Genetic diversity analysis and conservation strategy recommendations for ex situ conservation of Cupressus chengiana

Abstract

Background

Cupressus chengiana is mainly distributed in the Hengduan Mountains area in China. It is one of the Class II endangered plants, ex situ conservation is often used to the affected C. chengiana population due to the construction of the power station. However, population fragmentation and inbreeding in the ex situ conservation have led to decline in genetic diversity. It is therefore important to clarify the differences in genetic diversity between native populations and ex situ population.

Results

In this study, we used Genotyping-by-Sequencing to assess the genetic diversity of 30 C. chengiana trees from four populations in the Dadu River Basin, southwest China, including one ex situ conserved population (DK) and three native populations (BW, SA, RJ). The results showed that compared with the native populations, the DK population showed higher genetic diversity. Among the three native populations, SA population may experience inbreeding and has low genetic diversity. The population structure analysis further revealed that the DK population had higher gene flow and lower differentiation than other three populations. For ex situ populations, the primary determinant of genetic diversity is the genetic variation present in the seedlings sourced from natural populations.

Conclusion

These findings support the feasibility of ex situ conservation for C. chengiana conservation. This study provides a scientific foundation for the preservation, management, and restoration of C. chengiana, and would offer valuable insights for the conservation of other endangered plants.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12870-025-06581-z.

Background

The loss of biodiversity has becoming a global crisis and attracting significant international attentions towards conservation efforts [1]. Plant resources are a vital part of biodiversity and ecosystems and serve as an essential foundation and guarantee for the sustainable development of human society and the economy [2, 3]. However, increasing pressures such as habitat destruction, climate change, and human activities have put many plant species in risk of extinction, even led to the emergence of endangered plants like Araucaria nemorosa, Primula veris, and Roussea simplex [4–6]. In this context, ex situ conservation has become an important strategy to protect these endangered species effectively [7].

Ex situ conservation is a strategy to protect plant species by transplanting them from their native habitats to new, suitable environments [8, 9]. This approach not only offers plants a new living space but also serves as an emergency measure to address irreversible changes in their native habitats [10]. To protect these species from extinction, ex situ conservation has been conducted for endangered plants such as Vatica guangxiensis, Dianthus morisianus, and Typha minima Hoppe [10–12]. However, ex situ conservation may disrupt existing genetic structure [13, 14], alter their genetic diversity [15–17] and even a subsequent loss of genetic diversity [18]. Therefore, assessing the genetic relationship between ex situ and native populations is crucial for the success of ex situ conservation efforts [13, 19]. Population genetic analysis is used to understand the level of genetic diversity, which is essential for species conservation and management [20]. Generally, ex situ conservation is considered effective when the genetic diversity of an ex situ population is equal or greater than that of a native population [13, 21]. Yang et al. analyzed the genetic diversity and population structure of six Craigia yunnanensis populations, and concluded that ex situ conservation was effective for C. yunnanensis preservation [22]. Similarly, the conservation strategy for the endangered plant Zabelia tyaihyonii was evaluated based on its genetic diversity and population structure [23]. Currently, genetic diversity is often analyzed using restriction-site associated DNA sequencing (RAD-seq), double digest restriction-site associated DNA sequencing (ddRAD-seq), specific-locus amplified fragment sequencing (SLAF-seq), and genotyping-by-sequencing (GBS).

GBS is one of the most widely used simplified genome sequencing technologies [24, 25], for it bases on next-generation sequencing (NGS), and is particularly effective for analyzing genetic diversity in species with high levels of genetic redundancy [26]. GBS offers high specificity and reproducibility [27], enabling precise estimates of genetic differentiation with smaller sample sizes compared to traditional methods such as nuclear SSRs. This makes GBS an effective tool for evaluating genetic variation and population structure [28–30]. For example, GBS was used to study the evolution and germplasm conservation of Avena sterilis [31], the population genetic diversity of Aniba rosaeodora [32] and the ex situ conservation of Camellia tunghinensis [33].

