Structural characteristics of the Phoebe hui chloroplast genome and phylogenetic analyses of Phoebe plants Authors: Liangxin Zhang, Jianghong Qian, Yunjie Gu, Hongying Guo, Jian Peng, Minhao Liu, Xiandan Luo, Xueqin Wan, Lianghua Chen, Fang He, Yi Wang, Jinwu Li, Hanbo Yang Published: 2026 ABSTRACT AbstractIn this study, the chloroplast genome of Phoebe hui was sequenced and assembled, and comparative analyses with those of other Phoebe plants were conducted, with a focus on variations in inverted repeat boundary variations, nucleotide diversity, selection pressure, codon usage bias, and phylogenetic relationships. The length of the chloroplast genome of P. hui exhibited a conserved size of 152,831 bp, the GC content was 39%, and the number of genes was 126; all the genes maintained a quadripartite circular structure. Significant interspecific variations in the inverted repeat boundary were observed; ndhF in P. hui resided entirely within the small single-copy region, while the extended ψycf1 of P. zhennan overlapped with ndhF by 637 bp and served as a potential molecular marker for species identification. Noncoding regions showed higher nucleotide diversity, with hypervariable regions identified as valuable tools for population genetics. Ka/Ks analysis revealed strong purifying selection on photosynthesis-related genes (psbA and rbcL), whereas stress-response genes (ndhF and ndhD) exhibited elevated synonymous substitution rates. A preference for A/U-ending codons was noted, although P. hui exhibited unique codon avoidance, suggesting translational optimization. Divergence time estimation indicated that Phoebe diverged from Machilus approximately 42.366 million years ago (Mya), P. hui diverged from its congeners approximately 24.916 Mya, and the most recent divergence occurred between P. bournei and P. chekiangensis at approximately 0.113 Mya. These divergence events may be associated with major geological events and Quaternary climatic fluctuations. This study provides genomic insights into the phylogeny of conserved endangered plants and the development of molecular markers for hybrid breeding.Supplementary InformationThe online version contains supplementary material available at https://doi.org/10.1186/s12864-026-13070-6.Keywords: Chloroplast genome, Phoebe hui, Phylogenetics, Codon usage bias, Selection pressure Introduction Magnoliids, a lineage of angiosperms that emerged after true dicotyledons and monocotyledons, encompass taxa historically regarded as the “earliest angiosperms” [1]. Resolving the phylogenetic positioning of magnoliids offers critical insights into plant evolution and paleoenvironmental reconstructions [2], making their evolutionary relationships a central focus in plant phylogenomics. Previously, this clade has been investigated predominantly utilizing multiomics data, although consensus remains elusive [3]. Through rbcL gene analysis, Qiu et al. [4] identified Ceratophyllaceae as the sister group to all angiosperms, whereas Helmstetter et al. [5] proposed an independent familial status for Hydnoraceae using the 353 nuclear loci of angiosperms. This group comprises Laurales, Magnoliales, Piperales, and Canellales (~ 9,000 plants) [1, 6, 7], with Lauraceae (Laurales) ranking fourth in terms of tree diversity (50 genera, 2500–3000 species) [8]. The taxonomically complex genus Phoebe (avocado tribe, Lauraceae), first delineated by Nees von Esenbeck and subsequently revised [9–12], comprises approximately 100 species, predominantly in the Asian and South American tropics [13]. China hosts 35 Phoebe species (27 endemic), primarily in the Yangtze River Basin and areas to the south [14, 15]. Phoebe hui W. C. Cheng ex Yen C. Yang, also known as xiyenanmu, is a tall tree belonging to the Lauraceae family and Phoebe genus. This tree is native to regions such as Sichuan and Shaanxi Province, China. Wild specimens are frequently encountered in dense forests below an altitude of 1,500 m and are valuable native timber tree plants in China [16]. Valued for its golden-thread wood color and decay-resistant wood with aromatic properties, its wood is used to construct premium furniture and cultural artifacts, including Forbidden City structures where P. hui coexists with Phoebe zhennan S. K. Lee & F. N. Wei, with a minority of Phoebe bournei (Hemsl.) Yen C. Yang and Phoebe chekiangensis C. B. Shang [17]. Medicinally, its extracts demonstrate antitumor potential [18], but wild populations are critically endangered because of overexploitation [19]. The phenotypes of P. hui and P. zhennan are similar, particularly during the seedling stage, making them difficult to differentiate. Persistent taxonomic confusion arises from phenotypic convergence among P. hui, P. zhennan, and P. bournei, particularly in the juvenile stages, and from their sympatric distributions [20]. P. zhennan and P. bournei