Genomic structural variation is associated with hypoxia adaptation in high-altitude zokors

Xuan An; Leyan Mao; Yinjia Wang; Qinqin Xu; Xi Liu; Shangzhe Zhang; Zhenglei Qiao; Bowen Li; Fang Li; Zhuoran Kuang; Na Wan; Xiaolong Liang; Qijiao Duan; Zhilong Feng; Xiaojie Yang; Sanyuan Liu; Eviatar Nevo; Jianquan Liu; Jay F Storz; Kexin Li

doi:10.1038/s41559-023-02275-7

. Author manuscript; available in PMC: 2025 Jul 27.

Published in final edited form as: Nat Ecol Evol. 2024 Jan 9;8(2):339–351. doi: 10.1038/s41559-023-02275-7

Genomic structural variation is associated with hypoxia adaptation in high-altitude zokors

Xuan An ^1,⁶, Leyan Mao ^1,⁶, Yinjia Wang ^1,⁶, Qinqin Xu ^2,⁶, Xi Liu ¹, Shangzhe Zhang ¹, Zhenglei Qiao ³, Bowen Li ¹, Fang Li ³, Zhuoran Kuang ¹, Na Wan ¹, Xiaolong Liang ¹, Qijiao Duan ¹, Zhilong Feng ¹, Xiaojie Yang ¹, Sanyuan Liu ¹, Eviatar Nevo ⁴, Jianquan Liu ^1,^✉, Jay F Storz ^5,^✉, Kexin Li ^1,^✉

PMCID: PMC12301030 NIHMSID: NIHMS2090629 PMID: 38195998

Abstract

Zokors, an Asiatic group of subterranean rodents, originated in lowlands and colonized high-elevational zones following the uplift of the Qinghai–Tibet plateau about 3.6 million years ago. Zokors live at high elevation in subterranean burrows and experience hypobaric hypoxia, including both hypoxia (low oxygen concentration) and hypercapnia (elevated partial pressure of CO₂). Here we report a genomic analysis of six zokor species (genus Eospalax) with different elevational ranges to identify structural variants (deletions and inversions) that may have contributed to high-elevation adaptation. Based on an assembly of a chromosome-level genome of the high-elevation species, Eospalax baileyi, we identified 18 large inversions that distinguished this species from congeners native to lower elevations. Small-scale structural variants in the introns of EGLN1, HIF1A, HSF1 and SFTPD of E. baileyi were associated with the upregulated expression of those genes. A rearrangement on chromosome 1 was associated with altered chromatin accessibility, leading to modified gene expression profiles of key genes involved in the physiological response to hypoxia. Multigene families that underwent copy-number expansions in E. baileyi were enriched for autophagy, HIF1 signalling and immune response. E. baileyi show a significantly larger lung mass than those of other Eospalax species. These findings highlight the key role of structural variants underlying hypoxia adaptation of high-elevation species in Eospalax.

Animals native to high elevations are forced to cope with major environmental stressors, including low temperature, strong ultraviolet exposure and reduced oxygen availability¹. Genomic studies of highland humans and non-human mammals have suggested candidate genes for hypoxia adaptation^1–13, and experimental physiological studies have identified specific mechanisms of adaptation^14–16. However, much less is known about the extent to which structural variation in the genome has contributed to high-altitude adaptation.

Structural variants (SVs) can affect the orientation and linkage arrangements of genes, thereby potentially altering patterns of gene expression¹⁷. Structural changes can also affect gene expression by altering the distance between cis-acting regulators and their targets and by modifying chromatin accessibility to nuclear macromolecules^18,19. Moreover, SVs can affect gene expression and regulation by altering three-dimensional (3D) genomic structures, including active (A) or inactive (B) compartments, topologically associating domain (TAD) boundaries and chromatin loops. Large chromosome inversions, mediated mainly through non-homologous end joining and non-allelic homologous recombination (NAHR), are known to contribute to adaptive phenotypic change^20–22, and they can suppress recombination, thereby helping to maintain co-adapted combinations of alleles at linked loci (supergenes)^23–25. Recent studies have documented that SVs contribute to variation in physiological responses to hypoxia, thereby affecting susceptibility to altitude-related pathologies such as pulmonary oedema and cerebral hypoxia¹⁹.

