Chromosome-level genome of Neodon fuscus sheds light on the evolution and plateau adaptation of N. fuscus and Neodon

Abstract

Background

Neodon, a typical group of high-altitude rodents, originated in the high-altitude regions of the Tibet-Himalayan area. These rodents inhabit meadows, swampy grasslands, and shrublands at elevations ranging from 3,700 to 4,800 m in Qinghai-Tibet Plateau, where they experience environment stresses such as hypoxia, enhanced ultraviolet radiation, and low temperature.

Results

In this study, we assembled a high-quality chromosome-level genome of Neodon fuscus and use it to investigate the potential genetic bases of high-altitude adaptation. The genome of N. fuscus shows gene family expansion in genes related to immunity. Comparative genomics analysis revealed several positively selected genes and rapidly evolving genes associated with DNA damage repair (Atm and Pdia4) and male sperm quality (Idh3b and Atplb1) that aid in clarifying high-altitude adaptation of Neodon species. Additionally, we found that genes (Nrp2, Vav3, Cat and Adam8) associated with oxidative stress and cardiovascular regulation seem to have experienced convergent evolution among plateau species. Notably, we identified a Neodon-specific 24 bp deletion in Map3k6, which may regulate angiogenesis, aiding Neodon species adapt to hypoxic environments.

Conclusion

These findings elucidate molecular mechanisms underlying high-altitude adaptation, emphasizing the integrative roles of immunity, stress response, and reproduction. The high-quality genome of N. fuscus offers new insights into how Neodon adapt to high-altitude environments and will facilitate future research on Neodon.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12864-025-11709-4.

Background

The Qinghai-Tibet Plateau (QTP), the highest plateau in the world, with an average altitude exceeding 4,500 m, exerts tremendous survival pressures on its endemic mammalian species due to hypoxia, low temperatures, and intense ultraviolet radiation [1]. These conditions have driven notable phenotypic and genetic adaptations among QTP species [2]. Extensive research on QTP species has shown that genes related to hypoxic stress, energy metabolism, and angiogenesis are under selection for high-altitude adaptation [3–6]. Studies also suggest that different species may evolve distinct genetic mechanisms in response to the same environmental pressures [7, 8]. To date, a set of genes associated with high-altitude adaptation have been reported, such as EPAS1, EGLN1, HIF-1, etc. [9–12]. However, this high-altitude adaptation related research mainly focused on humans [13–17] and large domesticated animals such as cattle [18], yaks [19], pigs [20], sheep [21], and goats [22], with relatively few studies on other mammals [23–26]. This hinders the understanding of whether other plateau species have evolved convergent or unique mechanisms to adapt to the same environment.

Neodon is a genus of small rodents found exclusively in the Tibet-Himalaya region, primarily inhabiting meadows, swampy grasslands, and shrublands at elevations of 3,700–4,800 m in QTP. Previous studies have mainly focused on morphology and taxonomy of these rodents [27, 28]. Only recent morphological and molecular research have confirmed the monophyly of Neodon and identified multiple species within the genus [29–31]. In recent years, research on Neodon has increasingly focused on the molecular mechanisms underlying their adaptation to the QTP. Comparative transcriptome analyses have explored changes in the heart, lungs, and skeletal muscles of N. fuscus and found that N. fuscus exhibits various characteristics, including the conversion of energy metabolic substrates from fatty acids to glucose [32–34]. To date, only a scaffold-level genome (N. shergylaensis) of Neodon species was published [35], which significantly hinders research into adaptive changes in Neodon during their plateau adaptation.

In this study, we assembled the chromosome-level genome of N. fuscus, a basal species of the Neodon genus, using Nanopore and Hi-C sequencing technologies. We aim to explore the molecular mechanisms of adaptation of N. fuscus and Neodon to the plateau through comparative genomic analysis, convergent evolution analysis and structural variation analysis. This high-quality reference genome will provide valuable information for future genetic studies on Neodon.

Methods

Sampling and sequencing

The sample used in this study was a male N. fuscus, which was captured in Golog Tibetan Autonomous Prefecture, Qinghai Province, China. High-quality DNA was extracted from muscle tissue using the Genomic DNA Extraction Kit (Cat#13323, Qiagen, Germany). Extracted DNA that met quality and quantity standards was used to construct an Oxford Nanopore (PromethION) library (Table S1). RNAs from six different tissues (brain, kidney, heart, testis, liver, and lung) were also extracted and sequenced on the Illumina NovaSeq 6000 platform. Blood samples from the same individual were used for genome sequencing and Hi-C library construction, with sequencing performed on the Illumina NovaSeq 6000 platform. Approximately 300.5 Gb of data were generated from the Hi-C library sequencing. All sequencing services were provided by Haorui Genomics Technology Co., Ltd. (Xi’an, China). The euthanasia was carried out by cervical dislocation. Criteria for euthanasia were based on an assessment by our veterinary staff following AAALAC guidelines. The experimental protocols were approved by the Animal Care and Use Committee of Shaanxi Normal University and were conducted in accordance with the ethical principles of animal use and care in China.