have been suggested to possibly be the same species [21]. Although traditional classification often conflates these plants, RAD-seq data substantiate P. bournei and P. zhennan as distinct taxa with divergent biogeography: P. bournei predominates in the Fujian-Jiangxi-Wuling regions over the Sichuan-Qionglai distribution of P. zhennan [20, 22]. These taxonomic controversies underscore unresolved phylogenetic relationships within Lauraceae that require enhanced molecular resolution. Current studies on the phylogeny of Magnoliaceae have largely focused on constructing high-level taxonomic frameworks, whereas the phylogenetic relationships among Phoebe (Lauraceae) remain relatively underexplored. Inconsistencies among phylogenetic inferences from various studies highlight the necessity of incorporating more comprehensive genomic data to clarify the evolutionary placement of Magnoliaceae in relation to eudicots and monocots, as well as to solve relationships within Lauraceae [23]. Phoebe is of considerable economic and ecological importance, and its endangered status and challenges in species delimitation emphasize the urgency of elucidating interspecific phylogenetic relationships. Such efforts provide not only a theoretical basis for understanding the origin and evolutionary history of Phoebe but also critical guidance for the conservation and sustainable utilization of its genetic resources. Nevertheless, research on the diversity and phylogenetic relationships of Phoebe chloroplast genomes remains insufficient. Broader comparative and evolutionary genomic analyses across multiple Phoebe species are therefore essential for revealing their phylogenetic patterns. Plant cells contain three sets of genomes: the nuclear, plastid, and mitochondrial genomes [24]. Among these, the chloroplast genome typically exhibits a circular tetraploid structure ranging from 107 to 218 kb, characterized by a small molecular weight, high copy number, and ease of accessibility, which confer unique advantages in plant phylogenetic studies [25]. Its evolutionary rate lies intermediate between that of the nuclear genome (one-third) and the mitochondrial genome (three times), making the chloroplast genome widely applicable for resolving phylogenetic disputes at various taxonomic levels [26]. By revealing intraspecific and interspecific sequences and structural variations, chloroplast genomes provide crucial evidence for understanding plant adaptive evolution, facilitating the identification of closely related species and promoting conservation and breeding efforts [27, 28]. For example, Wang et al. [29] utilized the chloroplast genomes of Chinese Acoraceae to elucidate the phylogenetic pattern of the family. Through comparisons of the chloroplast genomes of Chenopodium and its allied genera, Wei et al. [30] successfully developed multiple efficient molecular markers. By employing sequencing technology, species differences at the DNA level can be precisely obtained and are highly favored in research domains such as phylogenetics [31]. Shi et al. [32] sequenced the chloroplast genomes of P. hunanensis and P. tavoyana and compared them with those of 22 published Phoebe species, inferring that Phoebe originated in Yunnan Province, China, and then disseminated to the southwestern and southeastern regions, providing a basis for the conservation of endangered plants. Li et al. [31] sequenced and comparatively analyzed the chloroplast genomes of P. chekiangensis and P. fortunei and reported that their genome structures were highly conserved but contained highly variable regions, which could be utilized for plant species identification and phylogenetic research. Currently, research related to P. hui focuses primarily on aspects such as community distribution characteristics [33], germplasm resources [20, 34], essential oil extraction [16, 35, 36], and seedling management [37, 38], with genomic research remaining largely unexplored. Key scientific issues such as phylogenetic position and genetic diversity warrant further investigation through chloroplast genome analysis. The phylogenetic placement of certain Phoebe species within Magnoliaceae remains ambiguous. In this study, the de novo assembly and annotation of the chloroplast genome of P. hui were performed, and the structural information of the chloroplast genome was analyzed. Moreover, on the basis of the chloroplast genome information of Phoebe species, the phylogenetic relationships of the genus Phoebe and its phylogenetic status within the Magnoliids were explored, revealing the origin and evolutionary history of P. hui and the genus Phoebe and providing a theoretical basis for the genetic conservation and breeding utilization of P. hui and Phoebe species. Materials and methods Plant material collection, chloroplast DNA extraction, and genome sequencing and assembly