Cupressus chengiana, a member of the cypress family, is a rare and endangered species in China, classified as a national second-class protected plant [34]. C. chengiana is distributed across three major river basins in China: the Minjiang River Basin and Dadu River Basin in Sichuan Province, and the Bailongjiang River Basin in Gansu Province [34]. Previous studies on its genetic diversity have primarily focused on native populations [34, 35]. While broader assessments of the species’ genetic structure (e.g., across all basins) are critical for comprehensive conservation planning, this study prioritizes populations directly impacted by recent anthropogenic disturbances to guide immediate ex situ conservation actions. C. chengiana is an important species for soil and water conservation in the arid valley areas of the upper Yangtze River, due to its high adaptation and extreme drought-resistance [36]. Thus, it is also one of the pioneering tree species for afforestation in alpine valleys and arid mountainous regions [35]. However, the native C. chengiana populations face severe survival challenges, including poor growing substrates, steep slopes, and anthropogenic disturbances, such as construction projects and indiscriminate logging [34].

There are three native populations of C. chengiana along the Dadu River in Sichuan province: Songgang (SA), Rezu (RJ), and Baiwan (BW). The construction of the power station destroyed their natural habitat, so ex situ conservation of C. chengiana in Dangka (DK) was conducted. The geographic and environmental conditions, such as altitude, temperature and precipitation in DK differ from other three native populations. How the genetic diversity and structure of C. chengiana would change in the ex situ population compared with the native populations? and which factors would influence genetic diversity? To answer these question, in this study, we used GBS to assess the genetic diversity, population structure and genetic differentiation between ex situ and native populations of C. chengiana, and also analyzed the impact of ecological factors on genetic diversity. These findings would provide a valuable reference for the conservation and utilization of C. chengiana, as well as for the ex situ conservation of other endangered plants in the region.

Materials and methods

Plant materials

In 2012, conservation efforts for the endangered C. chengiana were initiated outside its native habitat to DK due to the construction of the power station. Three native populations of C. chengiana (SA, RJ, BW) and one ex situ population (DK) were selected for this study. The details of these four populations are presented in Fig. 1 and Table S1. Sample collection was conducted in the summer of 2019. The plant materials utilized in our study were meticulously identified by our team. This process was primarily facilitated by cross-referencing with the existing specimens housed within the Herbarium of the Chengdu Institute of Biology, Chinese Academy of Sciences. Fresh leaves were collected from the C. chengiana populations at four sites: four plants from SA, RJ and BW, respectively, and 18 plants from DK (transplanted in 2012 at seedling stage from their native habitat to DK). To avoid duplicate genotypes due to proximity, in each site, the distance between sampled tree individuals was more than 50 m. After collection, the samples were immediately snap-frozen in liquid nitrogen and stored at -80 °C for subsequent analysis.

Fig. 1.

Growth (A) and habitat conditions (B) of C. chengiana, and sampling site distribution (C) in this study

DNA extraction, library preparation, and sequencing

Genomic DNA was extracted from C. chengiana leaves using a DNA extraction kit (TIANGEN Biotech, Beijing, China). The integrity of the DNA was analyzed using 2% agarose gel electrophoresis, and the concentration was precisely measured with a Qubit v2.0 fluorometer. The genomic DNA was randomly fragmented into short DNA segments by ApeKI restriction enzyme (TIANGEN Biotech, Beijing, China), followed by end repair. Then, dA-tails were added to both ends of the DNA fragments, and sequencing adapters were ligated. The adapter-ligated DNA fragments were purified using AMPure XP magnetic beads, and fragments in the range of 300–400 bp were selected for PCR amplification. The constructed library was further purified and quality-checked before being sequenced on the HiSeq X10 PE150 platform by Genedenovo Biotechnology Co., Ltd (Guangzhou, China).

Sequence analyses, bioinformatics, and SNP identification

The raw sequencing data were first converted into raw reads through base calling and then filtered. Initially, reads containing adapters were removed. Next, reads with an N ratio (indicating undetermined base information) greater than 10% were discarded. Finally, low-quality reads, where more than 50% of the bases had a Qphred score ≤ 10, were also removed. Following these steps, clean reads were obtained [37]. The clean reads were then clustered using the Stacks v1.43 software with the following parameters: -m 6, -M 6, -N 0, -i 1, -H [38]. The clustered tags were overlapped, and for tags without overlap, they were concatenated with N. These tags were subsequently used to create pseudo-genomes of uniform size using a custom script for further analysis.