Genome assembly

The genome size of N. fuscus was estimated based on the ~ 30 Gb publicly available Illumina reads (SRR10150281) using Jellyfish v2.3.0 [36] with different kmer sizes ranging from 17 to 25 in steps of 2. GenomeScope2.0 [37] was used to estimate heterozygosity and the rate of duplication with default parameters. All Nanopore reads were corrected using NextDenovo v2.5.2 [38] with default parameters and were assembled using Flye v2.9.3 [39] with parameters “--nano-raw --iterations 2”. Assembled contigs were further polished by NextPolish v1.4.1 [40] with default parameters for three rounds using Illumina reads. Redundant sequences were identified and removed using PURGE_DUPS [41]. All Hi-C reads were used to construct a chromosome-level assembly using Juicer v1.6 [42] and 3D-DNA v180922 [43]. Final assembly were curated manually with Juicebox v1.11.08 [44]. Benchmarking Universal Single-Copy Orthologs (BUSCO) v5.4.6 [45] was employed to assess the integrity of the gene sets in our assembly using mammalia_odb10 dataset. We also estimated the mapping rates of our transcriptome data to the genome assembly of N. fuscus [46, 47].

Repeat annotation

De novo and homology-based prediction were performed to annotate repetitive sequences. The de novo repeat database was built using RepeatModeler v2.0.3 (http://www.repeatmasker.org/RepeatModeler/) with default parameters. Then all transposable elements (TEs) were identified using RepeatMasker v4.1.0 [48], with parameters “-pa 32 -engine ncbi -species Rodentia -s -no_is -cutoff 255 -frag 20000 -gff”, combined with the data from the Repbase transposable element library and the constructed de novo repeat library. Tandem repeats were predicted using Tandem Repeat Finder [49] with default parameters. The evolution of TEs of species in Neodon was explored by calculated the insertion time of each TE groups using the formula T = k/2r, where T represents the insert time, K represents the kimura value, and the r represents the evolutionary rate [50].

Gene annotation

Protein-coding genes of N. fuscus were predicted using ab initio prediction, RNA-seq based prediction and homology-based prediction. For ab initio prediction, Augustus v3.4.0 [51], SNAP v2017-03-01 [52] and Genescan v1.0 [53] were used with default parameters. For homology-based prediction, the whole protein sequences of Microtus oregoni (GCF_018167655.1), Mesocricetus auratus (GCF_017639785.1), Onychomys torridus (GCF_903995425.1), Peromyscus leucopus (GCF_004664715.2), Mus musculus (GCF_000001635.27) and Homo sapiens (GCF_000001405.40) were downloaded from NCBI datasets. Protein sequences were aligned against N. fuscus genome assembly using BLASTP v2.14.1 + [54] and were identified using GeneWise v2.4.1 [55] with default parameters. For RNA-seq annotation, Trinity v2.15.2 [56] was used to assemble transcriptome sequences for each tissue with default parameters. PASA v2.5.2 was used to mapped assembled transcripts to the genome assembly for gene structure prediction. Finally, EVidenceModeler v1.1.1 [57] was used to combine the result of all above predictions to generate a consensus gene set with default parameters. For functional annotation, all annotated coding genes of N. fuscus were searched against SwissProt database and TrEMBL [58] using BLAST (evalue, 1e-5) [59].

Phylogenomic analysis and divergence time estimation

OrthoFinder v2.5.4 [60] with default parameters was used to identify one-to-one orthologues among N. fuscus, five other Arvicolinae species (Neodon shergylaensis, Microtus ochrogaster, Arvicola amphibius, Myodes glareolus and Ondatra zibethicus) and one outgroup (Cricetulus griseus). All orthologues protein sequences were aligned using Prank v170427 [61] and trimmed by pal2nal [62] with parameters “-nogap -output fasta”. Then four-fold degenerate sites were extracted using in-house script. The extracted matrix was used to construct phylogenetic tree using RAxML v8.2.12 [63] with parameters “-T 100 -f a -N 100 --bootstopn  perms = 1000 -m GTRCATI -x 12345 -p 12345 -o outgroup” and was used as an input file to infer divergence times using the MCMCTREE program in the PAML software (v4.9j) [64, 65]. A total of three fossil calibrations were used to calibrate the following nodes with uniform priors. The calibrated time nodes were “ > 3.6 Mya” for the Arvicolini lineage [66], “ > 4,13 Mya” for the Ondatrini lineage [66–68], and “ > 7 Mya < 10 Mya” for the split between the Arvicolinae lineage and the Cricetinae lineage [69]. The tree prior assumed a uniform birth-death process with default parameters. MCMC chains was run for totally twenty-five million cycles, sampling every 50, after the initial 200,000 cycles that were discarded as burnin. Convergence, was checked a posteriori in Tracer v.1.7 and all parameters obtained high ESS values > 200.

