Comparative Analysis of Chloroplast Genomes Reveals Molecular Evolution and Phylogenetic Relationships in Fraxinus (Fraxinus mandshurica) Authors: Wenxuan Liu, Jiawei Wu, Haonan Zheng, Guanzheng Qu, Sui Wang Published: 2026 ABSTRACT Fraxinus mandshurica (Manchurian ash) is an ecologically and economically valuable hardwood tree native to Northeast Asia, yet its genomic resources remain limited. We assembled its complete chloroplast (cp) genome (155,559 bp) using hybrid PacBio and Illumina sequencing and performed comparative, phylogenetic, and evolutionary analyses. The cp genome exhibits a typical quadripartite structure encoding 132 gene copies, comprising 114 unique genes (80 protein-coding, 30 tRNA, and 4 rRNA genes), with 18 genes duplicated in the inverted repeat (IR) regions. Simple sequence repeat analysis revealed dominance of mononucleotide A/T repeats. Phylogenetic analysis of 53 complete cp genomes strongly supported the monophyly of Oleaceae and resolved F. mandshurica as sister to the North American F. nigra, consistent with previously proposed Miocene intercontinental dispersal scenarios between East Asia and North America. Most protein-coding genes were under strong purifying selection (Ka/Ks << 1), whereas petB, rpl2, and several ndh genes showed elevated Ka/Ks values that are suggestive of altered selective constraint but are based on very few substitutions and are therefore not, on their own, evidence of positive selection. Nucleotide diversity (Pi) analysis identified 15 hypervariable intergenic spacers (mean Pi = 0.067), among which trnM-CAU-rps14, ndhJ-ndhK, and petL-petG represent promising candidate barcode regions requiring further validation. This study provides a high-quality, fully annotated cp genome of F. mandshurica and a valuable genomic resource for future phylogenetic, population genetic, and conservation studies of this important genus. 1. Introduction Fraxinus (ash) is a genus of approximately 43 species of deciduous or evergreen trees and shrubs distributed across the Northern Hemisphere [1]. Ash species are of outstanding ecological and economic importance, yet many have been severely threatened in recent decades by invasive pests and pathogens, most notably the emerald ash borer (Agrilus planipennis) in North America [2] and ash dieback (Hymenoscyphus fraxineus) in Europe [3]. Among them, Fraxinus mandshurica Rupr. (Manchurian ash) is one of the most cold-tolerant species of the genus and reaches the northernmost limit of ash distribution in northeastern Asia. In Northeast China, it is one of the “Three Hardwood” species (F. mandshurica, Juglans mandshurica, and Phellodendron amurense) and is economically significant for its hard, beautifully textured timber used in furniture, flooring and building, as well as for its bark (“qinpi”, Cortex fraxini), which is used as a traditional Chinese medicine containing compounds with immunosuppressive activity [4]. Because of long-term overexploitation and habitat loss, natural populations of F. mandshurica have declined sharply, and the species is now listed as a national second-class protected wild plant in China [5], making the characterization of its genetic resources an urgent priority for conservation and sustainable utilization. In recent years, genomic resources in Fraxinus have grown rapidly. A chromosome-scale genome is available for European ash (F. excelsior) [6], and chromosome-scale genome assemblies have been reported for several other species [7]. At the plastome level, complete chloroplast genomes have been assembled for multiple Fraxinus species, and comparative analyses have improved species identification and phylogenetic resolution within the genus [8]. However, plastid genomic information for F. mandshurica remains extremely limited: no fully annotated chloroplast genome has been reported for this species, and most existing studies focus on physiology, biochemistry, silviculture and tissue culture [9]. This gap constrains several research directions of direct relevance: (i) the molecular authentication of “qinpi” against adulterants using plastid barcodes; (ii) phylogeographic and population genetic studies of a species occupying the cold margin of the genus range, where chloroplast markers with predominantly maternal inheritance can trace glacial refugia and post-glacial expansion; and (iii) conservation genomics of a threatened timber species. Moreover, as a native Asian host that has co-evolved with H. fraxineus, F. mandshurica shows greater tolerance to ash dieback than European species [10], so its genetic resources are also of interest for breeding ash dieback resistance. Chloroplast genomes are typically 120–160 kb in length with a conserved quadripartite structure, in which two inverted repeats (IRs) separate the large and small single-copy regions (LSC and SSC), and contain approximately 100–130 genes [11,12]. Owing to their predominantly maternal inheritance [13], moderate substitution rates, and the presence of both conserved and variable regions, chloroplast genomes are powerful and cost-effective markers for species identification [14], phylogenetic reconstruction and population genetics [8,11]. For a species such as F. mandshurica that still lacks extensive nuclear genomic infrastructure, the chloroplast genome offers an immediately accessible first step toward understanding its evolutionary history and genetic resources. In this study, we assembled and annotated the complete chloroplast genome of F. mandshurica from whole-genome sequencing data, and characterized its genome structure, gene content, repetitive sequences and codon usage. We further reconstructed the phylogenetic position of F. mandshurica within Oleaceae based on complete plastome sequences. These results provide a fundamental genomic resource for species authentication, population genetics and conservation of F. mandshurica, and contribute to a more complete picture of plastome evolution in Fraxinus. 