1. Introduction

Mountain ranges are widely recognized as major drivers of evolution and biodiversity (Xie et al., 2004; Badgley et al., 2017; Antonelli et al., 2018; Li et al., 2019; Rahbek et al., 2019; Marder et al., 2025). By restricting gene flow, fragmenting populations, and promoting lineage divergence, mountains create opportunities for genetic divergence—a process initiated when uplift events create towering physical barriers that isolate once-continuous groups (Badgley et al., 2017; Antonelli et al., 2018). Most mountain systems, however, do not act as absolute barriers but rather as filters, permitting limited dispersal of certain lineages or modes while strongly constraining overall movement (Slechtova et al., 2004; Xie et al., 2004; DeChaine and Martin, 2005; Jensen et al., 2024; Salgado-Roa et al., 2024). In parallel, mountainous terrain generates pronounced environmental heterogeneity across elevation, slope, and aspect, exposing populations to contrasting climatic conditions that impose strong selective pressures and drive local adaptation—further reinforcing genetic divergence (DeChaine and Martin, 2005; Riddle, 2016; Badgley et al., 2017; Li et al., 2019). Globally, evidence from major mountain systems—including the Alps, Andes, and Rocky Mountains—demonstrates that the combined effects of restricted dispersal and environmental filtering shape population divergence and adaptive evolution (Slechtova et al., 2004; DeChaine and Martin, 2005; Jensen et al., 2024; Salgado-Roa et al., 2024).

Local adaptation is a key evolutionary process that generates and maintains plant diversity, particularly in fragmented and environmentally heterogeneous landscapes created by mountain barriers (Parmesan, 2006; Hoffmann and Sgro, 2011; Sloat et al., 2020). Through natural selection, populations shift allele frequencies to match local conditions, resulting in distinct physiological and genetic traits (Parmesan, 2006; Hoffmann and Sgro, 2011; Sloat et al., 2020). Adaptation commonly occurs at the level of individual genes or pathways, reflecting fine-scale responses to local environmental challenges (Messer and Petrov, 2013; Kelly, 2019; Irving-Pease et al., 2024). In some cases, strong or persistent selective pressures can drive large-scale genomic sweeps across extended genomic regions (Messer and Petrov, 2013). When similar selective pressures act across populations, parallel sweeps may occur (Huang et al., 2025), whereas divergent pressures can lead to broad genomic differentiation (Nosil et al., 2009). Such events often coincide with population declines, during which genetic diversity and effective population sizes still remain sufficient for selection to efficiently favour advantageous alleles (Messer and Petrov, 2013). Emerging evidence suggests that adaptation may operate not only at individual loci but also through large genomic blocks, facilitating coordinated responses to multifaceted environmental pressures (Gompert et al., 2025; Huang et al., 2025).

As the defining biogeographic feature of Arid Central Asia (ACA), the Tianshan Mountains—extending ~2500 km as the world’s longest temperate arid mountain system—provide an ideal natural laboratory for investigating how mountains drive population divergence and local adaptation (Xu et al., 2010; Yao et al., 2022). Acting simultaneously as a major physical barrier (Yao et al., 2022) and a regional “Water Tower” (Jin et al., 2024), the Tianshan Mountains divide Central Asia into two distinct climatic zones (Qiao et al., 2023). The northern slope intercepts moisture-laden westerlies, nourishing forests and meadows, whereas the southern flank casts a profound rain shadow over the hyper-arid Taklamakan Desert (Yao et al., 2022). Quaternary climatic oscillations likely triggered rapid changes in plant genetic diversity and population size across the region (Meng et al., 2015), while the stark north-south climatic divergence (Tang et al., 2022; Yao et al., 2022) imposed strong, contrasting selective pressures. Consequently, plant populations here may have employed diverse molecular mechanisms—from locus-specific selection to coordinated genomic responses— to achieve local adaptation (Nosil et al., 2009; Messer and Petrov, 2013; Wolf and Ellegren, 2017; Gompert et al., 2025; Huang et al., 2025). However, most studies on plants in this region have relied on limited molecular markers or localized ecological data, leaving the genome-wide basis of population divergence and adaptation across the Tianshan Mountains poorly understood (Zhang and Zhang, 2012; Wang et al., 2013; Jiang et al., 2014; Li and Yang, 2022; Hou et al., 2025; Yang et al., 2025). Consequently, key questions remain about how the Tianshan functions as a hard or filter barrier, and how genomic and environmental factors jointly shape divergence and plant diversity across the range.