Gene family expansion, contraction and adaptation analysis

With the classified orthologous genes and the estimated time tree, we performed a CAFÉ v5.0 [70] analysis to detect the gene family expansions and contractions, in which the default parameters of lambda (-s) and a P value cutoff of 0.01 were used to detect the significant expansions/contractions along each branch. All single copy orthologous identified above were used to identify positive selective evolution and rapid evolution on each gene at the branch of the ancestor of N. fuscus/N. shergylaensis. To identify the PSGs, we used branch-site model. To do this, we first estimated lineage-specific evolutionary rates for each gene using the Codeml module with a neutral evolution model in the PAML software (v4.9j). Next, we applied the positive selection model to identify genes with a higher ω (dN/dS) in specific lineages compared to the rest of the tree. A likelihood ratio test (LRT) was conducted to compare the two models. For REGs, we applied both the one-ratio and two-ratio models. Significance was tested using a chi-square test, and results with P ≤ 0.05 and P ≤ 0.01 were considered significant and very significant for PSGs and REGs, respectively. The candidate genes were passed to Metascape v3.5.20230501 [71] to perform functional and pathway enrichment analysis. Genes related to hypoxia entries in the GO database were also extracted for comparison. SIFT (https://sift.bii.a-star.edu.sg/) (amino acids with the scores of < 0.05 are predicted to be deleterious) [72], PROVEAN (https://provean.jcvi.org/) (variants with a score equal to or below -2.5 are considered deleterious) [73] and SWISS-MODEL (https://swissmodel.expasy.org/) [74] were further used to predict the effects of mutations on gene function.

Convergence analysis

The PCOC software v1.1.0 [75] was used to identify convergent signals for all residues of homologous genes across several high-altitude species. Convergence at a site was characterized by a shift from the ancestral biochemical role to a new biochemical role shared by multiple convergent lineages. These biochemical roles were modeled as “profiles,” consisting of amino acid frequency vectors built from large empirical datasets. PCOC assigned sites to profiles based on the observed amino acid frequencies. If a site was inferred to have a prior selection in convergent branches that differed from the background under maximum likelihood, it was considered to exhibit profile change (PC). Additionally, the site had to show at least one substitution or variation (OC) in each convergent branch. The combination of both provided a single estimate of the posterior probability of convergence in PCOC. The fit of the model was assessed based on a null model where all branches belonged to the same profile. We set the posterior probability threshold for a residue’s convergent evolution in high-altitude species lineages at 0.8 [76]. We performed statistical power calculations for each gene following the procedure recommended on the official website, retaining results with a False Discovery Rate (FDR) no higher than 2% (https://github.com/CarineRey/pcoc) [77]. All identified convergent genes were subjected to enrichment analysis using Metascape [71].

Structural variation detection

In order to find the structural variations unique to Neodon through genome comparison, Minimap2 v2.23-r1111 [78] was first used to map the three chromosome-level genomes (M. ochrogaster, A. amphibius and C. griseus) and the genome of the closely related species (N. shergylaensis) to the N. fuscus genome assembly with parameters “-ax asm20 --cs-r2k”. The haploid mode of the software SVIM-asm v1.0.3 [79] was then used with the parameter“-min_sv_size 10” to find structural variations for each species, generating VCF files. SURVIVOR v1.0.6 [80] was used to merge these VCF files with parameters “1 3 1 10 10”. Only those that do not have structural variations in the N. shergylaensis genome, but are present in all three other species, are considered candidates for structural variations.

Validation of Map3k6

To verify the conservation of Map3k6 in Neodon, we downloaded a total of 16 resequencing datasets from the NCBI database, including resequencing data from C. griseus, A. amphibius, and M. ochrogaster, as well as resequencing data from 13 different species within the Neodon genus. The downloaded resequencing data were first trimmed using Fastp v0.23.2 [81] with default parameters. The trimmed reads of each species were mapped to the genome assembly of N. fuscus using BWA-MEM v0.7.17-r1188 and SAMtools v1.6 [47, 59]. PICARD v2.26.4 (https://broadinstitute.github.io/picard) was used to remove PCR-induced duplicate reads with the parameters “REMOVE_DUPLICATES = true VALIDATION_STRINGENCY = LENIENT.”

Transcriptome analysis

We obtained transcriptome data of heart tissues under normoxia and hypoxia conditions for N. fuscus and M. musculus from PRJNA993829 in NCBI, with three replicates for each condition. For these transcriptome data, we similarly performed filtering using Fastp [81] with default parameters. The HISAT2-StringTie-DESeq2 pipeline [82] was employed for downstream analysis of the transcriptome data. HISAT2 v2.2.1 [46] was used with default parameters to align the transcriptome reads to the N. fuscus genome or M. musculus genome (GRCm39). After sorting the alignments, StringTie v2.2.1 [83] was used to assemble transcripts and quantify gene expression levels with default parameters. Raw read counts were obtained using prepDE.py (https://github.com/gpertea/stringtie/blob/master/prepDE.py) and provided as input file to DESeq2 v3.21 [84] for differential expression analysis with default parameters. Genes with |log2(fold-change)|≥ 1 and adjusted P < 0.05 were designated as differentially expressed genes (DEGs).