Zokors (genus Eospalax, Spalacidae: Rodentia) live most of their lives underground, where they have to cope with hypoxia and hypercapnia in closed burrow systems. O₂ and CO₂ levels in sealed tunnels have been recorded at 7.2% and 6.1%, respectively, during the rainy season when wet soil further impedes gas exchange with the outside air²⁶. Eospalax comprises six East Asian species that have well-documented phylogenetic relationships²⁷ (Fig. 1a,b). With the uplift of the Qinghai–Tibet Plateau (~3.6 million years ago (Ma)), Eospalax split into the high-altitude (including E. baileyi, 2,800–4,600 m, and E. smithi, 2,500–3,300 m) and low-altitude clades at ~3.22 Ma. E. baileyi and E. smithi split from their ancestor at ~1.18 Ma and 0.75 Ma, respectively, due to the generation of the Yellow River and a glaciation. For the low-altitude lineage, four Eospalax species have elevational ranges under 2,700 m. E. fontanierii speciated first, E. cansus diverged northwards from the ancestor at ~1.43 Ma, whereas E. rothschildi speciated southwards because of a later glaciation (Fig. 1a). Comparisons between highland and lowland species provide an opportunity to explore genetic changes that may have contributed to high-elevation adaptation.

Fig. 1 | — a, Phylogenetic trees show the relationship and elevation ranges of *Eospalax*. The bar shows the partial pressures of oxygen (pO₂) at each elevation. b, Primary subject of this study, *E. baileyi* (plateau zokor). Credit: Xuan An, Lanzhou University. c, Comparison of the relationship between lung mass and body mass among *Eospalax*. The linear regression line is shown as a solid red line and the 95% confidence intervals are shown as pink shadows. d,e, Difference in RBC counts (10¹² l⁻¹) (d) and blood Hb (g l⁻¹) (e) among *E. baileyi*, *E. rufescens* and *E. cansus*. There are also differences in *E. baileyi* populations at different elevations. The numbers in parentheses are sample sizes. In the box plots, the lower and higher hinges correspond to the first and third quartiles, respectively. The median is represented by the line inside the box. The whisker extends from the hinge to at most 1.5 × interquartile range (IQR). The other points are outliers. The difference in RBC and Hb supporting different populations was tested using a two-sided t-test. f, Functional enrichment of expanded-family genes (pink) and PSGs (green) in high-altitude *E. baileyi*. The size of the round shape represents the number of genes in this enrichment term. The PSGs (likelihood ratio test, P < 0.05, FDR < 0.05) and expanded family genes (likelihood ratio test, P < 0.05, FDR < 0.05) were identified with multiple corrections.

Here we investigated the role of SVs (deletion and inversion) in hypoxia adaptation of high-elevation species in Eospalax. We used a long-read sequencing approach to obtain a chromosome-level genome assembly of E. baileyi, and we assessed species-level variation using a long-read resequencing approach. The effects of SVs on hypoxia-related genes were examined using RNA sequencing (RNA-seq) and reporter assays. We also investigated TAD rearrangements and A/B compartment transitions to assess how evolved changes in 3D genomic architecture may have contributed to hypoxia adaptation.

Results

Phenotypic variation among zokors native to different elevations

We measured several phenotypes of relevance to high-altitude acclimatization and/or adaptation: the relationship between lung mass and body mass, red blood cell (RBC) counts, and blood haemoglobin concentration ([Hb]). The high-elevation E. baileyi showed a larger lung mass compared with other Eospalax species (Fig. 1c), which can be expected to enhance pulmonary diffusing capacity for gas exchange¹. RBC counts and [Hb] were also higher in E. baileyi relative to E. cansus and E. rufescens, suggesting a higher O₂-transport capacity per unit of cardiac output. In E. baileyi, RBC counts and [Hb] increased as a positive function of native elevation (Fig. 1d,e), consistent with expected acclimatization responses to chronic hypoxia²⁸.

