plastome). Isoleucine was encoded by the highest number of codons, with 2,264/6,726 codons (7.76%/8.50% of the total), followed by Glycine, with 2,019/5,420 codons (6.92%/6.85%). In contrast, Cysteine was encoded by the fewest codons, with only 416/900 codons (1.43%/1.14%). Among these codons, 29/30 in the mitogenomes/plastomes were used more frequently than expected (relative synonymous codon usage, RSCU > 1), while 30/29 were used less frequently than expected (RSCU < 1) (Fig. S3). These differences were primarily attributed to variations in the RSCU values of Alanine (GCA), Serine (UCC), and Threonine (ACC) between the organelles. Furthermore, only Methionine (AUG) and Tryptophan (UGG) exhibited no codon bias (RSCU = 1), indicating that most amino acids display a biased codon usage. A comparative analysis of RSCU values between the two organelles demonstrated a high degree of codon preference consistency in the mitogenomes and plastomes of Peganum. We calculated the GC content at the first (GC1), second (GC2), and third (GC3) codon positions of the shared PCGs in the organelle genomes (mitogenome/plastome). Across the three species, the GC content of PCGs was highly similar (Fig. 4A, B). The GC1 values ranged from 36.4% to 58.5%/34.4% to 62.2%, GC2 values from 31.0% to 52.2%/25.7% to 59.0%, and GC3 values from 28.6% to 57.2%/30.5% to 56.6%. The average GC content (GCall) across these positions was 42.4%/38.91%, indicating a preference for A/T bases at all codon positions in Peganum organellar PCGs (Fig. 4C, D). Additionally, we calculated the effective number of codons (ENC) for the PCGs. The ENC ranged from 45.15 to 60.66 in the mitogenomes and from 24.66 to 61.00 in the plastomes, with average ENC values of 53.56 and 47.18, respectively. These results suggest that Peganum organellar PCGs exhibit relatively low codon usage bias, approaching a random codon selection pattern. GC content of different positions based on the PCGs of (A) plastomes and (B) mitogenomes. Variation of ENC was indicated from (C) plastomes and (D) mitogenome. Value sizes were suggested using the different colors. E GC content of different positions from analysis of neutrality plot. (PH = P. harmala; PM = P. multisectum; PN = P. nigellastrum) Neutrality plot analysis showed a weak association between GC12 and GC3 in Peganum organellar genomes. In mitogenomes, GC12 exhibited a weak and non-significant negative correlation with GC3 (r2 = 0.004–0.012), whereas plastome genes showed a weak positive correlation with similarly low explanatory power (r2 = 0.019–0.021) (Fig. 4E). ENC-GC3 plots showed that most genes from both organelles deviated markedly from the expected standard curve and were mainly distributed below it (Fig. S4). This pattern suggests that nucleotide composition alone cannot explain the observed codon usage. PR2-plot analysis further indicated unequal usage of synonymous bases at the third codon position in all three Peganum species (Fig. S5), with A3 > T3 and G3 > C3 in mitogenomes, and A3 > T3 and C3 > G3 in plastomes. Selection analysis in Peganum Synonymous mutations do not alter amino acid sequences, whereas nonsynonymous mutations cause amino acid changes. Both types of mutations serve as indicators of evolutionary pressures. The dN/dS ratio (also known as Ka/Ks) represents the ratio of the nonsynonymous substitution rate (dN) to the synonymous substitution rate (dS), offering a measure of selection pressure on proteins during evolution [76]. dN/dS > 1 suggests positive selection, dN/dS = 1 indicates neutral selection, and dN/dS < 1 implies purifying selection [77]. Likelihood ratio tests were evaluated using FDR-corrected Q values, with genes showing Q < 0.05 considered statistically significant (Table S7). In mitogenome, only nad2 exhibited a significant difference between the foreground and background branches after multiple-testing correction. The estimated ω value for the foreground branch (ω = 0.56) was markedly lower than that of the background (ω = 1.64), indicating strengthened purifying selection along the Peganum genus. All other mitochondrial genes showed no significant branch-specific signals and were generally characterized by ω values below 1. In plastome, only rpoC1 (Q = 0.0167) and atpF (Q = 0.0218) retained significance after FDR correction. Both genes showed elevated ω values in the foreground branch (rpoC1: ω = 0.4286; atpF: ω = 0.6084) compared with the background, but remained below 1, suggesting that they more likely reflect relaxed purifying selection. No other plastome genes exhibited significant evidence of lineage-specific selection. Collinearity and