Healthy and mature leaves of P. hui were collected from a wild population in Qionglai city, Chengdu, Sichuan Province, China (103.351000°E, 30.371683°N, altitude 522.6 m). The specimen is a large arboreal tree (approximately 15 m in height) with greyish-brown bark, leathery lanceolate to oblong leaves, and short petioles. The species was formally identified by Hanbo Yang (Sichuan Agricultural University) based on morphological characteristics (leaf shape, indumentum, and inflorescence structure) following the Flora of China [13]. A voucher specimen has been deposited in the Forest Ecology and Conservation in the Upper Reaches of the Yangtze River Key Laboratory of Sichuan Province and is maintained by Hanbo Yang (specimen No. SAU-PH-20240943). The collection was conducted in accordance with the Regulations of the People’s Republic of China on the Protection of Wild Plants. No specific collecting permit was required for this study because the research activities were conducted exclusively for non-commercial scientific purposes and involved only minimal leaf sampling without causing any damage to the population or its habitat. All procedures complied with the CITES guidelines for non-commercial scientific research. Fresh leaves were collected, rapidly frozen in liquid nitrogen, brought back to the laboratory and stored at -80 °C for future use (Fig. 1). Chloroplast DNA was extracted using a Super Plant Genomic DNA DP360 Kit (Tiangen Biotech, Beijing, China). The quality of the chloroplast DNA was examined by 0.8% agarose gel electrophoresis, and the concentration was measured using a NanoDrop 2000 (Thermo Fisher Scientific, USA). Schematic and images of P. hui. Note: (A) General morphology of P. hui. B Branches, leaves, and flowers. Slender grayish-brown branches bear dense small conical inflorescences or young fruit clusters at the apex. C Grayish-brown bark with distinct longitudinal fissures and a rough, slightly flaky texture. D Leathery, lanceolate to oblong; apex acuminate, base cuneate, petioles short. Abaxial leaf surface pale, with distinct and regularly arranged lateral veins. E Inflorescences and a single flower. Small white flowers with yellow stamens and five petals, slightly fragrant at anthesis; attractive to bees, ants and other pollinators, typical entomophilous flowers Library construction was performed using the Illumina TruSeq DNA Sample Preparation Kit (Illumina, USA). The constructed libraries were subsequently sequenced on an Illumina Nova-Seq platform, generating paired-end reads with a length of 150 base pairs. Raw sequencing data from the Illumina platform underwent quality control processing using Fastp v 0.36 [39]. FastQC v 0.12.0 (https://www.bioinformatics.babraham.ac.uk/projects/fastqc/) was used to assess the quality of the sequencing reads. Gene annotation was conducted using GeSeq [40]. Genes that were reported as pseudogenes in previous studies were selected for manual examination of their putative protein products. If these gene products contained any premature termination codons (PTCs) that resulted in large fragment deletions, these genes were defined as “pseudogenes”. The annotation results of coding sequences (CDSs) were confirmed through the BLAST program on the NCBI website with the default settings, and the annotation of tRNA-encoding genes was inspected by tRNAscan software [41]. The structure of the P. hui chloroplast genome was visualized using Chloroplot software [42]. Comparative analysis of chloroplast genome structures among species of the genus Phoebe To analyze the variations in chloroplast genomes among the Phoebe species, we compared the structural variations in the chloroplast genomes of P. hui, P. zhennan (NC_036143), P. bournei (MF315088), and P. chekiangensis (NC_034925) [31]. The nucleotide compositions of the chloroplast genomes of the four Phoebe species were characterized and analyzed using BioEdit software [43]. The boundaries at the junctions of the inverted repeats (IRs) and the two single-copy (SC) regions were identified and visualized using the online tool IRscope [44]. Using the mVISTA [45] program and the Shuffle-LAGAN algorithm, with the chloroplast genome of P. hui as the reference, the sequence differences among the chloroplast genomes of the four Phoebe species were visualized. DnaSP v6 [46] was employed to calculate the nucleotide diversity (π) of chloroplast genes and intergenic spacers and to analyze the highly variable regions in the chloroplast. Analysis of gene selection pressure To analyze the differences in gene selection pressure among four Phoebe species (P. hui, P. zhennan, P. bournei, and P. chekiangensis) and three other Lauraceae species (Sassafras randaiense, Litsea veitchiana, and Cryptocarya metcalfiana), Cpstools v 2.6 was used to analyze the homologous protein-coding genes of these plants and investigate the