2. Results 2.1. Chloroplast Genome Assembly The complete chloroplast genome of F. mandshurica was assembled as a single circular molecule of 155,559 bp, with an overall GC content of 37.82%. The genome exhibits the typical quadripartite structure of angiosperm cp genomes, comprising a large single-copy (LSC) region of 86,429 bp, a small single-copy (SSC) region of 17,850 bp, and two inverted repeat (IR) copies of 25,640 bp each (IRa and IRb). The two IR copies were 100% identical inverted repeats, and the IR boundaries were localized at positions 86,418/112,058 (IRa) and 129,908/155,548 (IRb). Validation of the assembly with both short- and long-read data confirmed its high quality. Re-mapping of the Illumina reads (bwa mem) placed 9,169,356 of 579,112,560 reads (1.58%) onto the chloroplast genome, corresponding to a mean coverage of 7043× and complete (100%) breadth of coverage across the entire genome. Sequencing depth was uniform across the LSC, SSC, and IR regions (mean depth 6885.20×, 5594.30× and 7813.70×, respectively), with the IR copies showing elevated depth relative to the single-copy regions (approximately 1.1–1.4×), as expected for duplicated regions. All four IR/SC junctions were independently supported by PacBio long reads (340–6466 uniquely mapped reads per junction), and 1890 PacBio reads spanned the linearization point, providing direct evidence for the circularity of the assembled molecule. The complete chloroplast genome sequence of F. mandshurica has a total length of 155,559 base pairs (bp). Like most land plants, the circular cpDNA has a typical quadripartite structure. A large single-copy (LSC) region of 86,429 bp separated from a 17,850 bp small single-copy (SSC) region by two inverted repeats (IRs), each of 25,640 bp. The GC content of the whole genome is 37.82%, and the GC contents of the LSC, SSC and IR regions are 35.81%, 31.96% and 43.23%, respectively. Obviously, the IRs are relatively GC-rich regions (Figure 1). 2.2. Chloroplast Genome Annotation We annotate a total of 132 gene copies in the F. mandshurica cp genome, of which 18 are duplicated in the IR regions and therefore correspond to 114 unique genes. Among these 114 unique genes, there are 80 protein-coding genes, 30 tRNA genes, and 4 rRNA genes. The GC content of different types of genes varied greatly; it is higher in the rRNAs (55.28%) and tRNAs (52.80%) than in the protein-coding genes (37.82%). Although most of these genes are single-exon genes, 15 genes (6 tRNA, 9 mRNA) contain 2 exons and 3 mRNA genes contain 3 exons (Figure 1). Among the F. mandshurica cp genes, several are unusual with special structure. The rps12 gene which codes for the homologous ribosomal protein S12 is a trans-spliced gene [15], three exons of which are located in the LSC region and IR regions, respectively. Unlike ordinary genes using ATG as their initiation codons, there are 3 F. mandshurica cp genes that have unusual initiation codons, such as ACG in ndhD and GTG in rps19, ycf15; these genes code for a variety of proteins involved in photosynthesis, transcription and translation. Table 1 shows the gene functions and groups in the F. mandshurica cp genome. 2.3. Codon Usage Bias and RSCU Value Statistics Codon usage bias refers to differences in the frequency of occurrence of synonymous codons in coding DNA. Although codon degeneracy manifests as the multiplicity of three-base-pair codon combinations that specify an amino acid, synonymous codon usage frequency is always different, which can be measured by RSCU value. To some extent, this is influenced by mutational bias coupled with both natural selection for translational optimization and genetic drift [16]. There were 80 unique protein-coding genes in the F. mandshurica cp genome, including 26,485 codons in total. Genes duplicated in the IR regions were retained in the codon-usage analysis because they are actively transcribed and therefore contribute to the total codon pool, consistent with previous codon-usage surveys of Oleaceae plastomes. The three amino acids with the highest proportions were leucine (10.52%), isoleucine (8.48%) and serine (7.83%), while the lowest proportions were observed for cysteine (1.10%), tryptophan (1.77%) and the stop codons (0.33%). Actually, this codon usage bias is highly conserved among Oleaceae species [17]. Table 2 showed the RSCU value in F. mandshurica and some other species cp genes. Compared with the distantly related Nicotiana tabacum, the value of F. mandshurica is closer to that of Forsythia suspensa, but the differences were not obvious, especially the trend was completely consistent (greater than 1.00, less than 1.00 or equal to 1.00). Here we used RNA sequences without editing to compute the codon usage and RSCU values; because RNA editing was not incorporated, the values at edited sites may differ slightly from the protein-level codon usage. 