Results

Genome assembly and gene annotation

For the assembly of a high-quality chromosome-level genome of N. fuscus, we employed approximately 136 Gb of Oxford Nanopore long reads. The initial draft genome assembly was 2.26 Gb in size, consisting of 751 contigs with an N50 of 25.37 Mb (Figure S1). Approximately 301 Gb of Hi-C reads successfully anchored 99.16% of the contigs to 27 chromosomes (26 autosomes and one X chromosome), with a scaffold N50 of 83.80 Mb, showing a highly contiguous genome (Fig. 1A; Tables S1-S3). To assess the assembly quality, we aligned transcriptome data to the assembled genome, achieving a mapping ratio of 97.53%−98.30%, indicating a highly complete genome (Table S4). The completeness of the assembled genome was further validated by the BUSCO test, which showed a completeness of 97.2% (Complete: 97.2%, Complete and single-copy: 95.7%, Complete and duplicated: 1.5%, Fragmented: 0.5%). To annotate the genes of N. fuscus, we integrated ab initio, homology-based, and RNA-seq based approaches, ultimately identifying 22,957 protein-coding genes (Fig. 1B). Of all the predicted genes, functional annotation based on the SwissProt and TrEMBL databases was able to identify 97.47% (22,377 genes) (Table S5). These results of functional annotation indicate that the majority of genes in N. fuscus have homologs in public databases, and the predicted gene models have high accuracy.

Fig. 1.

Chromosome-level assembly of Neodon fuscus. A Hi-C interaction map of the whole genome, showing 27 chromosomes of N. fuscus. B Circos plot illustrating the features of the N. fuscus genome. From I to V, it shows GC content, repeat sequence types (LTR, LINE, and SINE), and gene density. C Phylogenetic construction and divergence time estimation of seven species

Compared to the N. shergylaensis genome [35], the continuity (N50) of the N. fuscus genome has improved approximately seven-fold (10.85 Mb vs. 83.80 Mb), making it the highest-quality genome currently available for the Neodon genus. The genomic structural comparison between N. fuscus and N. shergylaensis revealed a highly conserved collinearity between their genomes (Figure S2). Together, these results indicate that our high-quality genome is complete and suitable for downstream analyses.

Evolution of repetitive sequences in Neodon

In the N. fuscus genome, we identified a total of 868.01 Mb repetitive sequences, comprising various transposable element types of LTR (11.84%), SINE (11.82%), LINE (7.51%), and DNA (1.14%) (Fig. 1B; Table S6 and S7). Our assessment of TE insertion times in seven species showed that Neodon underwent expansion of LTR and LINE elements such as ERVK, ERV1, and L1 elements after two Neodon species diverged from other species (Fig. 1C; Figure S3, 4). Notably, we observed a specific expansion of the DNA element ZISUPTON in N. fuscus. The lineage-specific expansion of ERVs in the Neodon genus may be related to the highly diverse viral landscape in these species [85].

Rapid evolution of high-altitude adaptation genes

High-altitude species generally exhibit specific adaptations to extreme conditions, such as strong ultraviolet radiation (UV), hypoxia, and low temperatures. To evaluate the evolutionary conservation of species within the genus Neodon, we employed OrthoFinder software to identify 19,024 orthologous gene families and 10,683 single-copy orthologs across six species from four distinct clades within the Arvicolinae and one outgroup from Cricetinae (Table S8). We constructed a data matrix of single-copy orthologs using the maximum likelihood (ML) method, resulting in a tree topology that is consistent with previous reports (Figure S4 C, D) [66]. Our results showed that N. fuscus and N. shergylaensis diverged 1.12 million years ago, while the divergence between Neodon and Microtini tribe occurred ~ 2.01 million years ago (Fig. 2A).

Fig. 2.

Genomic features related to high-altitude adaptation in Neodon species. A Phylogenetic tree of seven species, with gene family expansions and contractions indicated on each branch. B GO and KEGG enrichment results for gene families expanded in N. fuscus. The bottom left section shows the clusters of these genes marked by different colors, and the network on the right displays the PPI (protein–protein interaction) enriched by the products of these genes. Each MCODE represents a protein complex recognized by Metascape. C Collinearity analysis reveals that N. fuscus has the longest MHC gene region among the four chromosome-level genomes compared. Blue and green boxes represent genes within the homologous region of the mouse MHC. D-F The REG Atm plays a role in DNA damage repair in Neodon species, with shared mutations in the conserved FAT domain, potentially affecting Atm function. G-I Pdia4 affects DNA repair and inhibits apoptosis, with Neodon-specific mutations in conserved domains that may impact its function