associated selection pressure [47, 48]. Through a comparison of homologous protein-coding genes, the quantity and specific names of the homologous protein-coding genes were obtained, the numbers of nonsynonymous (Ka) and synonymous (Ks) nucleotide substitutions were calculated, and the Ka/Ks ratio was determined. Multiple sequence alignment was conducted on the chloroplast genomes of four Phoebe species using MAFFT. SNP-sites were subsequently applied to identify single nucleotide polymorphism (SNP) loci from the aligned sequences, and the resulting data data were exported and stored as vcf files [49]. Analysis of codon usage bias and aversion indices Codon usage bias (CUB) is a crucial factor influencing gene expression and protein translation and can reflect the evolutionary history and relationships among different species [50]. To ensure the statistical reliability of the analysis of codon usage patterns, circumvent random errors of short sequences, and reflect the complete functional characteristics of genes, 52 protein-coding genes with CDS lengths greater than 300 bp were selected to infer the frequencies of synonymous codon usage and aversion indices in the chloroplasts of P. hui, P. zhennan, P. bournei, and P. chekiangensis. DnaSP v6 was employed to calculate the effective number of codons (ENC), codon bias index (CBI), and relative synonymous codon usage (RSCU) [51]. The optimal codons were estimated by the synonymous codon usage frequency difference method (△RSCU), and the codon aversion motif (CAM) method was used to determine the set of codons that were not present in the chloroplast genome of each Phoebe species, that is, the codon aversion motifs [52–54]. Phylogenetic analysis and divergence time evaluation To infer the phylogenetic status of Phoebe within Magnoliaceae, we performed a phylogenetic analysis based on chloroplast genomes, with representative species sampled from each genus of Lauraceae and Amborella trichopoda as the outgroup [55]. All sequence details are listed in Table S1 [31, 56–71]. The chloroplast sequences were aligned using MAFFT software with the default parameter settings [72]. The maximum likelihood (ML) phylogenetic tree was constructed using IQ-TREE v3.0.1 [73]. The built-in ModelFinder module of the software was used to select the optimal nucleotide substitution model on the basis of the Bayesian information criterion (BIC), so as to identify the evolutionary model best suited to the dataset of this study. The divergence times among Phoebe species were estimated using BEAST v2.7.7 [74]. On the basis of the multiple sequence alignment results obtained from MAFFT, the alignment file was converted into the NEXUS (.nex) format using BioAider v1.727 software [75]. Afterward, this file was imported into BEAUti v2.7.7 [74] to generate an input file compatible with BEAST 2 analysis. A dataset comprising 17 Phoebe species and representative species from each genus of Lauraceae was used to reconstruct the phylogenetic topology, under the maximum likelihood (ML) framework. The calibrated Yule model tree prior and relaxed clock log normal were employed. The counts of Gamma categories and iterations of Markov chain Monte Carlo (MCMC) were set at 4 and 100,000,000, respectively. The log files generated by BEAST 2 were analyzed using the software Tracer v1.7.2 [76]. When the effective sample size (ESS) values were all above 200, the results had converged. Once the results converged, LogCombiner v2.7.7 [74] within the BEAST 2 software package was employed to combine the run tree files, with the burn-in value set at 10%. The trees generated by BEAST 2 were aggregated onto the maximum clade credibility tree with the aid of TreeAnnotator v2.7.7 [74, 77]. The generated phylogenetic tree was refined and visualized using Figtree v1.4.4 (https://tree.bio.ed.ac.uk/software/figtree/) and Adobe Illustrator 2023. Results Characteristics and structure of the plastid genome The sizes of the four complete plastid genomes of Phoebe range from 152,831 bp (P. zhennan and P. hui) to 152,869 bp (P. bournei), featuring a typical quadripartite structure consisting of a large single-copy (LSC) region (93,749 − 93,772 bp) and a small single-copy (SSC) region (18,895 − 18,930 bp), which are separated by two inverted repeat (IR) regions (20,075 − 20,105 bp) (Fig. 2, Table S2). Among the four Phoebe species, P. bournei has the longest plastid genome, while P. chekiangensis has the longest LSC region. The four plastid genomes are highly conserved in terms of GC content and gene number, with a GC content of 39% and a total of 127 genes (126 genes in P. hui), including 81 protein-coding genes (80 genes in P. hui), 8 rRNA genes, and 36 tRNA-coding genes (Table S3). Among them, psbL is lacking in P. hui. Furthermore, all four species contain two pseudogenes. Among the 81 protein-coding genes (80 in