2.4. Comparison of Chloroplast Genome Sequences with Those of Other Species Despite high conservation of cp genome sequences, especially among closely related species, there are also some hotspots of variation on these sequences which can be used for species classification, phylogenetic analysis and domestication studies. In order to visualize the similarities and differences in the cp genome sequences among Fraxinus spp., we selected the F. mandshurica cp genome sequence we assembled and 5 other Fraxinus spp. cp genome sequences downloaded from NCBI organelle genome database for sequence alignment. These species are each from 6 of the 7 sections of the genus Fraxinus (the other section has no complete cp genome sequences in the NCBI database), which can represent the variation regions among different sections. The mVISTA plot was used to compare these Fraxinus spp. cp genome sequences with the gene annotation of F. mandshurica as a reference (Figure 2). Overall, these Fraxinus spp. cp genomes had little difference in size, ranging from 155,541 bp to 155,694 bp. Figure 2 shows that the structures of all the six Fraxinus spp. cp genomes are conserved and no obvious translocations or inversions are detected. Similar to previous studies, these sequences have a high level of overall similarity, especially in the coding region, but some hotspots of variation also exist on the sequences. Generally speaking, the most highly polymorphic regions were located in the intergenic regions (such as rpl2—trnH-GUG, rps16—trnQ-UUG, trnS-GCU—trnG-UCC, trnE-UUC—trnT-GGU, psbE—petL, rpl32—trnL-UAG). However, the ycf1 gene which had higher variability regions is an exception. 2.5. SSR Identification and Analysis SSRs identification is another standard analysis for cp genome analysis. Due to the existence of interspecific variation, this sequence is very useful in species classification, genome evolution and rearrangement [18,19,20]. In this study, we only focused on species representative from different sections of the Fraxinus genus. A total of 53, 43, 45, 50, 50 and 55 SSRs were detected in the F. americana, F. chiisanensis, F. chinensis, F. mandshurica, F. quadrangulata and F. xanthoxyloides cp genomes, respectively (Figure 3). Among the SSRs, most of them were mononucleotide repeats and few were dinucleotide repeats. No tri-, tetra-, penta- and hexanucleotides repeats were found in all the six Fraxinus spp. cp genomes. For F. mandshurica, considering sequence complementary, the mononucleotides were composed of 46 A/T and 1 C/G, and the dinucleotides were composed of 1 AG/CT and 2 AT/AT. To develop lineage-specific markers, CandiSSR was used to identify candidate polymorphic SSRs. Unfortunately, no candidate SSRs were identified among the Fraxinus species. Because only one plastome per species was examined, this result primarily reflects interspecific variation and does not directly evaluate population-level marker utility. 2.6. IR Expansion and Contraction Size variation in cp genomes is partly a result of expansion and contraction at the borders of the IR regions. The differences in IRs also reflect phylogenetic history. Here, we selected four representative species in Oleaceae and a detailed comparison of four junctions, LSC/IRA (JLA), LSC/IRB (JLB), SSC/IRA (JSA) and SSC/IRB (JSB), between the two IRs (IRA and IRB) and the two single–copy regions (LSC and SSC), was completed among Fraxinus mandshurica, Olea europaea, Schrebera alata, and Syringa oblata (Figure 4). The SSC/IRA junction is located in the ycf1 region in all Oleaceae species chloroplast genomes and extends a different length into the SSC region in all genomes; the IRB region includes 1090 bp (F. mandshurica), 1091 bp (O. europaea), 1091 bp (S. alata), and 1089 bp (S. oblata) of the ycf1 gene, respectively. Similarly, the trnH gene is located in the LSC region, 60, 64, 227, and 58 bp away from the IRA/LSC border in the four Oleaceae chloroplast genomes, respectively. The ψycf1 pseudogene was present in all four genomes: in S. oblata it extended several base pairs into the SSC region, whereas in the other three species it was entirely located within the IRb region. The distance between the ψycf1 pseudogene and the ndhF gene was similar in the four genomes (about 23 bp). The rps19 gene is located in the LSC region 44, 46, 46, and 43 bp away from the LSC-IRa border in these four Oleaceae chloroplast genomes, respectively. The rpl2 gene is located in the IR regions, and the IRa and IRb regions each contain 16 bp of the rpl2 gene. 2.7. Phylogenetic Relationships The maximum likelihood phylogenetic tree reconstructed from complete chloroplast genome sequences clearly demonstrates the evolutionary topology of the family Oleaceae and its related taxa (Figure 5). To rigorously root the tree, distantly related species including Solanum lycopersicum, Nicotiana tabacum, Arabidopsis thaliana, as well as Apiales species Eleutherococcus senticosus (a medicinal plant) and Panax ginseng, were selected as outgroups. This strategic choice not only stabilizes the topological direction at the root but also provides a macroscopic baseline for future comparisons of secondary metabolite evolution in medicinal plants from a broader angiosperm perspective (encompassing orders such as Apiales and Lamiales). The results show