Genome assembly, scaffolding and annotation

We generated the first chromosome-level genome assembly of E. baileyi. The initial assembly consisted of ~209.0 Gb of long reads and was subsequently corrected using ~255.1 Gb of Illumina paired-end reads (Extended Data Table 1). About 99.8% of corrected contigs were anchored onto 31 pseudochromosomes (2n = 62) through Hi-C data (Supplementary Table 1). This is consistent with the fibroblast karyotype analysis (Extended Data Fig. 1a,b,e). The finally assembled genome size was 2.47 Gb. Compared with the previously published scaffold-level assembly of the plateau zokor (v.2.1)³, the contig N50 increased from 323 kb to 38.0 Mb, the scaffold N50 increased from 6.7 to 82.3 Mb, and gaps decreased from 1.64% to 0.01% (Table 1). The completeness of the assembly, tested using Benchmarking Universal Single-Copy Orthologs (BUSCO), was 96.8%, indicating high precision, continuity and integrity. The proportion of genomic sequence consisting of repeats was 44.4%, where short interspersed nuclear elements (SINEs) and long interspersed nuclear elements (LINEs) accounted for 18.1% and 22.2%, respectively (Supplementary Table 2). We annotated 21,642 protein-coding genes, consistent with the annotation results for the genome of the closely related species E. fontanierii²⁷. The BUSCO evaluation of the annotation completeness showed that 95.6% of genes were predicted in this E. baileyi assembly (Extended Data Fig. 1c,d).

Table 1 |.

Comparison of assembly statistics between Plateau zokor v.2.1 and Plateau zokor v.3.0

Plateau zokor	Plateau zokor v.2.1	Plateau zokor v.3.0
Sequencing platform	Illumina, Pacbio (10x)	ONT, Hi-C
Assembly size (bp)	2,569,336,452	2,471,808,144
Total length of chromosome (bp)	–	2,385,578,628
Number of contigs	42,044	758
Contig N50 (bp)	322,513	37,952,775
Number of scaffolds	14,401	465
Scaffold N50 (bp)	6,701,949	82,317,615
Number of gaps	102,445	293
Total gap length (bp)	42,024,904	124,000
Ratio of gap (%)	1.64	0.01
GC content (%)	41.20	41.26
BUSCO (%)	93.30	96.80

Open in a new tab

Genome evolution and possible mechanisms of hypoxia adaptation

We estimated a phylogenetic tree and species divergence time of 9 rodent species based on 9,258 single-copy orthologues (Extended Data Fig. 1f). Compared with the other 8 rodents, we identified 469 gene families in the E. baileyi lineage that underwent significant expansions relative to the expectations of a null model of birth-and-death evolution (likelihood ratio test, P < 0.05). Functional enrichment analysis of the expanded gene families revealed an over-representation of the following terms: hypoxia-inducible factor (HIF) signalling pathway, DNA damage repair, autophagy, cell cycle, selenocysteine synthesis, purine and retinol metabolism, recruitment of nuclear mitotic apparatus (NuMA) to mitotic centrosomes, immune responses and nutrition functions (Fig. 1f and Supplementary Table 3). We also identified a total of 1,609 gene families that underwent significant contractions in the E. baileyi lineage. These genes were related to cell cycle, signalling by WNT, metabolism and meiotic recombination. Using the adaptive branch-site random effects likelihood (aBSREL) model, we tested for accelerated rates of non-synonymous substitution across all genes in the branch leading to E. baileyi and identified 71 lineage-specific positively selected genes (PSGs) (omega > 1, likelihood ratio test, P < 0.05, false discovery rate (FDR) < 0.05; Fig. 1f and Supplementary Tables 3 and 4). These PSGs were enriched for the following terms: double-strand break repair, circadian rhythm and ubiquitin-dependent protein catabolic functions.

SV identification, validation and distribution

Distribution of <1 Mb SVs in Eospalax genomes.