Wetland ecosystems are crucial reservoirs of biodiversity and server as ideal model systems for examining how climate-driven selection maintains diversity, owing to their transitional characteristics and high sensitivity to climate changes (Salimi et al., 2021). Among the dominant vegetation in these ecosystems globally are emergent macrophytes of the genus Typha (Typhaceae), which form foundational components of wetland habitats (Smith, 1987). Within this genus, Typha laxmannii Lepech (2n = 30; Chepinoga et al., 2008) occupies a distinct niche characterized by its slender morphology and high ecological plasticity, enabling it to naturally colonize extensive temperate and subtropical regions across Eurasia (Zhou et al., 2016; Volkova and Bobrov, 2021; Shevera et al., 2024). In the Tianshan Mountains, T. laxmannii is the most widespread representative of its genus, inhabiting low-altitude wetlands on both the northern and southern slopes (Zhou et al., 2016). Previous investigations based on limited chloroplast and nuclear gene fragments have shown that populations north of the Tianshan Mountains share haplotypes with those to the south while also harbouring private haplotypes, indicating that the Tianshan Mountains may act as a barrier to recent gene flow for this species (Zhou et al., 2016). Leveraging its widespread distribution across environmentally contrasting regions, we conducted a population genomic analysis of T. laxmannii to elucidate the evolutionary history of wetland biodiversity in ACA. Our study aims to: (1) determine how the Tianshan Mountains restrict gene flow and structure populations of T. laxmannii, and (2) identify the genomic signatures underlying population divergence and local adaptation. These findings elucidate how mountain systems actively shape biodiversity and highlight the adaptive mechanisms supporting population persistence amid environmental shifts. Furthermore, this study provides novel insights into how mountain barriers drive diversification, offering a methodological framework with broad implications for understanding biogeographic patterns in dryland regions of Central Asian.

2. Materials and methods 2.1. Summary of experimental and analytical procedures

Detailed descriptions of the experimental protocols and bioinformatic workflows are provided in the Supplementary Information. The following outlines the key steps concisely.

A chromosome-level reference genome for Typha laxmannii was generated by integrating long-read (Oxford Nanopore), short-read (Illumina), and chromatin conformation capture (Hi-C) sequencing data. Assembly quality was assessed using BUSCO for completeness and BWA-MEM (Li, 2013) for read mapping.

We performed whole-genome resequencing of 126 individuals sampled across the Tianshan Mountains. These data were used to characterize population genetic structure and infer demographic history. To investigate local adaptation, we employed an integrative framework combining Ecological Niche Modelling (ENM), genotype–environment association via Redundancy Analysis (RDA), and genome-wide scans for selective sweeps.

Software versions, specific filtering parameters, and detailed analytical configurations are provided in Appendix A.

3. Results 3.1. Chromosome-scale genome and whole-genome duplication events

We generated a high-quality chromosome-scale genome assembly of Typha laxmannii by integrating Oxford Nanopore Technologies (ONT) long reads (28.07 Gb, ~126× coverage), next-generation sequencing (NGS) short reads (31.07 Gb, ~140× coverage), and chromatin conformation capture (Hi-C) data (25.79 Gb, ~116× coverage) (Table S1). K-mer analysis of the NGS reads estimated the genome size to be approximately 206.38 Mb, with a heterozygosity rate of 0.65% (Table S2). The initial assembly based on ONT reads was polished twice using NGS data, yielding a contig-level assembly of 221.45 Mb with a GC content of 39.40% and a contig N50 of 6.58 Mb. Hi-C scaffolding anchored 217,987,256 bp (98.43%) of the sequence into 15 chromosomes, with strong interaction signals observed along the diagonal (Fig. 1a and b; Table S3). The final assembly size was 221,461,996 bp. The assembly achieved high reference quality, as evidenced by a BUSCO completeness score of 98.27% (1586 out of 1614 conserved genes) and a mapping rate of 99.57% for NGS reads using BWA-MEM.

Repetitive elements comprised 42.07% of the genome, predominantly long terminal repeat (LTR) retrotransposons (24.35%) and DNA transposons (2.26%). Non-LTR elements, tandem repeats, and unclassified repeats accounted for 1.36%, 0.11%, and 14.97%, respectively. A total of 20,699 protein-coding genes were predicted, with an average length of 5483.86 bp, an average coding sequence length of 1302.77 bp, and 6.11 exons per gene. Among these, 97.28% (20,135) were functionally annotated. Additionally, 1365 noncoding RNAs were identified, including miRNAs, tRNAs, rRNAs, and snRNAs.

Phylogenetic analysis placed Typha laxmannii within the Poales clade, with all major nodes strongly supported (posterior probability = 1.00) (Fig. 1c). The divergence time between Typha and its close relative genus Ananas comosus (pineapple) was estimated to be 87.75 million years ago (Mya) (95% Highest Posterior Density, HPD: 86.33–89.07 Mya). The divergence between T. laxmannii and the T. latifolia–T. angustifolia groups was estimated at 7.97 Mya (95% HPD: 7.59–8.38 Mya).

Using the ancient whole-genome duplication (WGD) event – a shared paleopolyploidization in seed plants dated to 310 Mya (Jiao et al., 2011) – as a calibration point, we identified three WGD events in Typha laxmannii (Fig. 1d–i). The distribution of synonymous substitution rate (Ks) revealed: 1) The ancient WGD, supported by Ks peaks in T. laxmannii (Ks = 3.185), T. angustifolia (Ks = 3.409), T. latifolia (Ks = 3.42), Ananas comosus (Ks = 3.71), Oryza sativa (Ks = 3.72), and Sorghum bicolor (Ks = 3.76); 2) The σ WGD (~120 Mya) marked by Ks peak in T. laxmannii (Ks = 1.44), T. angustifolia (Ks = 1.467), T. latifolia (Ks = 1.47), A. comosus (Ks = 1.57), O. sativa (Ks = 1.45), and S. bicolor (Ks = 1.55), consistent with early monocot radiation; 3) A recent, Typha-specific WGD event (Fig. 1c) with Ks values of 0.745 in T. laxmannii (~72 Mya), 0.734 (~67 Mya) in T. angustifolia, and 0.741 (~67 Mya) in T. latifolia. Based on the WGD calibration in T. laxmannii, where Ks = 3.185 ± 0.10 and T = 310 Mya, the species-level mutation rate of T. laxmannii was estimated as.