that all Oleaceae species cluster into a highly supported monophyletic group. The phylogenetic topology recovered in this study is broadly congruent with previous molecular phylogenetic frameworks of Oleaceae, although some differences are evident in the relationships among major lineages. Within the taxa sampled here, Jasminum fluminense was recovered as the earliest-diverging lineage, followed by a strongly supported clade comprising Schrebera alata and Comoranthus minor. The close relationship between Schrebera and Comoranthus is consistent with their placement within Schreberinae in previous molecular classifications of Oleaceae [21]. Olea europaea diverged subsequently, whereas the sampled species of Fraxinus and Syringa each formed well-supported monophyletic groups and were recovered as sister lineages in the present plastome-based analysis. Previous phylogenetic studies incorporating evidence from both nuclear and organellar genomes have generally supported the major tribal and subtribal framework of Oleaceae, while also revealing some discordance among genomic datasets in the resolution of deeper relationships [22]. Therefore, rather than indicating complete agreement with all previously proposed topologies, our results provide additional plastome-scale evidence for the relationships among the sampled Oleaceae taxa and demonstrate the strong phylogenetic signal of complete chloroplast genomes for resolving relationships within this taxonomic sampling. Within the genus Fraxinus, our finding that Fraxinus mandshurica and the North American endemic species Fraxinus nigra form a tight sister group carries profound biogeographical significance. They further form a highly stable evolutionary branch with other species of the section Fraxinus, including Fraxinus sogdiana, Fraxinus angustifolia, and Fraxinus excelsior. This specific grouping precisely reflects the traditional taxonomic classification of the section Fraxinus. The close sister relationship between F. mandshurica (East Asia) and F. nigra (North America) is highly consistent with the classic study by Hinsinger and colleagues [23] on the phylogeny and biogeographic history of Fraxinus who noted that both species belong to section Fraxinus and confirmed their close relationship at the molecular level. This East Asian–North American sister relationship is consistent with previously proposed Miocene intercontinental dispersal and subsequent vicariance scenarios between Eurasia and North America; however, divergence-time estimation and ancestral-area reconstruction would be required to test this hypothesis directly. The present phylogenetic tree recovers this intercontinental phylogenetic link, underscoring the value of complete chloroplast genomes for resolving both deep and shallow evolutionary relationships within Oleaceae. 2.8. Selection Pressure and Evolutionary Dynamics To assess the selective constraints acting on protein-coding genes of the F. mandshurica cp genome, we computed pairwise Ka/Ks ratios against three closely related Fraxinus species (F. nigra, F. excelsior, and F. americana). The analysis included 80 unique protein-coding genes, of which 14 to 20 yielded valid Ka/Ks values per species comparison (Figure 6, Table 3), after genes lacking synonymous substitutions (Ks = 0) or without a reliable single-copy orthologous alignment were excluded. Across all three pairwise comparisons, the vast majority of chloroplast genes exhibited Ka/Ks values considerably below 1, indicating strong purifying selection and high functional constraint, consistent with the overall conservation of cp genomes in Oleaceae. Photosystem-associated genes (e.g., psaA, psbA, psbB, psbI), most ribosomal protein genes (e.g., rps2, rpl20, rps14), and components of the cytochrome b6/f complex (e.g., petN, petG) showed Ka/Ks ratios close to zero, reflecting their essential roles in photosynthesis and translation. Notably, several genes displayed elevated Ka/Ks ratios suggestive of relaxed or potential diversifying selection (Table 3). (i) petB (cytochrome b6/f complex subunit) and rpl2 (ribosomal large subunit protein L2) showed high Ka/Ks estimates in the F. excelsior comparison (2.15 and 1.17, respectively). However, because these estimates were based on only one valid species comparison, they were treated as exploratory observations rather than robust evidence of altered selection. (ii) rps16 (ribosomal small subunit S16), with Ka/Ks = 0.92–1.04 in both F. excelsior and F. americana; (iii) clpP (ATP-dependent Clp protease proteolytic subunit), with Ka/Ks = 1.04 in F. excelsior and 0.91 in F. americana; (iv) ndhA and ndhB (NADH dehydrogenase subunits), which exhibited consistently elevated Ka/Ks across species pairs (0.66–0.85 for ndhB in all three comparisons; 0.82–0.89 for ndhA in two comparisons); and (v) ycf1 (hypothetical chloroplast ORF 1), with Ka/Ks = 0.62–0.78. The recurrent detection of elevated Ka/Ks in the ndh genes (ndhB in all three pairwise comparisons and ndhA in two) is particularly noteworthy, as the NDH complex plays a crucial role in chloroplast cyclic electron flow and redox regulation. Although ndhB and ycf1 are below the 0.8 highlight threshold, they are listed in Table 3 for completeness. 