Using long-read sequencing, we identified 398,087 uniquely located SVs in a sample of 2 or 3 individuals from each Eospalax species, including 218,607 deletions (DEL), 168,333 insertions (INS), 5,183 duplications (DUP), 5,062 inversions (INV) and 902 translocations (TRA) (Fig. 2a and Supplementary Table 5). The frequencies of these SVs were estimated from 146 individuals, including 55 E. baileyi, 25 E. cansus, 21 E. smithi, 18 E. rothschildi, 14 E. rufescens and 13 E. fontanierii (Extended Data Fig. 2a). We investigated the phylogenetic distribution of SVs across Eospalax and identified those that were species specific (Fig. 2b). Most of the INDELs were less than 300 bp, and there was another small peak at ~400 bp, which was mainly influenced by SINEs (Extended Data Fig. 2b,c). The majority of SVs (53.63%, 296,873) overlapped with simple repetitive elements and transposable elements. For transposable elements, SVs were enriched in regions with SINEs (35%) and long terminal repeats (LTRs, 22.8%). Most SVs (69.16%, 275,307) were distributed in intergenic regions. Meanwhile, 122,780 (30.84%) SVs were found within 68.15% (14,758 out of 21,656) of the genic regions (that is, in exons, introns, and 5ʹ and 3ʹ flanking regions within 2 kb of start and stop codons, respectively). Intronic SVs accounted for 28.25% of all SVs, and only 5.11% of SVs affected coding sequences (Extended Data Fig. 2d,e).

Large inversions (>1 Mb) in the E. baileyi genome.

We identified 18 inversions longer than 1 Mb in E. baileyi (Fig. 2c,d and Extended Data Fig. 3) based on the chromosome-level genome assembly. These large inversions were validated by aligning the breakpoint-spanning contigs of E. baileyi to the genome of E. fontanierii (Extended Data Fig. 4). Accurate breakpoints (Supplementary Table 6) were identified in three ways (Methods). We identified 81 telomere-enriched regions, with 60 appearing at expected chromosome ends. We found 11 inversions located at or near the chromosome termini (Fig. 2c), with most of the breakpoints near the telomere-enriched regions. Seven of these inversions contained telomeric regions, suggesting that the inversions probably altered the position of the telomeres.

Possible functional effects of SVs

Differentially expressed genes associated with SVs (<1 Mb).

To evaluate whether fixed SVs could potentially affect the expression of associated genes, we estimated the expression of all 3,748 fixed SV-associated genes using RNA-seq from animals after common garden experiments (Methods). Compared with E. fontanierii, we found 253, 187 and 122 differentially expressed genes (DEGs) in the lung, liver and heart of the high-altitude E. baileyi, respectively, accounting for 6.75%, 5.00% and 3.26% of all fixed SV-associated genes (Supplementary Tables 12–14). The DEGs in the lung were associated with immunity, inflammatory response, cell adhesion and cell developmental cycle (ILDR2, IL20, CDON, LAMB3, IL2RB, OPCM, RAB3C and NEK10). The DEGs of TBX20 and KCND2 in the lung contained intronic SVs and were involved in cardiac development, conduction and diseases. In the liver, differentially expressed SV-associated genes were involved in haematopoiesis, vascular development, platelet activation and vascular endothelial growth factor (GLRX5, CD109, NRP1 and ADTRP). Notably, immunity-related genes of SFTPD and NRP1, both of which contained intronic SVs, were significantly upregulated in the livers of E. baileyi. We found a number of differentially expressed SV-associated genes in the heart significantly enriched for the HIF1 pathway function (NOS1, NQO1, HIPK2 and BMP7), and genes related to cell adhesion and leucocyte proliferation (CD1D, NCAM2 and CDC26) were also upregulated in the high-altitude E. baileyi (Extended Data Fig. 5a–d).

Functional effects of SVs (<1 Mb).