3.2. Population genetic structure

To investigate the population genetic structure of Typha laxmannii, we analysed whole-genome resequencing data from 126 individuals across 31 sampling sites within the Tianshan Mountains region (Fig. 2a; Tables S4 and S5). Reads were mapped to the chromosome-scale reference genome, followed by stringent variant calling and filtering, yielding 3,191,263 high-quality SNPs for downstream analyses.

Principal Component Analysis (PCA) revealed a clear separation of all individuals into two distinct groups, corresponding to geographic regions north and south of the Tianshan Mountains (Fig. 2b). Phylogenetic analysis further confirmed this geographic structure, albeit with differing patterns of lineage consistency between regions. Southern populations formed a well-supported, cohesive monophyletic clade, while northern populations exhibited a more complex branching pattern. Specifically, upon the inclusion of outgroups, northern individuals appeared as a series of successive lineages, suggesting that the northern cluster has a deeper or more diverse ancestral structure than the southern cluster, which recently diverged (Fig. 2c). Population structure analysis performed using AdmixPipe (Mussmann et al., 2020) identified K = 2 as the optimal number of genetic clusters, corresponding clearly to the northern and southern populations. Most individuals exhibited nearly exclusive ancestry from their respective clusters (Fig. 2d). At higher K values (K = 3 and 4), population HM showed ambiguous and inconsistent grouping with other populations (Fig. 2d).

In summary, our integrated genomic analyses—spanning PCA, phylogenetics, and population structure—consistently support the division of Typha laxmannii in the Tianshan Mountains region into two genetically distinct clusters, geographically separated by the Tianshan Mountains (Fig. 2a).

3.3. Species distribution modelling and demographic history

Ecological niche modelling (ENM) was employed to reconstruct the current and historical suitable distribution of Typha laxmannii in the Tianshan Mountains region. Based on 31 natural collection sites and seven representative variables, we developed a model under current climatic conditions. This model was then projected onto three paleoclimatic periods: the Mid-Holocene (~6 Kya; Fig. 3a), the Last Glacial Maximum (LGM, ~21 Kya; Fig. 3b), and the Last Interglacial (LIG, ~130 Kya; Fig. 3c). The current ENM identified major suitable habitats around the Junggar Basin, the Ili River Valley, and the margins of the Tarim Basin (Fig. 3d). Demographic reconstruction using pairwise sequentially Markovian coalescent (PSMC) (Li and Durbin, 2011) analysis showed trajectories of effective population size (Ne) for both northern and southern genetic clusters: a slight decline from approximately 100 to 20 Kya, followed by sharp increase from the LGM to the Mid-Holocene, consistent with postglacial recolonization (Fig. 3e). The suitable distribution areas were in similar patterns, which were fragmented during the LIG and LGM but became more continuous during the Mid-Holocene. A pronounced expansion from the Mid-Holocene to the present occurred (Fig. 3f), after a contraction occurred from the LGM to the Mid-Holocene, reflecting postglacial recolonization (Fig. 3g). From the LIG to the LGM, suitable habitats decreased north of the Tianshan Mountains but increased to the south (Fig. 3h).

Estimated effective migration surface (EEMS) analysis (Petkova et al., 2016), which revealed pronounced spatial heterogeneity in effective migration, with significantly reduced migration rates (m < 0.1) across the Tianshan mountains, confirming their role as a major barrier to gene flow in this region (Fig. 4a). Demographic modelling estimated the divergence time between South and North groups at approximately 88.1 Kya, with no detectable gene flow between South and North groups after their split (Fig. 4b).

3.4. Environmental heterogeneity and local adaptation

Climate conditions vary significantly across the Tianshan Mountains, creating contrasting selective environments for northern and southern genetic clusters. Northern populations experience lower temperatures with greater seasonality, together with higher precipitation but reduced precipitation seasonality, whereas southern populations face warmer, more seasonally variable precipitation regimes (Fig. S3). PCA identified Bio2 (mean diurnal range, 77.88%) and Bio5 (maximum temperature of the warmest month, 21.56%) as the primary factors influencing the current distribution of Typha laxmannii (Fig. 4c), while RDA highlighted Bio4 (temperature seasonality), Bio12 (annual precipitation), and Bio15 (precipitation seasonality) as critical drivers of genetic differentiation between clusters (Fig. 4d).