P. hui), 75 employ the standard start codon AUG (the cemA gene begins with the alternative start codon GUG in both cases) (78 in P. hui). In P. hui, the rpl2 gene commences with the codon GCG, and the ndhD gene begins with the codon UAU. In the other three Phoebe species, in addition to the cemA gene, the rps19 gene also utilizes the alternative start codon GUG for initiation, the rpoC2 gene starts with the codon GAA, the psbL and rpl2 genes start with the codon ACG, the ndhB gene starts with the codon AGG, and the ndhD gene starts with the codon AGU. P. hui chloroplast genome map. Note: Circular map of the P. hui chloroplast genome. From the inside out, the map consists of three concentric layers: GC content distribution, base sequencing depth, and gene element annotation. Genes transcribed in the forward and reverse directions are shown in the inner and outer rings, respectively. The right legend indicates functional gene categories. This map clearly illustrates the genome structure, gene distribution, and basic characteristics of the P. hui chloroplast genome Comparative structural analysis of four Phoebe species plastid genomes Four species of the genus Phoebe (P. bournei, P. chekiangensis, P. hui, and P. zhennan) present marked differences in the structure of the IR-SC boundary region (Fig. 3). At the IRb-SSC boundary, ψycf1 is present in the IRb region in P. bournei and P. chekiangensis. The ndhF gene of P. bournei crosses the boundary by 30 bp and overlaps ψycf1 by 28 bp; the ndhF gene of P. chekiangensis crosses the boundary by 28 bp and overlaps ψycf1 by 24 bp. The ψycf1 gene of P. zhennan is 46%-61% longer than that of the other species and overlaps ndhF by 637 bp. The ndhF gene of P. hui is completely located in the SSC region and is spaced 27 bp from ψycf1. At the SSC-IRa boundary, the majority of the ycf1 sequences of P. bournei and P. chekiangensis are in the IRa region, whereas those of P. zhennan and P. hui are mainly in the SSC region. At the junction of IRa and LSC, the trnH-GUG genes of P. bournei and P. chekiangensis span the regions, with 23 bp located in the IRa region; in P. zhennan and P. hui, this gene is located only in the LSC region. Comparisons of the SSC, LSC, IRB and IRA regions in the chloroplast genomes of 17 Phoebe species. Note: The figure illustrates the structural features of the chloroplast genomes from 17 Phoebe species. From left to right, the regions are labeled as the LSC (large single-copy region), IRb (inverted repeat b), SSC (small single-copy region), and IRa (inverted repeat a), with their lengths labeled and boundaries highlighted in gray. Additionally, key genes and their boundary characteristics in selected areas are concurrently displayed The sequence discrepancies among the four Phoebe plastid genomes are depicted as follows. The plastid genome structures of the four Phoebe species are essentially identical and highly conserved (Fig. 4). According to the results of the nucleotide diversity analysis, multiple plastid loci exhibited varying degrees of sequence variation, with nucleotide diversity (π) ranging from 0.00000 to 0.04167. Among them, protein-coding genes such as rpl23 and several intergenic spacers (such as trnK-UUU-2-matK, trnG-UCC-2-trnR-UCU, trnY-GUA-trnE-UUC, and ccsA-ndhD) all demonstrated relatively low levels of variation. Notably, the trnH-GUG gene itself presented a certain degree of sequence variation (π = 0.20000), while the spacer sequence between it and the adjacent gene psbA was highly conserved (π = 0.00000), indicating extremely low variation (Fig. 5). Relatively speaking, the variations in intergenic spacers among different species were more pronounced, which agrees with the feature that intergenic regions in plastid genomes typically have greater variation. The sequence similarities of the chloroplast genomes of the four Phoebe species were compared using mVISTA, with P. hui used as the reference. Note: The figure has four rows, each representing different Phoebe species, with P. hui used as the reference genome. Horizontally, the figure displays linear genomic sequences and annotates gene names (e.g., matK and rpoC2). Color blocks, which are based on the legend (e.g., gene and exon), indicate the sequence conservation of elements such as genes, exons, conserved noncoding sequences (CNS), and mRNA. The percentages on the right represent the similarity threshold; higher values indicate greater sequence similarity in the corresponding regions among the species Nucleotide diversity (π) of chloroplast gene loci in four Phoebe species. Note: (A) Distribution of π values for intergenic spacer regions. B Distribution of π values for genes. The dashed line indicates the threshold of π = 0.070 Analysis of genes under positive selection To identify the genes that might be under positive selection, the Ka/Ks ratios of 74 plastid