In N. fuscus, we identified 292 expanded gene families, comprising 432 genes. Functional enrichment analysis detected several GO terms associated with high-altitude adaptation, such as positive regulation of sprouting angiogenesis (GO:1903672), retinoid metabolic process (GO:0001523) and regulation of tumor necrosis factor production (GO:0032680) (Fig. 2B; Table S9). Notably, among the expanded genes in N. fuscus, we found an expansion of genes related to antigen processing and presentation via MHC (Table S9). In fact, aside from these gene expansions, N. fuscus possesses the longest MHC region (0.77 Mb) and the highest number of genes (50 genes) among the four species with chromosome-level genomes used in our phylogenetic analysis (Fig. 2C; Table S10). In addition, unlike other chromosome-level genomes, N. fuscus has five copies of H2-EB1, whereas other genomes have one or none. Specifically, C. griseus has one copy, while A. amphibius and M. ochrogaster do not contain any copies of H2-EB1. H2-EB1 is thought to be associated with lung function and total lung capacity in mice [86]. The expansion of H2-EB1 genes in N. fuscus suggests an adaptive mechanism to maintain effective immune and lung function under hypoxic stress at high altitudes, potentially counteracting the challenges imposed by such environment.

To further investigate which genes in the Neodon species have undergone changes related to high-altitude adaptation, we calculated the synonymous and non-synonymous substitution rates of the 10,683 single-copy orthologs in the Neodon clade. We identified 559 positively selected genes (PSGs) and 710 rapidly evolving genes (REGs) (Table S11-S14). Similar to the expanded gene families in N. fuscus, these genes are highly enriched in high-altitude-adaptation-related processes, such as regulation of cold-induced thermogenesis (GO:0120161), sprouting angiogenesis (GO:0002040) and glutamine secretion and transport (GO:0014048, GO:0006868).

In the ancestral branch of Neodon, we identified two genes (Atm and Pdia4) related to damage repair and protein folding. Atm plays a crucial role in non-homologous end joining (NHEJ) and homologous recombination (HRR) [87, 88]. The cell cycle checkpoint kinase it encodes, which once phosphorylated can respond to DNA damage and maintain genomic stability, evolved rapidly in both Neodon species (Fig. 2D). The Atm gene has been reported to be related with the high-altitude adaptation in kiang [89]. We identified three missense mutations (A848V, K2404N, P2982L) in Atm (Fig. 2E). Among these three substitutions, the K2404N located in the conserved FAT domain, which would alter the charge and affect protein function (Fig. 2F). SIFT (score of 0.03) and PROVEAN (score of -−2.728) indicated that the K2404N substitution may have significant impact on protein function. Pdia4, located in the endoplasmic reticulum (ER), is responsible for protein folding, and its expression increases in response to ER stress (ERS) [90]. The ER generates ERS) under stress conditions such as hypoxia, leading to the accumulation of misfolded proteins. Increased expression of Pdia4 can induce platelet activation and participate in the ER response [91]. This process impacts DNA repair mechanisms and facilitates proper protein folding (Fig. 2G). In Pdia4, we detected a D533N substitution, which located in the conserved domain of Pdia4. The D533N substitution can lead to a change in protein charge and potentially affecting protein function, with a SIFT score of 0.03 (Fig. 2H, I). We also identified two high-altitude adaptation genes (Idh3b and Atp1b1) that are related to the reproductive system (see supplementary materials).

Convergent evolution of QTP species

As the highest plateau in the world, the QTP imposes significant selective pressures on native species, driving convergent evolution at the molecular level. We investigated two Neodon genomes, alongside other QTP lineages, to explore whether they exhibit shared signals of convergent evolution (Fig. 3A; Table S8). Although we did not find identical amino acid changes in these QTP species, we used the PCOC software to detect convergence across all protein residues and identified 363 genes with convergent signals in the two Neodon species, zokor, pikas, and wild yak (Fig. 3B; Table S15, 16). In the enrichment of these genes, we identified a set of significant overrepresented (corrected P < 0.05) GO and KEGG pathways related to the high-altitude adaptation, such as, DNA damage response (GO:0006974), DNA repair (GO:0006281), double-strand break repair (GO:0006302), DNA metabolic process (GO:0006259), glyoxylate and dicarboxylate metabolism (mmu00630), hypertrophy model pathway (WP202) etc., involving a set of genes (Nrp2, VaV3, Cat and Adam8) (Table S17). Notably, we observed transitions from ancestral amino acid profiles to convergent amino acid profiles at multiple sites within these genes in Neodon species and three other QTP species. These include one site in Nrp2 (596), one site in VaV3 (773), three sites in Cat gene (69, 423, 451) and five sites in Adam8 (535, 751, 776, 798, and 1048) (Fig. 3C, D; Figure S5). Under the PCOC model, these convergent sites exhibited high posterior probabilities. Nrp2, a downstream target gene of the hypoxia-inducible factor 2-alpha (EPAS1), shows reduced expression associated with a weakened hypoxic response [92, 93]. There is evidence supporting that Nrp2 interacts with vascular endothelial growth factor (VEGF), exhibiting convergent evolution in other high-altitude species [92, 93]. We incorporated RNA-seq data from heart tissue of N. fuscus and M. musculus under normoxia and hypoxia conditions to verify the role of Nrp2 in N. fuscus (Table S18). We found that under hypoxic conditions, the expression of Nrp2 was significantly reduced in mice, while no significant change was observed in N. fuscus (Fig. 3E). This suggests that Nrp2 also contributes to the hypoxic adaptation in N. fuscus. VaV3 promotes tumor growth, apoptosis, invasion and metastasis [94–96]. VaV3 is also known to play a crucial role in high-altitude adaptation in Ethiopian populations by contributing to angiogenesis [97]. Cat (catalase) is a key antioxidant enzyme essential for maintaining cellular redox balance, particularly in countering reactive oxygen species (ROS). Additionally, Cat as an antioxidant enzyme, is vital in maintaining redox balance within cells. Under hypoxic conditions, increased oxidative stress leads to heightened ROS production, and Cat helps cells survive by scavenging ROS [98]. Adam8 is strongly linked to various tumors, where it is highly expressed. In addition, Adam8 promotes angiogenesis in the hypoxic condition [99–101].