Genotype–environment analysis based on LFMM revealed that temperature-related bioclimatic variables are linked to distinct adaptive strategies in Typha laxmannii. Analysis of Bio2 identified enriched GO terms related to sulfur metabolism, polysaccharide turnover, and epigenetic regulation (Fig. S4a), together with KEGG pathways including circadian rhythm and purine metabolism (Fig. S4b), indicating rapid metabolic and transcriptional responses to daily thermal fluctuations. Candidate genes including Chr13G00183620.1 (altered phosphate starvation response 1-like), Chr4G00069440.1 (DNA methyltransferase 1B-like), Chr1G00012280.1 (peptidyl-prolyl cis–trans isomerase CYP26-2, chloroplastic), and Chr8G00128150.1 (sialyltransferase-like protein 1) highlight chromatin remodelling, protein folding stability, and sugar metabolism as key mechanisms underpinning adaptation to diurnal temperature variation (Fig. 4e and Table S6). For Bio4, GO enrichment pointed out polysaccharide and pectin catabolism as well as sphingolipid biosynthesis (Fig. S5a), and KEGG pathways included alkaloid biosynthesis, thiamine metabolism, and porphyrin metabolism (Fig. S5b), reflecting adaptation to seasonal temperature extremes. Candidate genes such as Chr6G00112330.1 (chalcone isomerase), Chr1G00023120.1 (pectinesterase-like), Chr2G00030030.1 (bifunctional aspartate aminotransferase/prephenate aminotransferase), and Chr4G00078080.1 (60S ribosomal protein L21-1) exemplify roles in flavonoid biosynthesis, cell wall modification, amino acid metabolism, and translational regulation under seasonal temperature variance (Fig. 4f and Table S7). Analysis of Bio5 showed GO enrichment in ligase activities, DNA replication and repair complexes, and histone demethylase activity (Fig. S6a), along with KEGG pathways related to phenylpropanoid and flavonoid biosynthesis and mismatch and base excision repair (Fig. S6b), highlighting mechanisms of resilience to extreme summer heat. Candidate genes including Chr15G00203810.1 (4-coumarate-CoA ligase-like 1), Chr1G00003150.1 (histone H3 demethylase KDM3), Chr11G00168930.1 (2-oxoisovalerate dehydrogenase subunit beta, mitochondrial), and Chr12G00173860.1 (probable methyltransferase PMT26) implicate phenylpropanoid biosynthesis, chromatin remodelling, amino acid catabolism, and methylation as critical adaptive responses to peak heat stress (Fig. 4g and Table S8). Collectively, temperature-related bioclimatic variables shape local adaptation in T. laxmannii through integrated adjustments in metabolic plasticity, structural resilience, and genomic integrity.

Precipitation-related variables further reveal the molecular basis of water-stress adaptation. Analysis of Bio12 highlighted GO terms such as NAD + ADP-ribosyltransferase activity, protein glycosylation, and microtubule cytoskeleton organization (Fig. S7a), along with KEGG pathways including carotenoid biosynthesis, MAPK signaling, and base excision repair (Fig. S7b), pointing to roles of hormonal signaling, stress transduction, and genomic stability under water-availability gradients. Candidate genes such as Chr5G00085580.1 (trans-cinnamate:CoA ligase, peroxisomal-like), Chr9G00141350.1 (carotene epsilon-monooxygenase, chloroplastic), and Chr10G00151180.1 (respiratory burst oxidase homolog protein C-like) exemplify phenylpropanoid metabolism, ABA biosynthesis, and ROS-mediated signaling in response to water stress (Fig. 4h and Table S9). Analysis of Bio15 identified GO categories in ligase and lyase activities (4-coumarate-CoA ligase, argininosuccinate lyase) and oxidative stress responses (Fig. S8a), and KEGG pathways such as sulfur metabolism, inositol phosphate metabolism, and phosphatidylinositol signalling (Fig. S8b). Candidate genes including Chr11G00163190.1 (mitotic checkpoint kinase BUB1), Chr2G00036270.1 (CAF1/NURF55/MSI1-like nucleosome remodeling factor), Chr1G00017370.1 (Mediator complex subunit 10), and Chr11G00165920.1 (kinesin-like protein KIN-12F) underscore the importance of cell cycle regulation, chromatin remodeling, transcriptional control, and cytoskeletal dynamics in adapting to seasonal precipitation fluctuations (Fig. 4i and Table S10). Thus, precipitation-related bioclimatic variables drive adaptation through hormonal signaling, stress-responsive metabolism, and regulatory mechanisms that ensure resilience to both annual water availability and seasonal precipitation variability.

3.5. Genetic basis of north–south differentiation

To investigate the genetic basis of differentiation between northern and southern populations, we applied two complementary approaches. RAiSD (Alachiotis and Pavlidis, 2018) detected three and nine candidate genes under positive selection in the northern and southern clusters, respectively (Fig. 5a; Tables S11 and S12). In the northern cluster, these included an alkaline/neutral invertase (chromosome 10), involved in sucrose catabolism and carbon allocation; a reverse transcriptase domain-containing protein (chromosome 12), indicative of transposon-mediated genome plasticity; and mitogen-activated protein kinase 9 (MKK9, chromosome 12), a regulator of MAPK signalling in stress responses (Luo et al., 2017) (Table S11). GO enrichment of northern RAiSD candidates revealed strong overrepresentation of carbohydrate catabolic and metabolic processes, including glycopeptide alpha-N-acetylgalactosaminidase activity, beta–fructofuranosidase activity, and various glucosidase functions (Fig. S9a). KEGG analysis further highlighted MAPK signaling and hormone signal transduction (Fig. S9b), suggesting that metabolic flexibility and stress-responsive signaling contribute to adaptation on the northern slope. In the southern cluster, eight candidate genes were identified, including pre-mRNA-splicing factor ATP-dependent RNA helicase DHX38/PRP16 (chromosome 11), which regulates RNA processing (Cona et al., 2022); SNI1 (chromosome 3), a negative regulator of systemic acquired resistance involved in DNA damage repair and defence suppression (Zhu et al., 2021); and a NAC domain-containing transcription factor (chromosome 7), associated with transcriptional regulation (Hussey et al., 2015). Other candidates featured proteins with ankyrin repeats and coiled-coil domains, likely involved in protein–protein interactions (Table S12). GO enrichment pointed to RNA helicase activity, negative regulation of defense responses, and chromosomal stability (Fig. S10a). KEGG analysis identified enrichment in spliceosome, linoleic acid metabolism, and proteasome pathways (Fig. S10b), indicating that post-transcriptional regulation and proteomic turnover are important under the warmer and more arid conditions of the southern slope.