For variant detection, multiple samples were analyzed using GATK v3.8.1 software [39]. The VariantFiltration tool was then used to filter the variants with the following criteria: variants with a quality score below 4.0, a Fisher’s exact test p-value above 60, a sequencing mapping quality below 40, and a genotype quality below 20 were all filtered out.

Climatic factors

Mean annual temperature (MAT) and mean annual precipitation (MAP) from 2001 to 2020 (Table S1 and Figure S1), were downloaded from the national meteorological information center of China (http://data.cma.cn/).

Analysis of population genetics

Data were transformed and processed using BCFtools v1.20 (https://github.com/samtools/bcftools). The genetic distance matrices were computed using PLINK v1.90, and phylogenetic trees were constructed and visualized with the aid of the ape and phangorn packages in R v4.4.1. The neighbor-joining method was applied, with bootstrap replications conducted 1000 times to enhance the robustness of the tree topology [40]. First, we used the --freq option in VCFtools v0.1.16 (https://github.com/vcftools/vcftools) to extract allele frequencies. Subsequently, we calculated the polymorphism information content (PIC) values using Python v3.13. We employed the --site-pi option in VCFtools v0.1.16 to compute nucleotide diversity at each site. Finally, we used R v4.4.1 to generate the nucleotide distribution plots. Population structure analysis was conducted with admixture v1.3.0 software (https://github.com/topics/admixture) and visualized using Tbtools v2.119 (https://github.com/CJ-Chen/TBtools-II) [41]. The matrices required for genotypic principal component analysis (PCA) were calculated using VCFtools v0.1.16, and the PCA results were visualized using the matplotlib and seaborn libraries in Python v3.13.

Pairwise fixation statistics (Fst) between populations were computed using VCFtools v0.1.16. Genetic diversity indices, including observed heterozygosity (Ho), expected heterozygosity (He), minor allele frequency (MAF), Nei’s gene diversity index (Nei), percentage of polymorphic loci (PPL), Shannon-Wiener index (I), fixation index (FIS), number of polymorphic loci (PLN), and effective number of alleles (Ae), were processed and calculated using VCFtools v0.1.16 and Perl scripts [33].

Results

GBS data analysis and SNP identification

Thirty C. chengiana samples were analyzed using GBS, and total 105,836,876,272 bp of clean data, with average Q20 value of 98.74%, Q30 value of 91.21% and GC content of 35.21% were observed (Table S2). After filtering, the average proportion of clean data obtained was 98.69%. The average clean data for the DK group was 98.78% (3,610,408,194 bp), while it was 98.6% (3,697,964,756 bp), 98.69% (4,140,300,007 bp) and 98.64% (2,374,117,431 bp) for the BW, RJ and SA group, respectively. After clustering by stacks, 1,947,047 tags were obtained, with lengths ranging from 163 to 314 bp. 1,308,445 variants were detected, including 1,259,610 SNPs and 48,835 indels (Table S3). The distribution of variants varied across chromosomes, with chromosomes 9 through 15 showing relatively high counts (each exceeding 120,000). Chromosome 15 had the highest number of variants at 130,982, while chromosome 7 had the lowest, with 49,750 variants.

Phylogeny and genetic diversity analysis

There were variations in genetic diversity among the four populations (Table 1). The Ae values ranged from 1.383 (SA) to 1.479 (DK), with a mean of 1.433 ± 0.035. Ho varied from 0.182 (SA) to 0.272 (RJ), with a mean of 0.236 ± 0.033, while He ranged from 0.223 (SA) to 0.298 (DK), averaging 0.259 ± 0.027. The I values spanned from 0.331 (DK) to 0.460 (SA), with a mean of 0.388 ± 0.047. The Nei value was highest in the DK population (0.308) and lowest in the SA population (0.265), with an overall mean of 0.293 ± 0.017. The FIS ranged from 0.187 (SA) to -0.031 (RJ). The FIS for BW (0.028) and RJ (-0.031) populations were close to zero, while DK and SA exhibited higher FIS of 0.172 and 0.187, respectively. The MAF ranged from 0.214 (DK) to 0.288 (SA), with a mean of 0.264 ± 0.029. The DK population had the highest PLN, Loci, and PPL values (1,229,691; 1,259,610; and 0.976, respectively), while the SA population had the lowest (696,039; 1,179,590; and 0.590, respectively).