protein-coding genes were analyzed. Most of the values were less than or equal to 1. Some genes, such as rbcL, ndhA, rpl16, ycf1, psaB, matK, cemA, and rps2, presented ratios greater than 1 in specific pairwise comparisons (such as P. hui and L. veitchiana, P. hui and S. randaiense, P. hui and P. bournei, and P. hui and P. chekiangensis), suggesting that these genes might have experienced positive selection in the corresponding pairwise comparisons. The ratios of the other genes, such as accD, were less than 1 in all the related pairwise comparisons except for P. hui and S. randaiense, where the ratio was 0.651001. The ratio of ndhA was 0.000000 in the pairwise comparisons of P. hui and P. bournei, P. hui and P. chekiangensis, and P. hui and P. zhennan. The majority of genes maintained a ratio of less than or equal to 1 in pairwise comparisons of all the species (Fig. 6). Heatmap of the Ka/Ks ratios among 74 plastid-encoded protein genes for species comparisons. Note: Heatmap illustrating the Ka/Ks ratios of 74 plastid protein-coding genes among seven species: P. hui (Phh), P. bournei (Phb), P. chekiangensis (Phc), P. zhennan (Phz), C. metcalfiana (Cr), L. veitchiana (Li), and S. randaiense (Sa). The color gradient (0–2.0) is shown in the right legend. Genes are plotted on the x-axis, and interspecific pairwise comparisons are plotted on the y-axis. Color intensity indicates variation in Ka/Ks ratios among genes across species, reflecting divergent selection pressures Codon usage patterns and aversion Analysis of the codon usage characteristics in four Phoebe species (P. bournei, P. chekiangensis, P. hui, and P. zhennan) revealed that their CDS lengths ranged from 70,359 bp to 70,688 bp, with the total number of codons varying between 23,453 and 23,560. Leucine (Leu) is the most abundant amino acid (with 4,733-5,066 codons), while cysteine (Cys) is the least abundant (1,155-1,250 codons). The most frequently employed codon is UUU (encoding phenylalanine), and less commonly used codons, such as UGC (encoding cysteine), appear 439 times in both P. hui and P. zhennan (Table S4). The synonymous codons of amino acids other than methionine (AUG) and tryptophan (UGG) exhibit distinct usage preferences. Codons with an RSCU > 1 (e.g., UCU for serine and UAU for tyrosine) were used more frequently than other synonymous codons and mostly ended with A or U, whereas those with an RSCU < 1 tended to end with G or C (e.g., UCG, CGC, and GCG) (Table S5). This codon usage pattern demonstrates a high level of consistency among the four Phoebe species and is similar to the codon preference patterns of chloroplast genomes in other plant groups. Further analysis of codon usage bias (CUB) revealed that CUB in the four Phoebe species exhibited both conservation and species specificity: P. bournei and P. zhennan shared high similarity in the extreme ENC (effective number of codons) groups (high and low ENC values), e.g., both contained psbA (41.96) and rps14 (36.35) in the low-ENC group and ycf4 (58.74) and cemA (57.14) in the high-ENC group, whereas P. hui and P. chekiangensis also had consistent ENC distributions (e.g., both included ycf3 (54.9) and clpP (55.93) in the high-ENC group). However, P. hui exhibited specificity: the ENC value of ndhD (49.62) was significantly lower than those of the other three species (49.73–50.21), and only P. hui had an ENC value of rpl2 (55.72) that deviated from the conservative range (55.7-55.92) of the others, which might be related to selection pressure or functional adaptability adjustments during its evolution (Table S6). Additionally, 18 optimal codons, such as GCT, TGT, and GAT, were identified in the chloroplast genomes of the four Phoebe species, with largely consistent optimal codon usage. Codon aversion motifs (CAMs) have been utilized to investigate plant phylogeny and evolution. Among the 52 chloroplast protein-coding genes, 50 (e.g., accD, atpA, and atpB) had completely consistent codon aversion motifs (CAM) across all four species; the other three species shared the same CAM patterns except P. hui, which additionally avoided the AGG codon in ndhB and GGT in ycf3, reflecting the close phylogenetic relationship but subtle differentiation between P. hui and the other three Phoebe species (Table S7, Fig. 7). Codon aversion motifs of ycf3 and ndhB in four Phoebe species. Note: The figure illustrates the codon avoidance patterns for four Phoebe species in the ycf3 and ndhB genes. The vertical axis is divided into two sections: the upper for ycf3 and the lower for ndhB, with specific codons (e.g., AGC, CGC, and CGT) on the horizontal axis. Different colored dots indicate the codons avoided by each species in the respective gene (i.e., codon avoidance motifs): red, purple, yellow, and cyan represent the four species in ycf3, whereas pink, green, orange, and blue represent