SVs specific to the two species with the highest range limits, E. smithi and E. baileyi, were located in 345 genes (that is, in exons, introns, or 5ʹ and 3ʹ flanking regions within 2 kb of start and stop codons), which were enriched in cellular response to hypoxia, DNA damage and repair, autophagy, heart morphogenesis and contraction, lung epithelium development, response to viruses, and regulation of vascular-associated smooth muscle cell migration. These same pathways have been implicated in hypoxia adaptation in other highland mammals^29,30 (Extended Data Fig. 5e and Supplementary Table 7).

We detected an intronic SV in EGLN1 with both heterozygous inversion (NI) and homologous inversion (II) existing only in E. baileyi (1,442 bp, LG01:1,889,355–1,890,797). Frequencies of inversion were 100% present at a frequency of 100% in animals above 4,300 m (n = 6) and 84.7% in those from <4,300 m (n = 49) (Fig. 3a). DEGs analyses showed that the E. baileyi EGLN1 allele with intronic inversion had a significantly higher expression compared with low-altitude species in Eospalax (Fig. 3b), and the dual-luciferase reporter assay (DLRA) confirmed that the inversion significantly increased luciferase activity (Fig. 3c and Extended Data Table 2).

A homozygous intronic inversion (2,212 bp, LG13: 44,667,409–44,669,626) inside HIF1A was fixed in E. smithi and E. baileyi (Fig. 3d and Extended Data Fig. 6a). The inversion allele in E. baileyi showed upregulated gene expression (Fig. 3e). As with EGLN1, the inversion within the HIF1A intron also resulted in significantly higher expression using DLRA (Fig. 3f and Extended Data Table 2). An intronic deletion in the HSF1 was identified (1,849 bp, LG09: 81,333,952–81,335,801) (Fig. 3g) and confirmed by polymerase chain reaction (PCR; Fig. 3h). DLRA suggested that the presence of this deletion significantly increased the HSF1 expression (Fig. 3i).

SVs were also detected in three genes related to DNA repair and immunity. A homozygous 12,732 bp deletion that included an intron and the second exon of SFTPD in E. baileyi resulted in a deletion of 55 amino acids from the N terminus of the protein. Protein structure predictions of E. fontanierii and E. baileyi suggested that the deletion resulted in missing α-helix and β-fold in E. baileyi (Extended Data Fig. 6b). We also detected a 9,280 bp intronic deletion in XRCC4 in E. baileyi and a 1,984-bp intronic deletion in UVRAG of E. smithi and E. baileyi, which were validated through read coverage and long-read sequence alignment (Extended Data Fig. 6c,d).

Limited deleterious effects of large inversions.

We found that 7 of the 36 large inversion breakpoints occurred in genic regions, with possible consequences for gene function and gene expression. For example, the breakpoints of LG01.1 disrupted the SPS1 gene, which has been reported to be involved in selenocysteine synthesis. Breakpoints for LG02.1, LG06.1, LG11.1 and LG17.1 disrupted the P2R2B, CB016, CFA91 and DLG5 genes, respectively. Breakpoints may also affect gene expression by disrupting regulatory elements (Fig. 4a). Inversions were not disproportionately associated with loss-of-function single nucleotide polymorphism (LOF SNP) (Fig. 4b and Supplementary Table 8). Some large inversions potentially rearranged TAD boundaries, which could alter interactions between genes/regulators. We found that most inversion breakpoints occurred near TAD boundaries (Fig. 4c).