DNA transfer of homologous sequences The plastomes of Peganum showed extremely high collinearity and were structurally conserved. Almost no genome rearrangements were detected. However, mitogenomes showed much lower collinearity and were characterized by frequent structural rearrangements (Fig. 5A, B, Table S8). Among the three species, the mitogenome structures of P. nigellastrum and P. multisectum were relatively similar, with only six large-scale rearrangements (> 1,000 bp) detected, totaling approximately 93 Kb. By comparison, the mitochondrial genome of P. harmala differed markedly from those of P. multisectum and P. nigellastrum, displaying numerous large-scale rearrangements. The cumulative rearranged lengths were approximately 318 Kb and 258 Kb, respectively. Collinear analysis of (A) mitogenomes and (B) plastomes from three Peganum species. C Homologous analysis based on the different organelles. The blue arc represents mitogenomes. The green arc represents plastomes. (PH = P. harmala; PM = P. multisectum; PN = P. nigellastrum) By fully assembling the plastomes and mitogenomes of Peganum and incorporating the published whole genome of P. multisectum, we evaluated patterns of sequence transfer within organellar genomes (identity ≥ 85%). In higher plants, intracellular transfer of genetic material is a common phenomenon in mitochondrial genomes, and sequence fragments derived from plastomes typically exhibit relatively low conservation [78]. Based on homologous sequence content and annotation, transfer locations were categorized into five types: exons, introns, rRNA, tRNA, and PCG regions (Fig. 5C). Sequence similarity analysis identified 22, 20, and 25 homologous fragments shared between the plastomes and mitogenomes of P. multisectum, P. nigellastrum, and P. harmala, respectively, with alignment lengths ranging from 75 to 981 bp, 75 to 1,005 bp, and 75 to 1,025 bp (Table S9). These homologous fragments were distributed across the organellar genomes, accounting for 1.17%, 1.33%, and 1.31% of the mitogenomes and 3.19%, 3.55%, and 3.67% of the plastomes, respectively. This distribution reflected the migration of sequences from plastomes to mitogenomes. Annotation of these homologous sequences revealed five complete shared genes (trnD-GUC, trnH-GUG, trnN-GUU, trnI-CAU, and trnM-CAU) and six shared incomplete PCGs (including ndhA-ndhI, rrn23, psbB, ycf2, psbC, and ycf3). Additionally, a homologous rRNA sequence (97–359 bp) was identified between the organellar genomes, suggesting that rrn26 in the mitogenome may have originated from rrn23 in the plastome. Since P. multisectum is the only Peganum species with a fully assembled nuclear genome, it was used as a model to explore the intracellular transfer of organellar genome sequences to the nuclear genome (Table S10, identity ≥ 95%). A total of 305 nuclear-organellar homologous sequences were identified (Table 3), covering all chromosomes of P. multisectum. Transferred sequences were detected regardless of whether they retained complete open reading frames (ORFs) or appeared as pseudogene fragments. Chromosome 2 contained the highest number of organelle-derived insertions. These included 17 complete ORFs and 58 pseudogene fragments, with a total aligned length of 16,179 bp. Chromosome 6 contained the second largest contribution, comprising 5 complete ORFs and 26 pseudogene fragments, totaling 15,312 bp. Chromosomes 9 and 11 contained the fewest organelle-derived sequences, with cumulative aligned lengths of only 1,442 bp and 1,480 bp, respectively. Table 3Statistics of fragments transferred from nuclear genome to organelle genome in P. multisectum. (Identity ≥ 95%)ChrLengthGene integrityComplete ORFsPseudogene fragmentsChr110,027721Chr216,1791758Chr32858416Chr45647421Chr53505215Chr615,312526Chr75588516Chr82514217Chr9144228Chr105948323Chr11148009Chr124635420Sum75,13555250 Statistics of fragments transferred from nuclear genome to organelle genome in P. multisectum. (Identity ≥ 95%) The maternal lineage of IHEs traced through phylogeny reconstruction The names of the three Peganum IHEs (MH, NH, and NM) were assigned based on their similarity to the presumed parental species (Fig. 6C). For example, the name of MH is derived based on its characteristics, which lie between those of P. multisectum and P. harmala. The previous analysis using nuclear data reconstructed the phylogeny of Peganum and its IHEs [35], however, extensive gene flow has left the parentage of the IHEs unresolved. Since organellar genomes are maternally inherited in plants, this study utilized newly assembled high-quality Peganum organellar genomes as