The second method, a genome-wide association study (GWAS) further identified 225 genes associated with North–South differentiation (Fig. 5b and Table S8). Notable candidates included MADS-box transcription factor 14 (chromosome 11), implicated in flowering time and development (Wang et al., 2025); PARP2/3/4 (chromosome 4), involved in genome maintenance and DNA repair (Grundy et al., 2016; Muoio et al., 2024; Li et al., 2025); and genes related to ion transport (e.g., chloride channel CLC-f-like, chromosome 2) (Park and MacKinnon, 2018), protein synthesis (histidyl-tRNA synthetase, chromosome 4; UTP10, chromosome 9) (Waldron et al., 2019), and cytoskeletal organization (tubulin beta, chromosome 11). Stress-related loci included auxin transport protein BIG/UBR4 (chromosome 8) (Grabarczyk et al., 2025) and root UVB sensitive 2 proteins (chromosome 4) (Ge et al., 2010) (Table S8). GO enrichment emphasized carbohydrate-derivative metabolism, proline metabolism, oxidoreductase activity, and lignin catabolism (Fig. S11a). KEGG pathway analysis highlighted ABC transporters, amino acid metabolism (e.g., arginine, proline, glutamate), and secondary metabolite biosynthesis, including flavone/flavonol, indole alkaloid, and betalain pathways (Fig. S11b). Together, these results indicate that both northern and southern populations exhibit region-specific and shared adaptive responses involving metabolic adjustment, stress signaling, and specialized secondary metabolism.

3.6. Selective sweep hotspots

Selective sweep analyses combining RAiSD μ statistics, nucleotide diversity (π), Tajima’s D, and pairwise F_ST identified two major sweep regions (Fig. 5a and c). Two significant selective sweeps–characterized by reduced diversity, negative Tajima’s D, and high RAiSD μ values–were detected: one on chromosome 5, present in both northern and southern clusters and characterized by low F_ST, indicating a parallel selective sweep (Fig. 5, Fig. 6a, and S12); and another on chromosome 6, characterized by high F_ST between clusters, with negative Tajima’s D in the northern population and positive Tajima’s D in the southern population, indicating a divergent sweep (Fig. 5, Fig. 6b, and S13).

The parallel selective sweep on chromosome 5 spanned a core region from 6.87 to 9.33 Mb and an extended region from 5.65 to 9.75 Mb (Fig. S12 and Fig. 6a). This region contained nine genes, most encoding transposon-related proteins such as reverse transcriptase, integrases, and retroviral proteases (Table S14). MYB6, a transcriptional repressor involved in stress responses and secondary metabolism (Liu et al., 2023), and glutaminyl-peptide cyclotransferase, which may enhance protein stability under stress (Yu et al., 2023), were identified as strong candidates. GO enrichment analysis revealed significant overrepresentation of DNA integration, RNA-directed DNA polymerase activity, DNA metabolic processes, and peptidase activity (Fig. S14). Overall, the parallel sweep likely targets genomic integrity, transposon regulation, and protein turnover, reflecting shared adaptive responses to environmental stress across populations.

The selective sweep on chromosome 6 covered a core region from 5.55 to 8.55 Mb and an extended region from 5.55 to 9.20 Mb (Fig. S13 and Fig. 6b), encompassing 21 genes enriched in functions related to oxidative stress response, signal transduction, and transcriptional regulation (Table S15). GO enrichment analysis revealed strong signals in DNA integration, glutathione peroxidase activity, tRNA dihydrouridine synthase activity, and nucleic acid metabolic processes (Fig. S15a). KEGG pathway enrichment highlighted glutathione metabolism, zeatin biosynthesis, glycosaminoglycan degradation, and spliceosome activity (Fig. S15b). Divergent selection likely reflects population-specific adaptation through oxidative stress management, hormone signaling, and transcriptional regulation.

Both sweep regions exhibited low recombination rates and higher CpG density, consistent with expectations under selective sweep theory and supporting the robustness of the inferred sweeps (Fig. 6a and b, Figs. S12 and S13). Molecular clock analysis dated both selective sweep events to the late Pleistocene. The common ancestor of the core and extended regions of the chromosome 5 sweep was estimated to date back 93,130 (Fig. 6c) and 56,696 years ago (Fig. 6d), respectively. For the chromosome 6 sweep, the common ancestor of the core and extended regions was estimated at 87,225 (Figs. 6e), 81,926 years ago (Fig. 6f), respectively.