2.9. Nucleotide Diversity and Hypervariable Regions To gain further insight into the patterns of sequence variation across the genus Fraxinus and to identify candidate molecular markers for downstream taxonomic and population genetic studies, we quantified nucleotide diversity (Pi) across 57 intergenic spacer regions of the F. mandshurica cp genome using orthologous sequences from 15 Fraxinus species. The genome-wide mean Pi across all analyzed intergenic spacers was 0.067, ranging from 0 to 0.363, indicating substantial heterogeneity in evolutionary rates across the cp genome (Figure 7, Table 4). The number of species contributing to individual intergenic spacers ranged from 4 to 14 (Table 4), and this variation in representation was taken into account when hypervariable regions were ranked. Consistent with the conserved nature of plastome inverted repeat (IR) regions, Pi within both IR copies was extremely low (essentially zero), reflecting strong sequence homogenization between the two IR segments (Figure 7a). In contrast, the LSC region harbored most of the variability, with multiple discrete Pi peaks exceeding 0.15 (Table 4). The 15 intergenic spacers with the highest Pi values were identified, with 14 located in the LSC region and one in the IRb region (Figure 7b). The top hypervariable region was located in the trnM-CAU-rps14 intergenic spacer (Rank 1; position 38,368–38,516 bp; Pi = 0.363), followed by the spacer between ndhJ and ndhK (Rank 2; Pi = 0.302; position 51,123–51,227 bp) and the spacer between petL and petG (Rank 3; Pi = 0.301; position 68,300–68,453 bp). Notably, multiple hypervariable regions are flanked by functionally important genes, including ndhJ, ndhK, petL, petG, psbI, psbJ, rpl33, atpI, rps14, cemA (Table 4), several of which are involved in photosynthesis or electron transport. Because high nucleotide diversity does not by itself demonstrate barcode performance, these hypervariable intergenic spacers are better described as promising candidate barcode regions requiring further validation for species identification and population genetic studies within Fraxinus. In particular, the ndhJ-ndhK (Rank 2, Pi = 0.302), psbI-trnS-GCU (Rank 6, Pi = 0.211), and trnM-CAU-rps14 (Rank 1, Pi = 0.363) spacers all show Pi values comparable to or exceeding those reported for the widely used trnH-psbA barcode region, making them strong candidates for future barcoding applications in Fraxinus. 3. Discussion In this study, we successfully assembled and annotated the complete chloroplast genome of F. mandshurica using a hybrid PacBio-Illumina strategy. Compared with approaches relying solely on short-read sequencing, long-read data enabled unambiguous resolution of the complex inverted repeat boundaries and tandem repeat structures, yielding a gap-free circular genome of 155,559 bp. The quadripartite architecture, gene content (132 gene copies; 114 unique genes, including 80 protein-coding genes), and GC content (37.82%) are highly consistent with other Oleaceae plastomes, confirming the strong structural conservation of chloroplast genomes within the family. The observation that rRNA genes (55.28% GC) are significantly more GC-rich than protein-coding (37.82%) and tRNA (52.80%) genes is also consistent with the known correlation between high GC content and the structural stability of ribosomal RNA molecules. SSR analysis revealed that mononucleotide A/T repeats overwhelmingly dominate the six Fraxinus cp genomes examined, while di-, tri-, and higher-order repeats are rare or absent. This composition is typical of angiosperm plastomes, where the high AT content of non-coding regions promotes poly-A or poly-T stretch formation through replication slippage. The absence of tri- to hexanucleotide repeats across all six species suggests that higher-order SSRs are under stronger selective constraint in the chloroplast genome, likely due to their greater potential to disrupt gene function when located in coding or regulatory regions. Because our comparison is based on a single plastome per species, the absence of candidate polymorphic SSRs among the six Fraxinus species mainly reflects interspecific variation; whether chloroplast SSR markers have discriminatory power at the population or intrasectional level remains to be tested with population-level sampling. Future population genetic studies may benefit from complementing plastid markers with nuclear microsatellites or genome-wide SNP data. The ML phylogenetic tree reconstructed from 53 complete cp genomes provides robust resolution of relationships within Oleaceae, with all major nodes receiving high bootstrap support [24]. The topology with J. fluminense at the most basal position, followed by S. alata and then the Syringa clade is congruent with recent molecular phylogenetic analyses of the family. Within Fraxineae, our data resolve two well-supported lineages in the genus Fraxinus. Critically, the placement of F. mandshurica as the sister species to the North American endemic F. nigra is consistent with previously proposed Miocene Beringian dispersal scenarios, a biogeographic pattern documented in numerous temperate tree genera. Complete plastome sequences provide robust phylogenetic signal; however, they constitute a single, predominantly uniparentally inherited genomic compartment. The phylogenetic framework recovered here should therefore be interpreted as a plastid genealogy rather than a full organismal species tree: chloroplast relationships may not always coincide with the species phylogeny, particularly in