references to identify shared PCGs in IHEs and reconstruct their maternal phylogenetic relationships. A The alignment ratio of IHEs with three Peganum species in different organelle genomes. B The topology of the species tree shows concordance among shared PCGs. (PH = P. harmala; PM = P. multisectum; PN = P. nigellastrum, **: Q < 0.05) A total of 48 and 111 shared genes (including PCGs, tRNAs, and rRNAs) were identified in the mitogenomes and plastomes of IHEs, respectively, and aligned against the reference organellar genomes of the three Peganum species (Fig. 6A, Table S11). The results showed that both mitogenome and plastome genes exhibited relatively low alignment rates with P. multisectum across all IHEs. For NM, the highest alignment rate was observed with P. nigellastrum. For MH and NH, the species showing the highest plastome alignment rates were P. harmala and P. nigellastrum, respectively. In contrast, in the mitogenome comparisons, P. harmala and P. nigellastrum consistently exhibited the highest alignment rates for both MH and NH (Q > 0.05, Wilcoxon test with FDR correction). The species tree, based on organellar PCGs, reconstructed the maternal phylogenetic relationships of Peganum and its IHEs, accurately reflecting their maternal lineage. As shown in Fig. 6B, NH and most NM individuals were more closely related to P. nigellastrum, indicating that both IHEs likely originated through hybridization events in which P. nigellastrum served as the female parent and P. harmala or P. multisectum as the male parent. MH, on the other hand, showed a closer relationship to P. harmala, strongly suggesting that MH originated through hybridization with P. harmala as the female parent and P. multisectum as the male parent. The topology of the species tree primarily supports a pattern in which P. multisectum forms the basal lineage, while P. harmala and P. nigellastrum are recovered as closely related sister taxa. Discussion Compared to the highly conserved plastomes, plant mitogenomes exhibit significant structural heterogeneity, characterized by a diverse range of circular and linear molecules. This complexity is primarily due to the presence of extensive repeat sequences, which facilitate gene recombination and the formation of subgenomic molecules in specific plant species [79]. Traditionally, obtaining plant organelle genomes involved isolating chloroplasts or mitochondria from plant tissues, followed by DNA extraction and the acquisition of complete genome sequences using first-generation sequencing technologies [80]. However, this isolation process was highly labor-intensive. With the rapid advancements in high-throughput second- and third-generation sequencing technologies, it is now possible to generate large amounts of high-accuracy sequencing data. These innovations have greatly streamlined the process of obtaining plant organelle genomes [81], facilitating robust investigations into species phylogeny, adaptive evolution, and speciation processes by providing high-resolution organellar markers. General features in Peganum organelle genomes The plastomes of Peganum exhibited a typical quadripartite structure, comprising 131 functional genes, with identical numbers and types of functional genes (Fig. S1). This study is the first to report the sequencing, assembly, and characterization of mitogenomes at the genus level for Peganum. The mitogenomes of the three Peganum species were assembled as single circular configurations of approximately 438 ± 10.2 Kb in length (Fig. 1). Recent studies have revealed substantial structural variation in mitogenomes across different plant species. For instance, the mitogenome of Corydalis pauciovulata Ohwi [82] consists of a single circular chromosome, while those of Cyperus stoloniferus Retz [83]. and Paphiopedilum micranthum Tang & F. T. Wang [84] contain two and 26 circular chromosomes, respectively. The assembly configurations (Fig. 1) recovered a highly contiguous and gap-free circular contig (Fig. S2), representing an assembled mitogenome configuration supported by long-read sequencing data rather than a definitive in vivo structure. Structural validation using long reads further supports this configuration: counts of reads spanning the major circularization junctions show substantial support in the final path (Table S1), with consistently high mapping quality. No alternative isoforms were detected above a conservative support threshold based on junction-spanning read counts. However, alternative low-frequency configurations cannot be entirely excluded given inherent sequencing and assembly limitations. The assemblies