Table 1.

Estimates of genetic diversity and population structure statistical within four Cupressus Chengiana populations: ex situ conserved population (DK) and three native populations (BW, SA, RJ)

Site	Ae	Ho	He	Nei	I	FIS	MAF	PLN	Loci	PPL
DK	1.479	0.246	0.298	0.308	0.460	0.172	0.214	1,229,691	1,259,610	0.976
BW	1.423	0.242	0.249	0.291	0.371	0.028	0.278	832,393	1,235,238	0.674
RJ	1.448	0.272	0.264	0.307	0.393	-0.032	0.276	894,268	1,246,343	0.718
SA	1.383	0.182	0.223	0.265	0.331	0.187	0.288	696,039	1,179,590	0.590

Note: Ae, Effective number of alleles; Ho, Observed heterozygosity; He, Expected heterozygosity; Nei, Nei’s gene diversity index; I, Shannon-Wiener index; FIS, Fixation index; MAF, minor allele frequency; PLN, Polymorphic Loci Number; PPL, Percentage of Polymorphic Loci

The DK population exhibited the widest range of PIC values and had a high median, while the other three populations (BW, RJ, and SA) showed more concentrated distributions of PIC values (Fig. 2B). Among the three native populations, the PIC distributions also varied: SA had the broadest range and the lowest median, while BW and RJ had narrower distributions but higher medians. Similarly, the nucleotide diversity in the DK population spanned a wide range and featured multiple peaks (Fig. 2C). In contrast, the nucleotide diversity in the other populations was more concentrated, particularly in the SA population.

Fig. 2.

Phylogeny (A), polymorphism information (B), and nucleotide diversity (C) of four Cupressus chengiana populations: ex situ conserved population (DK) and three native populations (BW, SA, RJ)

Genetic population structure and genetic differentiation analysis of C. chengiana

A phylogenetic tree based on genetic distances was constructed to illustrate the relationships among tree individuals (Fig. 2A). Thirty C. chengiana individuals were categorized into three main branches. The first branch primarily consisted of BW, SA, and five DK samples. The second branch mainly included RJ and six DK samples, while the third branch contained only seven DK samples. The clustering within three native C. chengiana populations (BW, RJ, SA) was relatively strong. However, there were also dispersion among the three native populations. Additionally, samples among all populations were mixed distribution in the phylogenetic tree.

Admixture analysis based on the SNP dataset, using cross-validation error (Figure S1), revealed that the CV error reached its minimum value at K = 1, indicating that genetic differences among the four C. chengiana populations were relatively small, and their genetic relationships were close. At K = 2, all individuals were grouped into two clusters, with DK, BW, and RJ mainly forming one group, while SA individuals primarily constituted the other. When K = 4, genetic mixing between the different populations became apparent, revealing a more complex genetic background. Notably, most SA individuals showed a distinct different genetic component (purple color) from other groups. In the PCA results, PC1 (4.29%) and PC2 (3.91%) together explained 8.2% of the total variation (Fig. 3B). Individuals from the three native populations exhibited higher clustering, while those from the ex situ population DK were more dispersed. The 30 C. chengiana individuals were not clearly separated on the score plot, particularly among the DK, BW, and RJ populations, mirroring the population structure observed at K = 2. However, the three native populations were well-separated on the score plot, with SA samples clustering on the negative axis of PC1, and BW and RJ samples clustering on the positive axis. The heatmap of genetic distances between the 30 C. chengiana samples (Fig. 3C) showed that SA population samples were more clearly differentiated from the other populations. The relatively small genetic distances between BW and RJ samples.

Fig. 3.