4. Discussion

Geographic heterogeneity and climatic gradients play central roles in shaping the evolutionary dynamics of plant species, particularly in mountain systems where strong environmental contrasts intersect with major topographic barriers (Slechtova et al., 2004; DeChaine and Martin, 2005; Yao et al., 2022; Jensen et al., 2024; Salgado-Roa et al., 2024). The Tianshan Mountains in ACA exemplify this interplay, combining pronounced elevational relief with strong north–south differences in temperature regimes and water availability (Fig. S3). By integrating genomic, population-level, demographic, and ecological analyses, this study provides insight into how geographic isolation and environmental heterogeneity jointly shape genetic divergence and adaptive differentiation in the wetland herb Typha laxmannii. Our results highlight the dual role of mountain systems as both barriers to gene flow and drivers of ecological diversification, with important implications for understanding plant adaptation and persistence in climatically extreme regions.

4.1. A high-quality genome provides a foundation for evolutionary analysis

The chromosome-level genome assembly of Typha laxmannii provides a critical foundation for investigating population divergence and environmental adaptation in this widespread wetland plant. Meanwhile, this genomic resource enables robust inference of demographic history, population structure, and genotype–environment associations, allowing evolutionary patterns to be interpreted within an explicit ecological and biogeographic framework. Such integrative genomic–environmental approaches are increasingly important for understanding how plant diversity is generated and maintained across heterogeneous landscapes, including those of the Tianshan Mountains region (Aitken et al., 2024; Bernatchez et al., 2024).

Comparative genomic analyses identified a Typha-specific WGD approximately at 67–72 Mya, coinciding with the Cretaceous–Paleogene transition. This polyploidization event may have increased genomic flexibility, potentially facilitating lineage persistence and diversification during the profound environmental disruptions of this period (Fawcett et al., 2009; Koenen et al., 2021). In addition, the estimated mutation rate for Typha laxmannii is comparable to those reported for other herbaceous plants, such as Arabidopsis thaliana (6.95 × 10⁻⁹ per site per generation; Weng et al., 2019), providing a reliable parameter for demographic modelling and divergence-time estimation. Together, these genomic features establish a solid evolutionary context for interpreting population divergence and local adaptation in this species.

4.2. The Tianshan Mountains act as a filter biogeographic barrier

Population genomic analyses reveal a pronounced north–south genetic split in Typha laxmannii across the Tianshan Mountains, indicating long-term spatial isolation between populations on opposing slopes. Although T. laxmannii is broadly distributed across Eurasia (Smith, 1987), our sampling is restricted to populations adjacent to the Tianshan Mountains. This regional focus enables a direct assessment of the barrier effect of the Tianshan Mountains, although broader-scale migration or gene flow from more distant parts of the species’ range cannot be fully evaluated without more extensive geographic sampling. The observed disparity in sample numbers between slopes reflects the natural distribution of suitable wetlands rather than sampling bias, and our fieldwork covered most accessible and suitable sites (Zhou et al., 2016). Given that Typha species frequently reproduces clonally, individuals within each site were spaced sufficiently to ensure sampling of independent genets, and the sample size was adequate to represent each locality. Collectively, population structure, PCA, phylogeny, demographic modelling, and EMMS analyses consistently demonstrate that, within the Tianshan region, the mountains have imposed persistent spatial isolation since late-Pleistocene climatic fluctuations, allowing substantial genetic differentiation to accumulate (Fig. 2, Fig. 3a and b).

Comparisons with other regional taxa reveal, however, that the Tianshan Mountains act not as a hard barrier but as a filter (Xie et al., 2004), with permeability determined by ecological tolerance, habitat dependence, and dispersal capacity (Zhang and Zhang, 2012; Wang et al., 2013; Jiang et al., 2014; Li and Yang, 2022; Ali et al., 2025; He et al., 2025; Hou et al., 2025; Yang et al., 2025). Species with broad niches or high physiological resilience typically show weak genetic differentiation across the range (Hou et al., 2025; Yang et al., 2025), whereas wetland-dependent or climate-sensitive taxa often exhibit pronounced genetic breaks (Zhang and Zhang, 2012; Wang et al., 2013; Jiang et al., 2014; Li and Yang, 2022; Ali et al., 2025; He et al., 2025). This contrast indicates that the Tianshan Mountains do not impose uniform isolation; instead, they selectively filter lineages according to their ability to traverse or persist within the cold, arid, and topographically complex environments that dominate the range.

For Typha laxmannii, the barrier effect is particularly pronounced. As a shallow-wetland specialist reliant on stable hydrological conditions (Zhou et al., 2016), the species is poorly adapted to cross steep elevational gradients or survive in the fragmented, seasonally dry habitats characteristic of the Tianshan Mountains interior. These ecological constraints intensify the barrier effect, resulting in a pronounced genetic discontinuity where more flexible species maintain connectivity. Yet the Tianshan Mountains are not uniformly impermeable: the Ili River Valley serves as a major corridor, where broad valley morphology and milder climate facilitate movement. Similar corridor-mediated connectivity has been observed in other regional taxa (Zhang et al., 2021; Li and Yang, 2022; Ali et al., 2025; He et al., 2025), underscoring that the Tianshan act simultaneously as a barrier and a conduit. This spatial heterogeneity in permeability is a hallmark of filter-type biogeographic barriers, and illustrates how mountain systems shape evolutionary trajectories by coupling strong isolation with localized opportunities for dispersal.