recover nearly all expected mitochondrial genes and show consistent long-read support across junctions, providing a reliable framework for downstream analyses. The functional gene content of Peganum mitogenomes was not entirely consistent, with notable differences in gene composition (Table 2). For example, P. nigellastrum lacked the ccmFN gene, which was present in the other two species. This gene is thought to enhance drought tolerance and plays an important role in the development of plant embryos and endosperms [85, 86]. However, the functional significance of its absence in P. nigellastrum remains unclear and requires further investigation. Guanine-cytosine (GC) content plays a critical role in shaping amino acid composition in proteomes during land plant evolution [87]. The GC content of the Peganum plastomes ranged from approximately 37.3% to 37.5% (Table 1), closely matching that observed in the plastomes of four Nitraria species (N. roborowskii Kom., N. sibirica Pall., N. sphaerocarpa Maxim., and N. tangutorum), which ranged from 37.15% to 37.32% [88, 89]. In contrast, the GC content of the Peganum mitogenomes was consistently 45.5%. Previous studies suggested that GC content is relatively conserved among plant mitogenomes. For example, the GC content of the mitogenomes of Astragalus membranaceus (Fisch.) Bunge [90], C. pauciovulata [82], and Angelica biserrata (R. H. Shan & Yuan) C. Q. Yuan & R. H. Shan [91] were 44.47%, 46.1%, and 44.8%, respectively. These findings indicate that the GC content of Peganum mitogenomes aligns with general characteristics observed in plant mitogenomes, reflecting their evolutionary stability and functional attributes. Comparative organelle genome analysis and intracellular gene transfer Through a comprehensive comparative analysis of the organellar genomes of three Peganum species, we found that their plastomes exhibited high collinearity (Fig. 5B). In contrast, the mitogenomes displayed lower collinearity and a greater number of genomic rearrangements (Fig. 5A). These rearrangements, involving homologous syntenic blocks, are crucial in the formation of genetic libraries in angiosperms, genome evolution, gene expression regulation, and species evolution [92]. Both SSRs and dispersed repeats were detected in the mitogenomes and plastomes of Peganum. These repeats serve as important markers for evolutionary analysis and play a key role in genomic plasticity and adaptive evolution [93–95]. In three Peganum species, we identified 824 mitogenomic and 619 plastid repeat sequences, including 518/266 SSRs and 234/262 dispersed repeats (Fig. 2, Table S2). The abundance of repeat sequences suggests that molecular recombination occurs more frequently in the mitogenomes compared to the plastomes [96]. These repeats are valuable not only for reconstructing interspecies phylogenetic relationships but also for analyzing intraspecific phylogeography [97, 98]. The expression of plant organellar genes involves various co-transcriptional or post-transcriptional RNA modifications, including RNA processing at the 5’ and 3’ ends, cis- and trans-splicing, and RNA editing [75]. RNA editing primarily occurs in organellar transcripts and has been widely documented in plants [99]. In land plants, the number of RNA editing sites varies significantly, ranging from none to several hundred. As plants evolve, the number of RNA editing sites generally decreases, with editing events being more frequent in early-diverging plants. Comparative studies suggest that RNA editing sites tend to be more abundant in early-diverging lineages and reduced in later lineages [100]. In this study, we predicted 442–479 RNA editing sites in the mitogenomes and 249–255 in the plastomes (Fig. 3), similar to previous studies on the organellar genomes of Hypopitys monotropa Crantz [101], Cymbidium ensifolium (L.) Sw [102]., and Bupleurum chinense Franch [103]. The number of predicted RNA editing sites varies across genes, genomes, and species. Among these genes, mttB and ycf1 exhibited the highest number of predicted RNA editing sites in the mitogenomes and plastomes, respectively. All predicted RNA editing sites were located at the first two positions of codons and exhibited C-U type editing, which is common in plants [104]. These predicted editing events were associated with codon changes that would alter encoded amino acids (Table S3). Codons play a crucial role in recognition and transitions in seed plants, and they are significant in the context of genetic mutations. During protein translation, the usage of certain synonymous codons is influenced by species-specific variations, leading