groups in which historical introgression or chloroplast capture may have occurred. This caveat is especially relevant to the sister relationship recovered between F. mandshurica and F. nigra, whose congruence with the nuclear phylogenomic signal warrants confirmation in future genome-wide studies. The stable phylogenetic signal from complete plastome sequences suggests that chloroplast phylogenomics is a powerful tool for resolving deep and shallow relationships within Fraxinus [25,26], complementing the limited resolution provided by single-locus markers such as rps16 and trnL-F used in earlier studies. The Ka/Ks analysis revealed that the vast majority of chloroplast genes in F. mandshurica are under strong purifying selection (Ka/Ks << 1), consistent with the essential nature of photosynthetic and translational functions in the chloroplast. However, several genes showed notable deviations from this pattern. Among the single-comparison estimates, petB showed the highest Ka/Ks (2.15 in the F. excelsior comparison), a value exceeding the neutral threshold of 1.0. Because this value derives from a single species comparison, it is more appropriately interpreted as elevated Ka/Ks suggestive of altered selective constraint than as evidence of positive selection. petB encodes a core subunit of the cytochrome b6/f complex, which mediates electron transfer between photosystems II and I. Amino acid changes in this subunit could alter electron transport efficiency, potentially reflecting local adaptation to different light regimes or temperature conditions. The ribosomal protein gene rpl2 likewise displayed an elevated Ka/Ks (1.17 in the F. excelsior comparison); this value is also based on a single species comparison. Among the ndh genes, ndhB showed consistently elevated Ka/Ks across all three pairwise comparisons, whereas ndhA yielded valid estimates in two of them. The NDH complex mediates cyclic electron flow around photosystem I and is critical for maintaining redox balance under environmental stress. The elevated Ka/Ks signal observed in the ndh genes is consistent with, although it does not by itself demonstrate, a role in the adaptation of F. mandshurica to the cold, high-latitude environments of Northeast China, where efficient photoprotection is essential for survival through prolonged winter dormancy; this hypothesis calls for further functional validation. These findings collectively suggest that while the chloroplast genome is globally conserved, specific functional modules—particularly those involved in electron transport and redox regulation—harbor signatures of adaptive evolution that may contribute to ecological differentiation within Fraxinus. The genome-wide Pi analysis revealed a striking dichotomy in sequence variation across the F. mandshurica cp genome. Intergenic spacers in the LSC region were the primary reservoirs of nucleotide diversity, whereas IR regions exhibited essentially zero variation, reflecting the well-known copy-correction mechanism that homogenizes the two IR copies through gene conversion. This pattern of localized hypervariability embedded within a globally conserved genome is consistent with the emerging view that cp genome evolution is characterized by punctuated variation rather than uniform divergence. The identification of the 15 intergenic spacers with the highest Pi values provides a rich resource for marker development in Fraxinus. The top three candidates, trnM-CAU-rps14 (Pi = 0.363), ndhJ-ndhK (Pi = 0.302), and petL-petG (Pi = 0.301), exhibit Pi values comparable to or exceeding those of the widely used trnH-psbA barcode, and their flanking genes are involved in translation initiation (trnM), ribosome structure (rps14), and electron transport (ndh, pet), lending support to the hypothesis that regulatory sequences adjacent to functionally critical genes may experience reduced purifying selection. The high Pi observed in the ndhJ-ndhK spacer (N = 14 species) is particularly promising for species-level barcoding applications, as its broad representation across the genus minimizes the risk of amplification failure in divergent lineages. 4. Materials and Methods 4.1. Plant Materials and Sequencing Tender leaves were collected from an adult F. mandshurica plus tree located in the Experimental Forest Farm of Northeast Forestry University, Harbin, Heilongjiang Province, China (approximately 45°43′ N, 126°38′ E), in June 2022. A voucher specimen was deposited at the Herbarium of Northeast Forestry University, and the species identity was confirmed by Professor Li Kailong. Genomic DNA for PacBio and Illumina sequencing was prepared from these leaves with a modified CTAB-DNA extraction method followed by QIAGEN Genomic-tip 100/G (Qiagen, Hilden, Germany; 10243) cleaning and filtering. A long fragment PacBio library (insert sizes = 20 kbp) and a short fragment paired-end Illumina library (insert sizes = 500 bp, read length = 250 bp) were prepared and sequenced on the PacBio RS-II platform (PacBio, Menlo Park, CA, USA) and the HiSeq 2500 platform (Illumina, San Diego, CA, USA) at Feisha (Wuhan, China). 4.2. Data Filtration To remove adaptor and low-quality reads, raw Illumina data were filtered using fastp tool v0.13.1 (trim reads from 5′ = 10, trim reads from 3′ = 20, qualified quality phred = Q20, unqualified percent limit = 20%) [27]. Clean reads were also checked using FastQC (v.0.11.8) [28] tool to make sure all these satisfied the quality-control requirements. 