Fig. 4 | — a, Distance from inversion breakpoints to the transcriptional start sites of the nearest genes. Seven (19%) of these breakpoints disrupted genes. b, Proportion of LOF SNP in inversion/non-inversion regions compared with that in the respective genome-wide regions. A one-sample two-sided t-test was used (n = 18 inversions). In the box plots, the centre line represents the median, whereas box boundaries represent the upper and lower quartiles. The whisker reaches no farther than 1.5 × IQR. c, Distance from the nearest TAD boundary to each inversion breakpoint in the *E. fontanierii* genome. d, Phylogenetic trees are constructed based on the long-read-SV dataset. The previously reported divergence time of species was plotted on phylogenetic trees. The synteny comparison showed that the rearrangement types on chromosome 1 (0–50 Mb) of *E. cansus*, *E. rothschildi*, *E. rufescens*, *E. baileyi* and *E. smithi* aligned with that of *E. fontanierii*, respectively. The terminal region with the *EGLN1* gene (red arrow) was inverted to the middle of the chromosome arm in the three high-altitude species and the humid-region species. e, Comparison of A/B compartments between *E. fontanierii* and *E. baileyi* on chromosome 1. The background represents the regions of chromosome 1, excluding the area of rearrangement. ‘A–B’ means the homologous bins were compartment A in *E. fontanierii* but compartment B in *E. baileyi*. ‘B–A’ was the opposite. ‘Stable’ means the homologous bins in two genomes had the same compartment type. f, TAD distribution in the 21.6 Mb region of chromosome 1. Distribution of genes associated with adaptation to hypoxia in the inversion region of chromosome 1 in the *E. fontanierii* genome. The shaded area contains three consecutive TADs in the *E. fontanierii* genome, and the homologous bin appears as a single TAD in the *E. baileyi* genome. After TAD fusion, *TSNAX* and *DISC1* were located in the same TAD. g, Comparison of the significant cis interaction distance between *E. fontanierii* (n = 1,002) and *E. baileyi* (n = 1,771) in the rearrangement region. A two-sided t-test was used. In the box plots, the centre line represents the median, whereas box boundaries represent the upper and lower quartiles. The whisker reaches no farther than 1.5 × IQR.

Rearrangement on chromosome 1 in Eospalax.

Compared with the E. fontanierii genome, we noticed a 50 Mb chromosomal rearrangement involving two parts with lengths of 21.6 Mb and 28.4 Mb. The chromosomal rearrangement was observed in each of the four possible homozygous or heterozygous genotypic combinations (Fig. 2d). The two lowest-altitude species of E. fontanierii and E. cansus showed the same genotype I. Although very close to E. rufescens, the humid-region species of E. rothschildi showed a specific genotype II. The third highest-altitude species of E. rufescens showed a unique genotype III, and the two highest-elevation species showed the same genotype IV (Fig. 4d). Notably, repeated elements were present around their breakpoints. Remarkably, the hypoxia-relevant EGLN1 gene located in this region shifted from the chromosome terminal to the middle of the chromosome arm through inversions in all three variant genotypes (genotypes II, III and IV) in species from the top four highest elevations (Fig. 4d). To further verify the frequency of the rearrangement, we identified the breakpoint in three unrelated individuals per species on chromosome 1 (Supplementary Fig. 2). Interestingly, the species corresponding to genotypes II–IV lived in more hypoxic environments caused by higher elevation or greater soil humidity. Remarkably, most genes located in large inversions, such as EGLN1, have been reported to be involved in hypoxia adaptation, DNA damage repair, inflammation, immunity, apoptosis, platelets and vascular development (Fig. 4f). The number of genes located in the remaining 17 inversion regions ranged from 35 to 384.

To further clarify the potential impact of SV on the 3D chromatin architecture at compartment A/B, TADs and significant interaction levels (loop), we constructed a chromosome interaction matrix at different resolutions with homologous bins across two genomes. The GC content and gene density in compartment A were significantly higher than in compartment B (Extended Data Fig. 7a). Although the compartment types (A or B) were consistent between the high-altitude E. baileyi and low-altitude E. fontanierii across most of the genome, those within the rearrangements in chromosome 1 were less consistent (Fig. 4e). Within the chromosome 1 rearrangement, many homologous bins were identified as compartment B in E. fontanierii but as compartment A in E. baileyi, indicating altered gene expression. In E. baileyi, the breakpoints of this SV were located at TAD boundaries with significantly lower insulation scores than those of other nearby TAD boundaries, indicating that interactions were weakest at the inversion breakpoints (Extended Data Fig. 7b). There were more species-specific TADs in the rearrangement region than found in other regions of chromosome 1 (Extended Data Fig. 7c). We examined the genes in the E. baileyi-specific TADs and found that six genes (AKR1C1, AKR1C18, AKR1C13, AKR1C5, AKR1C6 and AKR1C15) of the AKR1 family were in the same TAD. These genes were reported to be related to apoptosis and autophagy³¹. Five genes (CYP3A25, CYP3A1, CYP3A11, CYP3A6 and CYP3A9) in the CYP3A subfamily were also located in the same TAD, and these reportedly catalyse cholesterol and lipid synthesis reactions^32,33 (Extended Data Fig. 7d). Three consecutive TADs within the rearrangement of E. fontanierii fused into a single TAD in E. baileyi, and multiple genes within this TAD, including EGLN1, GNPAT, EXOC8, SPRTN, DISCI and TSNAX, are known to play a role in the cellular response to hypoxia^12,34–36. The DISC1, TSNAX and GSK3B complex was reportedly associated with DNA repair. After TAD fusion, DISC1 and TSNAX entered the same TAD (Fig. 4f).