to codon usage bias. This preference for certain synonymous codons plays a key role in shaping the genetic traits of organisms [105]. In this study, we assessed the GC content and RSCU of various positions in the mitochondrial and plastid PCGs of Peganum. The results showed a preference for A/T bases and A/T-ending codons at the third codon position, with significant consistency in codon usage patterns between mitogenomes and plastomes (Fig. 4A, B). This finding was consistent with previous studies on the organellar genomes of A. biserrata, Euphorbiaceae Juss., and Mesona chinensis Benth, indicating that both the mitogenomes and plastomes share similar codon usage preference [106–108]. The near-zero slopes in neutrality plot analyses (Fig. 4E) and the widespread deviation of ENC values from the standard curve collectively (Fig. S4) indicate that natural selection likely plays a more important role than mutation pressure in shaping codon usage bias in Peganum organellar genomes [109]. However, the low r2 values suggest that this tendency is weak rather than decisive, and that mutation pressure alone or selection alone cannot fully explain the observed patterns. In addition, the asymmetric base usage revealed by PR2-plot analyses implies that codon usage may be influenced by multiple selective constraints (Fig. S5). The dN/dS ratio is an important indicator for studying how genes in angiosperms respond to environmental stress and evolutionary pressures [110]. It provides insights into patterns of sequence evolution across lineages. In the mitogenome, nad2 was the only gene showing a significant signal after FDR correction, with a reduced ω value in the foreground branch (Table S7). As nad2 encodes a core subunit of mitochondrial complex I, this pattern may reflect functional constraints associated with mitochondrial activity [111]. In the plastome, rpoC1 and atpF exhibited significantly higher ω values in the foreground branch. However, these values remained below 1, consistent with a pattern of relaxed purifying selection rather than positive selection. rpoC1 encodes a core subunit of the plastid-encoded RNA polymerase [112], and atpF is a component of the ATP synthase complex. The observed patterns may indicate subtle lineage-specific variation, although no clear evidence for substantial functional divergence was detected [113]. Sequence migration and maternal phylogeny of IHEs in Peganum The analysis of DNA sequence transfer from plastomes to mitogenomes in Peganum revealed significant genetic interactions between the two organellar genomes (Fig. 5C). This finding aligns with previous studies on intracellular gene transfer (IGT) in other angiosperms [93, 114, 115]. Moreover, our study identified instances of IGT between organellar and nuclear genomes (Table S10). However, most gene sequences transferred from plastomes to mitogenomes or from nuclear genomes to organelles appear to be non-functional in the mitogenomes (Table S9, S10), suggesting that these sequences may have been lost or evolved into pseudogenes during organellar genome evolution [116]. Using high-quality assembled Peganum organellar genomes, we explored the maternal inheritance relationships of MH, NH, and NM by integrating PCGs, tRNAs, and rRNAs, providing critical insights into the mechanisms of speciation in IHEs. Alignment ratio analyses revealed contrasting inheritance signals between plastid and mitochondrial genomes in IHEs (Fig. 6A). Plastome alignments showed a clear bias toward a single Peganum species, supporting a predominantly maternal origin of plastids in IHEs. In contrast, mitochondrial gene alignments for both MH and NH showed elevated similarity to P. harmala and P. nigellastrum simultaneously, suggesting a more complex inheritance pattern. One possible explanation is occasional paternal leakage. This phenomenon has been reported in plants, particularly under interspecific hybridization, and may result in sporadic transmission of paternal mitochondria to hybrid offspring [117, 118]. Alternatively, mitochondrial heteroplasmy may also contribute to this pattern, as the coexistence of multiple mitochondrial haplotypes within individuals could reduce apparent sequence similarity between IHEs and their maternal parent [119]. The ASTRAL species tree further clarifies these patterns. The backbone topology, which resolves major lineages and the putative maternal origins of IHEs, is supported by relatively high local posterior probabilities, quartet scores, and gene- and site-concordance factors (Fig. 6B, S6, S7). However, substantial conflict is observed within individual IHE lineages, where