4.3. Genome Assembly, Annotation, and Visualization To complete the assembly, the chloroplast genome sequence of Fraxinus excelsior (GenBank: MG214254.1) was used as a guide. A total of 66.78 Gb of PacBio subreads (N50 = 11,847 bp) was generated for assembly. All sequence data generated that were derived from PacBio RS-II platform were assembled with Organelle_PBA pipeline [29]. The Illumina reads were mapped onto the initial cp genome of F. mandshurica using BWA [30] software (v. 0.7.17) with mem algorithm. Pilon (v.1.23) was then used to polish the initial assembly [31]. This procedure was repeated until no correction occurred. We used the online Chloroplast Genome Annotation, Visualization, Analysis and GenBank Submission Tool 2 (CPGAVAS2) to generate the draft annotations [32,33]. All tRNA genes were predicted by tRNAscan-SE 2.0 [34]. We finally corrected the draft annotations by using the Apollo Genome Annotation and Curation Tool (v1.11.8) [35] manually. The corrected chloroplast annotation was submitted to CPGAVAS2 for re-analysis. The mean sequencing depth of the chloroplast genome was approximately 7043×, with 1.58% of the total Illumina reads mapping to the final assembly. The IR/LSC and IR/SSC junctions and the circular conformation of the plastome were directly confirmed by PacBio long reads spanning these boundaries. 4.4. Codon Usage Statistics We analyzed the codon usage for all protein-coding genes. Although RNA editing may occur at some sites, we ignored them here. MEGA X [36] software (version 10.1.6) was used to calculate the codon usage and the relative synonymous codon usage (RSCU) value. Because the genetic code of the chloroplast genome differs from the standard code, Plant Plastid was selected as the genetic code in MEGA X. Genes duplicated in the IR regions were retained in the analysis because they are actively transcribed in the plastome and contribute to the overall codon pool; this approach is consistent with previous codon-usage surveys of Oleaceae plastomes. Genes with non-standard initiation codons (ndhD, rps19, and ycf15) were analyzed using the plastid genetic code, in which these alternative start codons are translated as methionine. RNA editing was not incorporated; consequently, the codon-usage and RSCU values are based on unedited RNA sequences, which may slightly affect the values at edited sites. 4.5. Genome Comparison The complete cp genome sequences of F. mandshurica and five other species, F. americana (NC_042449.1), F. chiisanensis (MF980720.1), F. chinensis (MK299391.1), F. quadrangulata (NC_042451.1), F. xanthoxyloides (NC_042452.1) from different sections of the Fraxinus genus, were aligned using the program mVISTA [37]. The SLAGAN algorithm with translated anchoring was used to form VISTA plots. The RankVISTA probability threshold was set at 0.5 to obtain accurate identity and similarity. 4.6. SSR Analysis SSR identification was performed using the Perl script MISA [38] (MIcroSAtellite identification tool) and categorized using standard parameters (the minimum number of repeats of 10, 5, 4, 3, 3 and 3 for mono-, di-, tri-, tetra-, penta- and hexanucleotides, respectively). Meanwhile, CandiSSR [39], an efficient pipeline used for identifying candidate polymorphic SSRs based on multiple assembled sequences, was used to identify PolySSRs and to automatically design primer pairs for each identified PolySSR in F. mandshurica and five other species from the genus Fraxinus above. 4.7. Analysis of IR Expansion and Contraction To examine expansion or contraction of the IR regions for Oleaceae taxa, four species Syringa oblata (MT025818.1), Schrebera alata (NC_042467.1), Fraxinus mandshurica and Olea europaea (GU931818.1) were selected to represent the Subtribes Ligustrinae, Schreberinae, Fraxininae and Oleinae, respectively, depending on the APG (Angiosperm Phylogeny Group) IV system of flowering plant classification. These four taxa were selected as a representative comparison to illustrate the range of IR/SC boundary configurations across the four Oleaceae subtribes (Ligustrinae, Schreberinae, Fraxininae, and Oleinae), rather than to support broad evolutionary conclusions. 4.8. Phylogenetic Analysis The phylogenetic analysis was based on the complete chloroplast genome sequence of F. mandshurica and 52 other dicotyledonous species downloaded from the NCBI database (https://www.ncbi.nlm.nih.gov/; accessed on 15 August 2025). These downloaded sequences comprise representative species from the family Oleaceae, primarily belonging to the genera Fraxinus and Syringa. Furthermore, five distantly related species, namely Arabidopsis thaliana, Panax ginseng, Eleutherococcus senticosus, Solanum lycopersicum, and Nicotiana tabacum, were included and designated as outgroups. The whole chloroplast genome sequences were aligned using MAFFT [40]. To remove poorly aligned regions and spurious sequences, the alignments were subsequently trimmed using trimAl [41]. The program IQ-TREE [42] was employed to automatically find the optimal nucleotide substitution model via ModelFinder [43] and to reconstruct a maximum likelihood phylogenetic tree. Ultrafast bootstrap resampling with 1000 replicates was used to evaluate the branch supports. The best-fit nucleotide substitution model selected by ModelFinder under the Bayesian information criterion was TVM + F + I + R4, and the final trimmed alignment comprised 155,234 bp from 53 chloroplast genomes. 