Fig. 3.

Convergent evolution in Neodon and other QTP species. A Phylogenetic relationships of selected species, with QTP species highlighted in yellow. B GO and KEGG enrichment results for genes showing convergent evolutionary signals. C-D Results of three high-altitude adaptation genes identified by PCOC, with amino acids showing convergent signals in QTP species marked with black boxes. The numbers above each column indicate convergent amino acid sites, with posterior probabilities from different methods (PCOC_v1, PC_v1, OC_v1). The yellow lines represent the QTP species and the blue represents the other species. E Comparative analysis of Nrp2 expression changes in heart tissue under normoxia and hypoxia between N. fuscus and M. musculus. N. fuscus exhibits nondifferential expression, whereas M. musculus shows downregulation. Species abbreviations: Asa: Loxodonta africana. Eca: Equus caballus. Bmu: Bos mutus. Fca: Felis catus. Cfa: Canis lupus familiaris. Ppa: Pan paniscus. Has: Homo sapiens. Ocu: Ochotona curzoniae. Eba: Eospalax baileyi. Cgr: Cricetulus griseus. Aam: Arvicola amphibius. Moc: Microtus ochrogaster. Nsh: Neodon shergylaensis. Nfu: Neodon fuscus

Map3k6 may play a role in maintaining angiogenesis rate in Neodon

Structural variants (SVs) are believed to play an important role in species’ adaptation to their environment [102]. Recent research on zokors revealed that SVs contribute to physiological responses in hypoxic environments, where chromosomal rearrangements and intragenic SVs alter the expression of key genes involved in hypoxia response [25]. To investigate which genes might be affected by SVs during the high-altitude adaptation of Neodon species, we used N. fuscus as the reference genome and compared it with N. shergylaensis alongside three other chromosome-level genomes to identify Neodon-specific structural variants (Fig. 4A; Table S19). We identified a total of 4,169 SVs, including 3,235 insertions and 934 deletions, which are intersect with 1,493 genes. The enrichment analyses identified several pathways linked to high-altitude adaptation, including those involved in the morphogenesis and development of the heart and blood vessels (Table S20). These pathways also include a set of previously reported genes (Table S21).

Fig. 4.

Neodon-specific SVs. A Strategy for identifying Neodon-specific SVs. B Map3k6 antagonizes the function of HIF-1 in VEGF regulation. C Comparative Map3k6 expression changes in heart tissue under normoxia and hypoxia between N. fuscus and M. musculus, with N. fuscus showing downregulation, and M. musculus showing upregulation. D A 24 bp deletion in exon 26 of Map3k6 is present in two Neodon species but absent in the other three chromosome-level genomes. E Using C. griseus genome as the reference, we mapped the resequencing data of C. griseus, A. amphibius, M. ochrogaster, and the 13 Neodon species to assess the sequencing coverage of SV in the homologous region of C. griseus. Resequencing data from 13 Neodon species show consistent patterns in the SV region. The non-gray bars represent the SNPs of that species at the respective positions. Cgr: Cricetulus griseus. Aam: Arvicola amphibius. Moc: Microtus ochrogaster. Nsh: Neodon shergylaensis. Nfu: Neodon fuscus

Notably, there are 81 genes containing Neodon-specific SVs that were also identified as PSGs and REGs (Table S22). Among these two gene sets, Arnt2 has been previously reported to be associated with hypoxia adaptation [97]. It plays an essential role in the nervous system’s response to hypoxia by forming heterodimers with HIF-1α, which mediate transcriptional regulation during hypoxic conditions [103, 104]. Additionally, differential expression analysis of heart transcriptomes in N. fuscus and M. musculus revealed that 21 of these 81 SV genes exhibited altered expression. Among these 21 genes, only Map3k6 had an SV overlapping with an exon. A 24 bp deletion was identified in exon 26 of this gene (Table S19, S23). The decrease of Map3k6 expression can significantly inhibit the tumor growth, vessel formation and the VEGF expression [105] (Fig. 4B). To validate the role of Map3k6 in N. fuscus, we obtained heart transcriptomic data of N. fuscus and M. musculus from public databases (Table S18). We found under normal and hypoxic conditions, N. fuscus and M. musculus displayed opposite expression patterns: N. fuscus showed a significant downregulation of Map3k6 expression, while M. musculus showed a notable upregulation (Fig. 4C; Table S23).