support values are generally low and fail to resolve well-supported subclades. Such discordance is consistent with recurrent hybridization among the three Peganum species, which is expected to generate heterogeneous phylogenetic signals across loci [120]. Overall, the organellar genome patterns in IHEs are largely consistent with maternal inheritance, particularly based on plastome evidence. However, the mitogenomes in some IHEs show elevated similarity to multiple parental species, suggesting more complex inheritance patterns (Fig. 6A). Minor heteroplasmy or rare paternal leakage events cannot be excluded given the limitations of genome-scale data. Future studies incorporating deeper sequencing or experimental crosses will be necessary to fully resolve the contributions of these processes and to better understand the ecological and evolutionary drivers of hybridization in Peganum. Conclusions We successfully assembled the mitochondrial genomes of P. multisectum and P. nigellastrum using Illumina, HiFi, and Nanopore sequencing platforms, filling a gap in mitochondrial genome data for Peganum in ACA. We also combined the previously published P. harmala mitogenome with the newly assembled plastomes of Peganum to conduct a comprehensive study of organellar genomes. The structures, collinearity, and gene content of Peganum plastomes were nearly identical, whereas significant differences existed in the mitogenomes, with numerous DNA sequence rearrangements. Comparative analysis of the organellar genomes revealed consistent results in terms of repetitive sequences, codon usage biases, and RNA editing. Integrating selective pressure results from both organellar genomes, we identified nad2, rpoC1, and atpF as genes showing changes in selective constraints. DNA transfer analysis supported genetic exchange among the plastome, mitogenome, and nuclear genome, although most sequences transferred to the mitogenome have likely evolved into pseudogenes. Phylogenetic reconstruction based on organellar genomes, together with plastome alignment signals, supported the inferred maternal origins of IHEs. However, phylogenetic discordance within IHE lineages and mitochondrial alignment signals suggest potential complexity in inheritance patterns. These findings provide a framework for further investigation of the evolutionary relationships involved in hybrid speciation in Peganum. Authors’ contributions Faqi Zhang, Xiaofeng Chi, Shilong Chen, and Hao Xu conceived and supervised the study. Hao Xu analyzed the data and wrote the manuscript. Hao Xu, Shuang Han, Yu Niu, and Yun Han collected plant samples. Jingya Yu and Qingqing Zheng provided technical support and theoretical support for this work. Guangyuan Zhang prepared figures and tables. All authors contributed to the experimental design, analysis of the data, and final paper. Supplementary Information Supplementary Material 1. Supplementary Material 2. Authors’ contributions Faqi Zhang, Xiaofeng Chi, Shilong Chen, and Hao Xu conceived and supervised the study. Hao Xu analyzed the data and wrote the manuscript. Hao Xu, Shuang Han, Yu Niu, and Yun Han collected plant samples. Jingya Yu and Qingqing Zheng provided technical support and theoretical support for this work. Guangyuan Zhang prepared figures and tables. All authors contributed to the experimental design, analysis of the data, and final paper. Funding This research was supported by the Qinghai Provincial Science and Technology Major Project (2023-SF-A5), the Xining Science and Technology Major Project (2023-Z-13), the CAS - Qinghai on Sanjiangyuan National Park (LHZX-2022-03), the CAS “Light of West China” Program (2024), the leading talents of the Kunlun talents in Qinghai Province (2024). Data availability The raw sequencing data for HiFi platform have been deposited in the National Genomics Data Center (CRA022121). The complete plastomes and mitogenomes have been deposited in NCBI (https://www.ncbi.nlm.nih.gov/) with accession number PQ859556-PQ859558 and PQ859559-PQ859560, respectively. Declarations Ethics approval and consent to 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 Xiaofeng Chi, Email: xfchi@nwipb.cas.cn. Faqi Zhang, Email: fqzhang@nwipb.cas.cn. References Associated Data Supplementary Materials Supplementary Material 1. Supplementary Material 2. Data Availability Statement The raw sequencing data for HiFi platform have been deposited in the National Genomics Data Center (CRA022121). The complete plastomes and mitogenomes have been deposited in NCBI (https://www.ncbi.nlm.nih.gov/) with accession number PQ859556-PQ859558 and PQ859559-PQ859560, respectively.