4.9. Selection Pressure Analysis To evaluate the evolutionary selection pressure acting on protein-coding genes of the F. mandshurica chloroplast genome, we calculated the ratio of nonsynonymous to synonymous substitutions (Ka/Ks) using the Nei–Gojobori method as implemented in custom Python scripts (Biopython v1.88 [44]). For each of the 80 unique protein-coding genes identified in the F. mandshurica cp genome, orthologous coding sequences (CDS) were extracted from three closely related Fraxinus species: F. nigra (NC_042450.1), F. excelsior (MG214254.1), and F. americana (NC_042449.1). As chloroplast genomes are highly collinear within the genus Fraxinus, orthologous CDS were localized using the conserved gene coordinates of the F. mandshurica reference genome. Genomic sequences were aligned using the Needleman–Wunsch algorithm at the protein level, and codon alignments were reconstructed accordingly. Synonymous (Ks) and nonsynonymous (Ka) substitution rates were calculated for each orthologous gene pair, with the Jukes–Cantor correction applied to account for multiple substitutions. A gene was excluded from a given comparison (recorded as N/A) when the estimated Ks was zero (no synonymous substitutions, rendering Ka/Ks undefined), when the gene was absent or pseudogenized in the compared species, or when the aligned region contained too few informative sites to reliably estimate substitution rates. Genes were classified as evolving under purifying selection (Ka/Ks < 1), neutral evolution (Ka/Ks ≈ 1), or positive/diversifying selection (Ka/Ks > 1). Pairwise Ka/Ks estimates were interpreted descriptively. Genes with an average Ka/Ks of 0.8 or higher across valid comparisons were highlighted, when supported by at least two species comparisons; estimates based on a single comparison were reported separately with caution. No pairwise Ka/Ks value was considered, by itself, evidence of positive selection. Table 3 reports the genes with interpretable pairwise estimates, including several below the 0.8 threshold; the complete per-comparison estimates are summarized in Figure 6. 4.10. Nucleotide Diversity Analysis To identify hypervariable regions with potential utility as molecular markers or DNA barcodes within the genus Fraxinus, we computed nucleotide diversity (Pi, π) across 57 intergenic spacer regions of the cp genome. Complete cp genome sequences of 15 Fraxinus species (F. americana, F. angustifolia, F. chiisanensis, F. chinensis, F. excelsior, F. insularis, F. lanuginosa, F. latifolia, F. mandshurica, F. nigra, F. pennsylvanica, F. quadrangulata, F. sieboldiana, F. velutina, and F. xanthoxyloides) were retrieved from NCBI. For each intergenic spacer of the F. mandshurica cp genome (as annotated by CPGAVAS2), homologous sequences were extracted from each species using coordinate-based mapping followed by local alignment refinement within ±500 bp of the expected position. Each set of orthologous intergenic sequences was globally aligned using the Needleman–Wunsch algorithm (Biopython v1.88), and nucleotide diversity was calculated per region as the average number of pairwise nucleotide differences per site. The 15 intergenic spacers with the highest Pi values were considered hypervariable. 5. Conclusions In this study, we assembled and characterized a complete, fully annotated chloroplast genome of F. mandshurica through an integrated PacBio-Illumina sequencing approach. The 155,559 bp genome exhibits the conserved quadripartite structure and gene repertoire characteristic of Oleaceae plastomes. Our multi-faceted comparative and evolutionary analyses yielded several key findings: (1) the phylogenetic reconstruction based on 53 complete cp genomes robustly resolved intergeneric relationships within Oleaceae and identified F. mandshurica as the sister species to the North American F. nigra, consistent with previously proposed Miocene Beringian dispersal scenarios for temperate tree taxa; (2) selection pressure analysis revealed that most plastid protein-coding genes are under strong selective constraint, whereas a small subset, including petB, rpl2, and several ndh genes, exhibited comparatively elevated pairwise Ka/Ks estimates; because some estimates were based on very few substitutions or only one valid species comparison, these patterns should be regarded as exploratory evidence of altered selective constraint rather than as evidence of positive selection; (3) genome-wide nucleotide diversity analysis identified 15 hypervariable intergenic spacers, with trnM-CAU-rps14, ndhJ-ndhK, and petL-petG emerging as the most promising candidate barcode regions, which nevertheless require further validation before their practical utility for species identification can be confirmed; and (4) SSR profiling confirmed the dominance of mononucleotide A/T repeats but revealed limited interspecific polymorphism, underscoring the need for complementary nuclear markers in population-level studies. Collectively, this work substantially expands the genomic toolkit available for Fraxinus and provides a foundational resource for future studies on the molecular evolution, phylogenomics, conservation genetics, and sustainable utilization of this ecologically and economically important hardwood genus.