Alignments of five genomes visually demonstrated the structural variation present in Map3k6 (Fig. 4D; Figure S6). To determine whether the 24 bp deletion at Map3k6 found in the two Neodon genomes is shared across all Neodon species, we used C. griseus genome as the reference and resequencing data from 13 different species within the genus to validate this SV (Table S18). The sequencing coverage results indicate that this region shows a consistent deletion across all 13 species of Neodon, while it is present in C. griseus, A. amphibius, and M. ochrogaster (Fig. 4E). This supports the hypothesis that this SV in Map3k6 gene was occur in the ancestor of the Neodon genus, correlating with the origin of Neodon species on the QTP [26, 35]. We hypothesize that this ancestral SV is responsible for the change in Map3k6 expression, but future experimental validation is needed.

Discussion

In this study, we sequenced and assembled the chromosome-level genome of N. fuscus using Nanopore and Hi-C sequencing technologies. Species within the genus Neodon represent a unique group of high-altitude mammals outside of primates and large animals [14, 16, 17, 19, 20]. Investigating the genetic mechanisms underlying high-altitude adaptation in these species can enhance our understanding of how diverse species respond to extreme environments. This high-quality genome assembly provides significant insights into Neodon, contributing to our understanding of the biological mechanisms underlying high-altitude adaptation.

Analysis of TEs in N. fuscus and N. shergylaensis revealed unique compositional proportions of LINE elements (N. fuscus: ~ 12%; N. shergylaensis: ~ 11%) and LTR elements (N. fuscus: ~ 8%; N. shergylaensis: ~ 8%), which differ from those in other mammals [106]. A similar pattern was observed in deer mice, which host various exogenous retroviruses. Exogenous retroviruses (ERVs) and LINEs compete for genomic integration sites [107]. ERVs often outcompete LINEs for these sites, frequently disrupting autonomous LINE copies, directly influencing the evolutionary trajectory of LINEs. The greater difference in the ratios of LTRs and LINEs in N. fuscus and N. shergylaensis compared to deer mice (LTR: ~ 11% vs. LINE: ~ 10%) may indicate enhanced competition between LTRs and LINEs.

Unlike other species that migrate to the QTP, Neodon has a much longer history of high-altitude adaptation [17, 35, 108–111]. Comparative genomic analyses of N. fuscus, N. shergylaensis, and five other high-quality genomes from different tribes revealed that expanded gene families, PSGs, and REGs in species within the genus Neodon are associated with DNA damage repair, energy metabolism, immune response and reproductive capability. Among the gene families expanded in N. fuscus, we identified expansion of MHC genes, which has also been reported in other species with strong immune responses [112, 113].

The immune and/or reproductive abilities of native species in high-altitude environments with low oxygen, low temperature, and strong UV radiation show significant changes compared to lowland species, and these changes are closely related to the altitude of the environment [114–117]. For example, compared to before high-altitude migration, high-altitude-adapted sheep show significant changes, with tissues involved in short-term high-altitude adaptation primarily associated with immune system and rapid energy metabolism regulation. Additionally, the offspring mortality rate of ewes adapted to hypoxia at high altitudes significantly decreases, indicating that maternal hypoxia adaptation may enhance offspring tolerance to hypoxia [118]. In this study, many expanded gene families, PSGs, and REGs were significantly enriched in immunity-related items, eg, immune response, inflammatory response and tumor regulation. Notably, we identified that MHC genes underwent significant expansion. The expansion of immune genes has been reported in many high-altitude species [112, 113, 119, 120], which indicated an adaptive mechanism to maintain effective immune function under hypoxic stress at high altitudes, potentially counteracting the challenges imposed by such environment.

High-altitude hypoxia environments can also negatively impact sperm concentration, function, testosterone levels, and ovarian function in populations from low-altitude regions [121–123]. However, high-altitude populations appear to have evolved strategies to mitigate these effects. For instance, Sherpa high-altitude populations exhibit higher levels of follicle-stimulating hormone and testosterone, which may play a crucial role in maintaining reproductive function under hypoxic conditions [124]. In Andeans, PRKAA1 undergoes natural selection, which is associated with normal fetal birth weight in high altitude regions, thereby maintaining fertility [125]. Similarly, Tibetans have evolved genetic mechanisms to protect fetal development in high altitude environments [126]. We identified two genes, Idh3b and Atp1b1, in Neodon related to sperm development and cold tolerance. Given that high altitude conditions can impair sperm concentration and quality in non-native individuals, Neodon may ensure normal reproductive capacity by improving male sperm quality and cold tolerance [121, 122, 127]. This adaptation likely represents a strategy for males to maintain reproductive success during long-term adaptation to the plateau environment.