Population structure, principal component analysis (PCA) and nucleotide diversity of four Cupressus chengiana populations. (A) Proportions of genetic clusters for accession from K = 2, 3, 4, and 5. Different colors represent distinct ancestral genetic clusters. (B) PCA of 30 Cupressus chengiana individuals based on filtered single nucleotide polymorphism (SNP). (C) Genetic distance heatmap among the four populations: ex situ conserved population (DK) and three native populations (BW, SA, RJ). Colors from green and white to orange indicate low, medium, and high genetic distance

Pairwise Fst values among C. chengiana populations ranged from 0.024 to 0.091, with an average of 0.057 (Table 2). The Fst values for DK_vs_RJ (Fst = 0.024), BW_vs_DK (Fst = 0.037), and BW_vs_RJ (Fst = 0.046) were all below 0.05. In contrast, the Fst values for DK_vs_SA (Fst = 0.068), RJ_vs_SA (Fst = 0.076), and BW_vs_SA (Fst = 0.091) were all between 0.05 and 0.15. Correspondingly, the Nm values for DK_vs_RJ (10.017), BW_vs_DK (6.464), and BW_vs_RJ (5.173) were the highest. And the Nm values for DK_vs_SA (Nm = 3.450) and RJ_vs_SA (Nm = 3.024) were relatively lower. The BW_vs_SA pair had the lowest Nm value (Nm = 2.497).

Table 2.

Pairwise Fst values among four Cupressus Chengiana populations: ex situ conserved population (DK) and three native populations (BW, SA, RJ)

Compare	Fst	Nm
BW_vs_DK	0.037	6.464
BW_vs_RJ	0.046	5.173
BW_vs_SA	0.091	2.497
DK_vs_RJ	0.024	10.017
DK_vs_SA	0.068	3.450
RJ_vs_SA	0.076	3.024

Note: Fst, Fixation statistic; Nm, Gene flow

Genome-wide SNP transition and transversion patterns with population-specific variations

Among the 1,259,610 SNPs, transition mutations were more common, representing 73.57% (926,754 SNPs), while transversions accounted for only 26.43% (332,856 SNPs), resulting in a transition/transversion ratio (Ti/Tv) of 2.78 (Fig. 4A). The G/A transition was the most frequent at 21.33%, whereas the G/C transversion was the least common at 1.54%. The frequencies of the two transition types were similar (A/G: 15.59%, T/C: 15.37%), with C/A having the highest frequency among transversions at 4.72%. Across populations, the SNP variant distribution was generally consistent with the overall pattern, though there was a slight increase in the proportion of transversions, particularly G/A and C/T (Fig. 4B-E).

Fig. 4.

Summary result of the single nucleotide polymorphism (SNP) types of four Cupressus chengiana populations: ex situ conserved population (DK) and three native populations (BW, SA, RJ)

Discussion

Impact of ex situ conservation on genetic diversity of C. chengiana

The key genetic diversity indicators among populations include He, Ho, and Nei [42–44]. Comparsion of Ho and He can reveal the genetic structure of a population and the potential risk of inbreeding. Inbreeding may occur within the population when Ho < He [45]; the population is in genetic equilibrium when Ho = He [42, 46]. Nei is primarily used to measure genetic variation within or among populations [47, 48]. In the present study, C. chengiana had a low-level He of 0.259 (Table 1), which is lower than many of other endangered plants such as Taxus cuspidata (0.261), Pinus gerardiana (0.323), Torreya grandis (0.563), and Pinus squamata (0.373) [49–51]. Additionally, the Nei value for C. chengiana was 0.293, which is higher than that of Taxus cuspidata (0.038), Parrotia subaequalis (0.203), and Geodorum eulophioides (0.267) [49, 52, 53]. These findings suggest that C. chengiana possesses moderate genetic diversity, implying a certain degree of adaptability and survivability [54].