them in ndhB. It allows visual comparison of codon usage preferences among the four species in terms of these two genes, with species easily distinguished by color. The plant abbreviations are as follows: Phh: P. hui; Phb: P. bournei; Phc: P. chekiangensis; and Phz: P. zhennan Phylogenetic relationships and divergence times among Phoebe species To elucidate the phylogenetic position of P. hui within Lauraceae and Magnoliaceae, a phylogenetic tree encompassing 49 species was constructed. The results showed that Lauraceae species do not form a simple bifurcating pattern but instead exhibit multiple major clades with successive divergence, with most genera forming well-supported monophyletic groups that reflect a clear genus-level phylogenetic structure. Basal lineages such as Cassytha diverged early, while the remaining taxa were further resolved into several major lineages, including those represented by Cryptocarya, the Litsea-Lindera-Laurus-Sassafras group, and the clade comprising Cinnamomum, Phoebe, Machilus, and Persea. Within this framework, Phoebe formed a well-supported monophyletic clade closely related to Machilus. Divergence time estimates revealed that the genus Phoebe diverged from Machilus approximately 42.366 million years ago (95% HPD: 25.933–60.951 Mya). Additionally, P. zhennan split from other Phoebe species within the same clade approximately 6.436 million years ago (95% HPD: 1.929–12.595 Mya), and P. hui diverged from other Phoebe species approximately 24.916 million years ago (95% HPD: 13.552–38.218 Mya). This timing may correspond with the uplift of the Qinghai-Tibet Plateau, but further geological and climate data are needed to verify this association. The most recent divergence occurred between P. bournei and P. chekiangensis approximately 0.113 million years ago (95% HPD: 0.00001–0.4092 Mya) (Fig. 8). Phylogenetic relationships and divergence time estimates of Lauraceae species based on chloroplast genomes. Note: The figure illustrates the phylogenetic relationships and divergence time estimates of representative species of Lauraceae based on their chloroplast genomes. The dataset includes multiple genera within Lauraceae, with each genus represented by one or two species, with particular emphasis on the genus Phoebe. Amborella trichopoda was used as the outgroup [55]. Node colors (green, orange, and red) indicate the 95% highest posterior density (HPD) divergence time ranges (in millions of years ago, Mya), illustrating the evolutionary relationships and divergence sequence among lineages. Chloroplast genome sequences of all sampled species were used for phylogenetic reconstruction. Detailed information for all taxa is provided in Table S1 [31, 55–71] Discussion This study presents a comparative analysis of the chloroplast genomes of P. hui and its three closely related Phoebe species, revealing both conservation and species-specific variations in their genomic structures while providing new insights into their phylogenetic relationships and evolutionary history. Overall, the chloroplast genomes of the four species were highly conserved in size (152,831 − 152,869 bp), GC content (39%), and gene number (126–127), all of which maintained the typical quadripartite circular structure, reflecting the stability of angiosperm chloroplast genomes [31, 32]. However, significant interspecific differences were observed in the IR-SC boundary regions; for example, the ndhF gene of P. hui was entirely located in the SSC region, whereas the pseudogene ψycf1 in P. zhennan showed abnormal extension and extensive overlap with ndhF. Such boundary variations might be driven by genomic recombination or replication slippage mutations [44] and could serve as molecular markers for differentiating morphologically similar species (P. hui and P. zhennan), especially when identification challenges during the seedling stage and in folk taxonomy are addressed [17, 20]. Additionally, nucleotide diversity analysis revealed that variations in coding regions were generally lower than those in noncoding regions, which is consistent with the evolutionary rule of conservation in coding regions and dynamics in noncoding regions for chloroplast genomes [26]. The screened hypervariable regions can provide new tools for population genetic and phylogeographic studies of Phoebe [32]. Ka/Ks analysis further revealed that core photosynthesis genes (e.g., psbA and rbcL) were under strong purifying selection to maintain functional stability. In contrast, stress-response genes (e.g., ndhF and ndhD) achieved adaptive evolution through high synonymous substitution rates (Ks) [44]. Notably, the high Ks value of ψycf1 in P. zhennan also suggested that its pseudogenization might be related to relaxed selection [20]. Codon usage bias (CUB) analysis revealed that all four species prefer codons ending with A/U (e.g., UUU or