4.3. Climatic heterogeneity across the Tianshan Mountains shapes local adaptation

Beyond restricting gene flow, the Tianshan Mountains create steep environmental gradients that shape the adaptive landscape of Typha laxmannii. Marked differences in temperature seasonality, precipitation regimes, and thermal extremes between the northern and southern slopes closely align with population genetic structure (Fig. S3). Environmental association analyses consistently identify temperature- and moisture-related variables as major predictors of allele-frequency variation, underscoring the central role of climatic heterogeneity in driving adaptive divergence (Fig. 4, Fig. 5; Tables S6–S13).

These climatic disparities hold particular ecological significance for wetland plants in continental arid regions, where pronounced seasonal fluctuations constrain growth, phenology, and stress tolerance (Magee and Kentula, 2005). Functional enrichment of environmentally associated loci suggests that northern and southern populations have evolved distinct physiological strategies to cope with divergent thermal and hydrological conditions, involving pathways related to stress response, metabolic regulation, and flowering-time control (Fig. 5; Tables S11–S13). Such findings align with broader patterns observed in global mountain systems, where steep climatic gradients can drive rapid adaptive differentiation even in species with considerable dispersal capacity (Slechtova et al., 2004; DeChaine and Martin, 2005; Jensen et al., 2024; Salgado-Roa et al., 2024).

Notably, environmental associations are spread across multiple genomic regions rather than localized to a few loci, suggesting that adaptation to the heterogeneous climate of the Tianshan Mountains is polygenic (Nosil et al., 2009; Messer and Petrov, 2013; Gompert et al., 2025; Huang et al., 2025). This observation matches theoretical expectations for complex traits such as cold tolerance, drought resistance, and phenological adjustment (Nosil et al., 2009; Messer and Petrov, 2013). Together, these results indicate that the Tianshan Mountains act not only as a physical barrier to gene flow but also as a generator of ecological heterogeneity, reinforcing population divergence through climate-driven selection.

4.4. Genomic architecture of adaptation: parallel and divergent sweep regions

Our genome-wide analyses reveal that adaptation in Typha laxmannii is not only confined to isolated loci, but also involves extended genomic regions influenced by linked selection, recombination landscape structure, and long-term climatic forcing in ACA (Fig. 4, Fig. 5, Fig. 6, Figs. S12 and S13). These large genomic regions exhibit concordant signatures of selection—including reduced nucleotide diversity, skewed allele-frequency spectra, elevated haplotype-based statistics, and localized recombination suppression—indicating that they function as cohesive evolutionary units.

The ~2.5 Mb parallel sweep on chromosome 5 is estimated to have occurred at ~91 Kya (Fig. 6a, c and d), marking an early, range-wide adaptive event in Typha laxmannii that coincides with late-Pleistocene climatic shifts in the Tianshan Mountains (Meng et al., 2015; Wei et al., 2025). Strong sweep signatures and a deep recombination trough indicate a constrained haplotype block that rose to high frequency after the onset of the population decline (~100 Kya) but before the north–south divergence (~88.1 Kya). This timing aligns with a phase of intensified continentality, enhanced winter cooling driven by the Siberian High, and repeated glacial–interglacial aridification cycles in the Tianshan region (Wei et al., 2025). Such prolonged harsh environmental conditions likely imposed persistent, range-wide selective pressures on stress-response pathways, favoring genomic configurations that stabilize gene expression amid abrupt thermal and moisture fluctuations (Meng et al., 2015; Wei et al., 2025). Demographic contraction likely further increased the efficiency of selection (Fig. 3, Fig. 4b), enabling beneficial regulatory architectures to sweep rapidly and uniformly across the species’ range. The presence of MYB6 and glutaminyl-peptide cyclotransferase, along with transposon-related protein-coding genes, supports the interpretation that this sweep favored a coordinated stress-response module enhancing transcriptional robustness and metabolic homeostasis (Liu et al., 2023; Yu et al., 2023).

The ~3.0 Mb divergent sweep on chromosome 6 represents a northern-specific adaptive region that has become a genomic island of divergence under the extreme climate of northern ACA (Fig. 5, Fig. 6b). Northern populations show sharply reduced nucleotide diversity and strongly negative Tajima’s D, consistent with a recent selective sweep, whereas southern populations exhibit positive Tajima’s D, indicating balancing selection or retention of ancestral variation (Fig. 5, Fig. 6b). Elevated F_ST further supports a tightly linked haplotype block selectively maintained in the north (Fig. 5c). This region is enriched for genes involved in oxidative stress response, hormone signaling, and transcriptional regulation—functions essential for coping with intensified winter cooling, high temperature seasonality, and prolonged freeze–thaw cycles driven by the strengthened Siberian High (Fig. S15 and Table S15). Glutathione-related pathways, zeatin biosynthesis, and spliceosome-associated functions point to enhanced stress signaling and transcriptional plasticity under severe cold and desiccation stress (Fig. S15). High CpG island density and a pronounced recombination trough suggest that epigenetic features and structural constraints help maintain haplotype integrity (Fig. S13 and 6b). The estimated sweep timing (~85 Kya) aligns with intensified northern cooling and aridification after the late Pleistocene (Jouzel et al., 2007; Wei et al., 2025), indicating that this haplotype arose as a localized adaptive response after north–south environmental differentiation had begun. In contrast to the chromosome 5 parallel sweep, the chromosome 6 sweep reflects spatially heterogeneous climatic pressures that drove population-specific genomic restructuring across the Tianshan Mountains (Wolf and Ellegren, 2017).