The DK population showed higher genetic diversity and heterozygosity potential (Table 1). In contrast, the RJ population had a higher Ho than He, indicating a larger number of heterozygous individuals and higher diversity. The BW population’s genetic data suggests stable genetic structure due to random mating, while the SA population’s lower Ho and genetic diversity indicators point to inbreeding. This is consistent with FIS observed in DK and SA populations. Similar declines in genetic diversity due to inbreeding have been seen in other species, like Pulsatilla patens [55]. Geographic isolation and inbreeding often coursed the reduction of diversity [56]. Ex situ populations often exhibit lower genetic diversity compared to their native populations, as seen in endangered species like Primula reinii, Attalea crassispatha, and Ammopiptanthus nanus [57, 58]. However, the DK population of C. chengiana, possibly due to wild seedlings transplantation, exhibits higher genetic diversity than expected for ex situ conservation. Wild seedlings transplantation is more effective in preserving the genetic diversity of the species [33]. The primary determinant of the genetic diversity of DK is the genetic variation in the seedlings sourced from its natural populations. This indicates that ex situ conservation can be effective in preserving genetic diversity of C. chengiana, as supported by recent studies [59].

Among the three native populations (BW, RJ, and SA), the RJ and BW populations exhibit higher genetic diversity and more open mating systems, suggesting a healthier genetic condition. Notably, the RJ population has higher heterozygosity, which may indicate that it has faced stronger selective pressures or maintained a larger population size during its evolutionary history. Conversely, the SA population has lower genetic diversity and shows evidence of inbreeding, signaling a potential risk of genetic diversity loss [60]. This situation could be linked to a smaller population size, geographic isolation, or human interference [61]. Moreover, the loss of genetic diversity may reduce the SA population’s ability to adapt to environmental changes and elevate its risk of extinction.

The population structure and genetic differentiation of ex situ C. chengiana population

Population structure in C. chengiana is influenced by mutation, gene flow, and habitat fragmentation, as observed in this study. However, our findings are limited to populations within the Dadu River Basin. To fully understand the species’ genetic architecture, future studies should incorporate populations from the Minjiang and Bailongjiang River Basins, where distinct ecological conditions may drive divergent genetic adaptations [34]. The current focus on Dadu River populations is driven by their immediate vulnerability to hydropower development, which necessitates urgent conservation interventions. C. chengiana populations in the Bailongjiang River Basin have undergone dramatic declines due to unmitigated anthropogenic pressures, particularly unregulated logging [34]. Despite this limitation, our results provide actionable insights for preserving genetic diversity in anthropogenically threatened habitats. Population structure is influenced by mutation, gene flow, natural selection, and genetic drift, which are essential for conservation strategies [62]. Genetic differentiation, a measure of genetic differences between populations, is a key indicator of population structure [63]. We found that the ex situ population DK exhibited lower differentiation compared to the native populations BW and RJ, which may share a common origin (Table 2). Based on genetic diversity parameters, PCA, and structure analyses, no significant genetic differentiation was observed among the three populations DK, BW, and RJ, except for the SA population (Figs. 2A and 3). However, higher genetic differentiation was observed between the SA population and the others, likely due to lower gene flow, as indicated by an Fst value > 0.05 and an Nm value < 4 (Table 2). This is further supported by the significant lower mean annual temperature of SA population compared to the other three populations (Figure S1).

Higher gene flow can reduce genetic differentiation between populations by inhibiting genetic drift [64]. The relatively lower genetic differentiation observed in the DK population may be attributed to its origin from a larger wild population that harbors a wealth of genetic resources. This rich genetic reservoir likely facilitates effective gene pool mixing and enhances genetic diversity [65, 66]. And populations with diverse genetic origins tend to exhibit more complex genetic backgrounds [13]. Furthermore, the continued random mating within these mixed gene pools promotes genetic fusion, which further diminishes differentiation between the DK and native populations [67, 68]. Additionally, the geographic proximity of DK to the BW, SA, and RJ populations could also enhance gene flow between them, preserving genetic similarity across these groups [69].

The genetic similarity between the BW and RJ populations suggests that they share a common gene pool and have undergone similar evolutionary processes. In contrast, the SA population shows significant genetic differentiation from the other two populations (BW and RJ) and exhibits low gene flow. This differentiation may be due to ecological barriers caused by lower temperatures [70], which restrict gene exchange between the SA population and the others (Figure S1).