We also compared the significant chromatin interactions of E. baileyi and E. fontanierii in the rearrangement region. The results showed that there were few significant interaction regions shared by two species (Extended Data Fig. 7e). Consistent with the TAD fusion event in the rearrangement region, some TADs in the region of E. baileyi were expanded and interaction regions extended significantly (Fig. 4g). In summary, the chromosome rearrangement in E. fontanierii and E. baileyi showed differences in compartment type, TAD boundary and significant interaction.

Mechanisms of inversion formation

Mechanism of short inversion (<1 Mb) formation by LINE-1.

LINE-1 insertions can result in genomic inversions generated during retrotransposition via ‘twin priming’³⁷. Of all transposable elements identified in short inversions, 31%, 25%, 25% and 19% were SINE, LINE, LTR and DNA elements, respectively. Almost all LINEs belonged to the LINE-1 superfamily (Extended Data Fig. 8a,b). More than 200 inversions were found to consist almost entirely of LINE-1 retrotransposon sequences, suggesting that the retrotransposition of LINE-1 elements is an important mechanism of chromosomal evolution in Eospalax.

Mechanism of large inversion (>1 Mb) formation by NAHR.

To determine the mechanism of inversion formation in E. baileyi, we identified inverted repeats near each breakpoint. At least one repeat ranging from 200 bp to 70 kb in length was identified near both breakpoints in 17 of the 18 inversions (Extended Data Fig. 8c,d), suggesting that most inversions were mediated by NAHR, consistent with patterns documented in deer mice²⁵. To explore whether breakpoints were in repetitive regions, we identified segmental duplications across the genome, and we found significant enrichment of segmental duplications (75%) (Extended Data Fig. 8e) near the inversion breakpoints (1 Mb) compared with the whole genome (Kolmogorov–Smirnov test: P < 0.0001). In conclusion, segmental duplication-enriched regions may be particularly prone to structural change via NAHR in Eospalax.

Discussion

Genome-wide scan of nucleotide variation has been used extensively to identify candidate genes for hypoxia adaptation in high-altitude humans and other animals^{12,13,35,38,39}, but the role of structural genomic variation has not been assessed systematically. Here we capitalized on technological advances in long-read sequencing to investigate how SVs may contribute to hypoxia adaptation in Eospalax.

The high-quality, chromosome-level genome assembly of E. baileyi enabled us to identify SVs, gene family expansions and PSGs. The genes we identified (Fig. 1f and Supplementary Tables 3 and 4) have been reported to play important roles in hypoxia adaption^30,40. In E. baileyi, significantly expanded lungs were suggested to maximize the pulmonary diffusing capacity for respiratory gases with proliferated alveolar and increased surface area¹, suggesting more efficient oxygen absorption^1,41–43.