Taken together, the parallel sweep on chromosome 5 and the divergent sweep on chromosome 6 show that adaptation in Typha laxmannii has been shaped by both shared and region-specific selective pressures across the heterogeneous landscapes of the Tianshan Mountains region. The chromosome 5 region reflects a range-wide adaptive response established under prolonged climatic harshness and reinforced by demographic contraction, whereas the chromosome 6 region illustrates how spatially structured climatic regimes—particularly intensified northern winter cooling—can drive localized selective sweeps that form genomic islands of divergence.

4.5. The Tianshan Mountains biodiversity: heterogeneity, sorting, and evolution

Our findings suggest that a “spatio-temporal environmental sorting” mechanism underlies this dynamic. Unlike high-productivity biomes such as tropical rainforests or coral reefs—where diversity is often sustained by resource abundance and niche partitioning (Connell, 1978), biodiversity in the Tianshan Mountains arises from adaptive refinement under persistent climatic stress (Fig. 3, Fig. 6 and Fig. S3). The complex vertical and horizontal terrain of the Tianshan creates a diverse mosaic of microclimates and habitats, establishing a physical template of varied selective pressures (Tang et al., 2022; Yao et al., 2022). This spatial complexity, epitomized by the stark north–south contrast in temperature and moisture regimes, is further intensified by deep-time climatic oscillations, such as the Pleistocene glacial–interglacial cycles (Meng et al., 2015; Qiao et al., 2023). These historical fluctuations acted not merely as disturbances, but as potent evolutionary filters. As regional climates shifted, the mountains did not simply impede dispersal—they selectively sorted lineages according to physiological tolerance, promoting the persistence and local adaptation of those capable of withstanding region-specific stresses. The outcome of this prolonged sorting process is the emergence of distinct, region-specific adaptive genomic architectures (Fig. 5, Fig. 6). In Typha laxmannii, this is reflected in divergent selective sweeps (e.g., on chromosome 6) and well-defined north–south genetic clusters, which represent genomic legacies of long-term adaptation. In this context, topographic and climatic severity have not merely constrained life but have actively accelerated the fixation of resilient genotypes, turning the mountains into a cradle of evolutionary novelty.

Consequently, this region can be viewed as an archipelago of “evolutionary islands,” where each mountain range functions as a geomorphological laboratory—dictating evolutionary trajectories through persistent environmental filtering (Muellner-Riehl et al., 2024). Our findings underscore that these mountains are not passive barriers but active drivers of diversification, powered by the spatio-temporal environmental sorting mechanism. In this way, they continually shape and safeguard the unique biotic heritage of Central Asia in the face of ongoing global change.

5. Conclusion

This study demonstrates that the Tianshan Mountains act as a dual force in plant evolution, functioning as a selective filter for gene flow and a primary generator of adaptive ecological gradients. By generating a chromosome-scale genome and integrating population genomic, demographic, and environmental analyses, we reveal restricted gene flow since the late Pleistocene and distinct adaptive trajectories between northern and southern populations across the Tianshan Mountains. Climatic heterogeneity emerged as a major driver of genomic differentiation, and both shared and region-specific sweep regions indicate that adaptation involves coordinated selection across extended genomic blocks. Our findings collectively support a model of biodiversity governed by spatio-temporal environmental sorting. This model provides a pivotal framework for deciphering a key mechanism in the ACA mountains: how their topography and climate work in concert to both generate evolutionary novelty and temper the impacts of climate change, thereby underpinning genetic resilience.

Acknowledgements

This study was supported by the National Water Pollution Control and Treatment Science and Technology Major Project, China (No. 2015ZX07503005). The calculations in this paper were performed using the super computing system at the Supercomputing Center of Wuhan University.

CRediT authorship contribution statement

Chichi Zhao: Investigation, Formal analysis, Data curation, Visualization, Writing—original draft, Writing—review and editing. Lei Huang: Investigation, Formal analysis, Data curation, Writing—review and editing. Huijun Wang: Investigation, Formal analysis, Data curation, Writing—review & editing. Xinwei Xu: Conceptualization, Methodology, Resources, Supervision, Project administration, Writing—review & editing.

Data availability statement

The genome sequencing data generated in this study have been deposited in the National Genomics Data Center (CNCB-NGDC) under BioProject accession number PRJCA045581. The raw sequencing reads are available under GSA accession number CRA029658, and the assembled reference genome is accessible under GWH accession number GWHGPYE00000000.1. The analysis workflow and scripts used in this study are available on GitHub at https://github.com/chichizhao/Tianshan-Mountains-barrier-to-Typha-laxmannii.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

Supplementary data to this article can be found online at https://doi.org/10.1016/j.pld.2026.03.002.