Genetic variations among populations primarily stem from inherent genetic differences in seedlings prior to transplantation. In ex situ conservation, environmental discrepancies between habitats can shift genetic diversity in C. chengiana [36, 71]. Maintaining native-like conditions is critical to preserve genetic stability [72]. Notably, the SA population, experiencing lower MAT, showed the lowest genetic diversity (He = 0.223) and higher differentiation (Fst > 0.05), suggesting temperature may acts as an ecological barrier limiting gene flow. Temperature directly impacts physiological processes (e.g., photosynthesis) and genetic stability [16, 73], while extreme highs may restrict growth [74]. Although altitude differences among populations were minor [34], combined effects of altitude and precipitation could still influence acclimatization. Consistent SNP profiles across populations further indicate shared genomic backgrounds (Fig. 4). Mirroring native environments in ex situ conservation is thus essential to safeguard adaptive potential.

Overall, the successful establishment of DK population demonstrates the effectiveness of ex-situ conservation strategies in maintaining genetic diversity. First, the DK population exhibits high genetic diversity, with a rich genetic resource pool that maintains genetic similarity to wild populations through gene flow. Second, the environmental conditions of the DK population, particularly temperature, are similar to those of wild populations, which helps maintain genetic stability and adaptability. Third, the distribution of SNP variation in the DK population is consistent with that of wild populations, indicating a representative genetic background.

Our GBS analysis showed no significant genetic differentiation between ex situ and native C. chengiana populations, suggesting that ex situ conservation preserves genetic diversity without causing genetic drift or bottleneck effects (Fig. 5). Based on the genetic diversity and population structure of C. chengiana observed in this study, several strategies are proposed to optimize its conservation: (1) Strengthen the conservation of existing native populations: Protect habitats, prohibit logging, and implement sustainable management practices to maintain the genetic diversity of the species. (2) Select sites for ex situ conservation that closely resemble the native environment to maximize the preservation of C. chengiana ‘s genetic diversity. (3) Combine ex situ and native populations to conserve C. chengiana resources while enriching its genetic diversity. (4) Integrate population genetics with molecular biology (e.g., transcriptomics) and physiology (e.g., photosynthesis rate, stress tolerance) to assess the species’ ability to cope with future changes and to plan conservation strategies.

Fig. 5.

Schematic diagram illustrating changes in genetic diversity and differentiation between ex situ (DK) and in situ (BW, SA, RJ) populations of Cupressus chengiana. The red arrows represent significant gene flow between populations, while the green arrows indicate nonsignificant genetic differentiation

Our findings are based on data gathered from the first seven years of ex situ conservation, offering a short-term view. As previous research indicates, adaptive evolution can occur in the short term under ex situ conditions [75]. Therefore, longer-term genetic monitoring is needed to ensure that ex situ populations can continue to maintain healthy genetic diversity and to assess the long-term effects of ex situ conservation of C. chengiana.

Conclusion

In the present study, the ex situ conserved population (DK) exhibited the higher genetic diversity and lower differentiation than native populations, while the native SA population showed inbreeding and reduced genetic diversity. Overall, these results demonstrate the feasibility and potential effectiveness of ex situ conservation for C. chengiana. Based on the study, we recommend strengthening long-term genetic testing of both ex situ and native populations. When necessary, new genetic groups can be introduced to enrich and enhance genetic diversity. This study provides valuable insights into the genetic diversity of C. chengiana populations and highlights the feasibility of ex situ conservation for C. chengiana.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Author contributions

C.Y. designed methodology; C.C., X.P., Q.L., L.X., and H.C. collected the data; C.C. analyzed the data; C.Y., C.C., H.C., X.P., and led the writing of the manuscript. All authors contributed critically to the drafts and gave final approval for publication.

Funding

This research was supported by China Renewable Energy Engineering Institute (ZS-KJHB-20220014) and the National Natural Science Foundation of China (No. 32171756).

Data availability

The raw sequencing data are available at the National Genomics Data Center database under the accession number CRA021699 (https://ngdc.cncb.ac.cn/gsa/browse/CRA021699).

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.