Many high-altitude species have evolved cancer resistance mechanisms as a response to various environmental stressors. In the zokor, the tumor-associated p53 gene has undergone mutations, enabling the species to adapt to the harsh conditions of the QTP [128]. In another QTP-dwelling animal, Anan’s rock agama, mutations in the TIE2 gene are believed to contribute to suppressing tumor growth [129]. These adaptations underscore the selective pressures imposed by high-altitude environments and the evolutionary strategies species have developed to mitigate the risk of cancer while surviving in such extreme conditions. Our convergent evolution analysis of QTP species identified several intriguing genes, such as Vav3 and Adam8, which promote angiogenesis under hypoxic conditions [97, 99, 100]. In the high-altitude adaptation of Tibetan goats, numerous genes are enriched in the VEGF signaling pathway and melanoma-related pathways [130]. In Anan’s rock agama, TIE2 is also thought to play a crucial role in angiogenesis and tumor suppression [131]. Both Vav3 and Adam8 are strongly associated with tumor progression and angiogenesis, suggesting that directly altering the functions of tumor-related genes may represent a cost-efficient adaptive evolutionary strategy common among QTP species. Studies of different populations have found that cancer incidence and mortality rates are lower among high altitude populations in the United States and China, but higher in high altitude residents of Ecuador and India [132–135]. This ability to adjust the balance between tumor progression and angiogenesis may play a role in the low cancer incidence and mortality observed in the Chinese populations.

Previous comparative transcriptomic study on N. fuscus found significant downregulation of genes in fatty acid oxidation pathways and upregulation of the core glucose oxidation gene in its heart under hypoxic conditions, indicating a switch in metabolic substrates from fatty acids to glucose [33]. This hypoxia adaptation strategy has also been reported in high-altitude Andean mice and Sherpas [23, 136]. However, this strategy is not without its drawbacks. On one hand, elevated blood glucose levels can lead to atherosclerosis or accelerate heart failure, while also promoting VEGF expression and angiogenesis, thus accelerating tumor progression [137–139]. On the other hand, the preference shift from fatty acid oxidation to glucose utilization is believed to impair the failing heart [140]. We observed a significant downregulation of Map3k6 under hypoxic conditions in N. fuscus, which contrasts with the situation in M. musculus, where it is thought to favor the regulation of angiogenesis [105]. We propose that this lineage-specific structural variation represents an alternative adaptive strategy for regulating the rate of angiogenesis in Neodon as part of its long evolutionary history of adaptation to high-altitude environments. However, further experiments are needed to validate this hypothesis.

Conclusion

In conclusion, this study identifies key genes in N. fuscus and demonstrates their critical roles in facilitating the adaptation of Neodon species to high-altitude stress, offering a valuable reference for future research on plateau-dwelling rodents.

Abbreviations

BUSCO

Benchmarking Universal Single-Copy Orthologs

TEs

Transposable elements

GO

Gene Ontology

KEGG

Kyoto Encyclopedia of Genes and Genomes

Mya

Million years ago

NR

Non-redundant protein

BWA

Burrows-Wheeler aligner

PSGs

Positively selected genes

REGs

Rapidly evolving genes

SVs

Structural variations

Authors’ contributions

G.L. and T.S. designed research; J.Q., T.F., G.C., R.D., L.X. and D.M. provided the samples; P.Y. performed the genome assembly, comparative genomics analysis, convergent analysis and structure variation analysis; P.Y., Y.W. and Z.X. performed the genome annotation; P.Y. wrote the raw manuscript; T.S., G.L. and Y.R. revised the manuscript. All authors read and approved the final version of the manuscript.

Funding

Foundation items: This work was supported by National Natural Science Foundation of China (32200337) and the fellowship of China Postdoctoral Science Foundation (2022M712003) to T.S., National Natural Science Foundation of China (32470445) to G.L., the Postdoctoral Fellowship Program of CPSF under Grant Number GZC20240965 to L.X., Fundamental Research Funds for the Central Universities (Grant No. GK202304017) to Y.R. and Doctoral exploration project of Shaanxi Normal University (2021TS065) to P.Y.

Data availability

Both raw sequenced sequences under the BioProject accession number PRJNA1172575 and the reference genome assembly (accession number: JBIEOQ000000000) of the N. fuscus have been deposited and released in the NCBI database. Codes for this manuscript can be found in https://github.com/GanglabSnnu/Comparative_analysis.

Declarations

Ethics approval and consent to participate

All samples were handled following the guidelines in accordance with the Regulations for the Administration of Affairs Concerning Experimental Animals approved by the State Council of the People’s Republic of China. The experimental protocols were approved by the Animal Care and Use Committee of Shaanxi Normal University (Ethic approval No: 2022–039).

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.