UUA) [32]; however, the accD and atpA genes of P. hui exhibit a unique codon avoidance pattern, particularly in the accD and atpA genes, which may be linked to translational optimization or tRNA adaptability. Further experimental verification is needed to confirm these potential functional implications. Consistent with previous studies on chloroplast codon usage patterns [78], the conserved codon usage bias and species-specific codon aversion motifs (CAMs) in four Phoebe species were highly consistent with their phylogenetic relationships, indicating that chloroplast codon characteristics can be used as reliable auxiliary molecular markers for exploring interspecific differentiation and microevolution of Phoebe. Phylogenetic analysis based on chloroplast genome data from 49 species across Lauraceae and Magnoliaceae was conducted to elucidate the phylogenetic position of P. hui [68]. The results showed that Lauraceae does not exhibit a simple bifurcating pattern; instead, it comprises multiple major clades with successive divergence. Basal lineages such as Cassytha diverged early, while the remaining taxa were further resolved into several major lineages, including the Cryptocarya clade, the Litsea-Lindera-Laurus-Sassafras group, and a larger clade comprising Cinnamomum, Phoebe, Machilus, and Persea. Within this framework, Phoebe formed a well-supported monophyletic group that is closely related to Machilus and Cinnamomum. Divergence time estimation revealed that the most recent divergence between P. bournei and P. chekiangensis (~ 0.113 million years ago) may have been driven by Quaternary glacial climate fluctuations [22, 32]. This study supports the status of P. hui as an independent species [22], as its genomic specificity can be used to develop molecular markers to facilitate the identification of endangered species and the regulation of the illegal timber trade [17]. However, given the complexity of species delimitation, further studies incorporating morphological and ecological data are necessary to definitively confirm its distinctiveness. Prioritizing the conservation of populations with genetic uniqueness (e.g., those in Qionglai, Sichuan) is also recommended. We acknowledge several limitations of the study, such as its small sample size. Future research should include more Phoebe plants to refine the phylogenetic framework and integrate transcriptome data to verify the association between codon bias and gene expression. Supplementary Information Supplementary Material 1. Supplementary Material 2. Supplementary Material 3. Acknowledgements The authors thank the staff at the Sichuan Academy of Forestry and Sichuan Agricultural University for their assistance in sample collection and laboratory work. Authors’ contributions Hanbo Yang, Yunjie Gu, and Liangxin Zhang were responsible for research activity planning and execution, collect the data, performed the statistical analysis, and wrote the first draft of the manuscript. Jianghong Qian, Hongying Guo, and Jian Peng prepared the materials, collect the data, and performed the statistical analysis. Minhao Liu, Xiandan Luo, and Xueqin Wan contributed substantially to the conception and design of the study. Lianghua Chen, Fang He, Yi Wang, and Jinwu Li performed the statistical analysis and provided critical revision of the article for the final version. All authors reviewed the manuscript. Funding This work was supported by the National Key Research and Development Program of China (2024YFD2200103), Key Research and Development Project Program of Sichuan Province (2021YFYZ0032), and University Student’s Innovation and Entrepreneurship Training Program Project (202410626053). Data availability Data supporting the results of this study are publicly available in the Genome Sequence Archive (GSA). The project number is CRA030922. The dataset supporting the conclusions of this article is available in the ScienceDB repository, https://www.scidb.cn/detail?dataSetId=f95d154f8d1b405288e795fc9a18bf74. Declarations Ethics approval and consent to participate The authors declare that the collection of samples complies with the rules and regulations of relevant institutional, national guidelines and legislation, and they did not cause damage to the local environment. 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 Yunjie Gu, Email: guyunjie7901@163.com. Hanbo Yang, Email: yanghanbo6@sicau.edu.cn. References Associated Data Supplementary Materials Supplementary Material 1. Supplementary Material 2. Supplementary Material 3. Data Availability Statement Data supporting the results of this study are publicly available in the Genome Sequence Archive (GSA). The project number is CRA030922. The dataset supporting the conclusions of this article is available in the ScienceDB repository, https://www.scidb.cn/detail?dataSetId=f95d154f8d1b405288e795fc9a18bf74.