EGLN1 and HIF1A are known to play important roles in cellular responses to hypoxia, and mutations within these genes (even in introns) are reported to have functional significance^44,45. Intronic inversions in EGLN1 and HIF1A in high-altitude Eospalax were associated with significant upregulation in the lung. Increased EGLN1 gene expression could result in increased PHD2 expression, which could blunt the expression of HIF1A in conditions of chronic hypoxia⁴⁶. HSF1 was a key transactivator of stress-inducible genes, regulating cellular and oxidative stress, immune, autophagy and other such responses to hypoxia^47,48. Moreover, E. baileyi-specific intronic deletion of the SFTPD genes was possibly involved in response to hypoxia. XRCC4 and UVRAG genes were associated with DNA damage. Thus, these results suggested that SVs could affect gene expression and take part in hypoxia response. Additionally, E. baileyi has been divided into two lineages. E. baileyi2 split from E. baileyi1 at ~1.18 Ma earlier than the split of E. smithi at ~0.75 Ma. We speculate that the shared SVs may account for the same phenotype in E. baileyi and E smithi. We think the adaptations evolved in the ancestry of the two highland species.

We speculate that the rearrangement in chromosome 1 occurred to avoid silencing by telomeres^49,50, thereby maintaining the required gene expression under hypoxia. Heterochromatin has been reported to be dynamic at telomeres⁵¹, which could alter adjacent nucleosomal chromatin to silence the expression of some nearby genes. The 3D chromosome model of chromosome 1 suggested that rearrangement was more tightly structured in E. fontanierii than E. baileyi (Supplementary Fig. 3). More compartment transformation from B to A in E. baileyi in the rearrangement suggested that genes within these regions might be more highly expressed than that in E. fontanierii³⁹. This rearrangement contributed to a fusion of TADs and expansions of interacting regions in E. baileyi (Fig. 4f,g), suggesting changes could alter patterns of some gene expression. Although we speculate that large rearrangements harbouring important genes such as EGLN1 should have functions facilitating hypoxia adaptation, functional tests of even whether they were fixed require further investigation.

Long-read sequencing enabled us to identify SVs and breakpoint regions, which were mainly distributed in repeated regions. A large proportion of short SVs and 56% (10 out of 18) of large SVs occurred at chromosomal ends (Fig. 2c), possibly because telomeric concentrations of repetitive elements facilitated NAHR^25,52. Alternatively, inversions with breakpoints near telomeres could be less likely to be removed by purifying selection because these regions are relatively sparse in genes^25,52. Large inversions mainly originated via inverted repeats through NAHR in Eospalax, similar to other mammals^25,53.

In this study, the patterns of genomic structural variation in E. baileyi and the documented effects on the expression of hypoxia-relevant genes suggest numerous hypotheses about possible mechanisms of hypoxia adaptation. Such hypotheses can now be tested using genome-editing tools in conjunction with experimental studies of organismal physiology.

Methods

Sample information

We collected 17 specimens of Eospalax in northern China: 3 E. baileyi, 3 E. fontanierii, 3 E. cansus, 3 E. rufescens, 3 E. rothschildi and 2 E. smithi (Supplementary Table 9). Live zokors were caught and euthanized using chloral hydrate. Tissues were collected and preserved in liquid nitrogen.

Phenotypic data of Eospalax

We measured the mass of the body and tissues for all specimens. We did a linear regression analysis of the body and lung mass for all samples. Blood was collected with EDTA as anticoagulant. RBC and Hb were measured using Mindray BC-2800Vet. Three replicates were measured for each individual. The data for E. baileyi came from ref. 3.

Fibroblast culture and karyotype analysis

Skin fibroblast culture.

The skin of zokor was sterilized with 75% alcohol, washed with PBS, clipped by adding a DMEM/F12 medium and digested with collagenase at 37 °C for 1 h. After centrifugation at 3,700g for 10 min, the supernatant was poured off and the precipitate was mixed with DMEM/F12 medium using 10% FBS, and it was spread out flat in a Petri dish.

Cellular chromosome karyotype analysis.

Vigorous growing cells with 70–80% confluence were selected and treated with 0.1 µg ml⁻¹ of colchicine for 1 h and 0.075 mol l⁻¹ of KCl solution at low osmolality. Trypsin was used to digest the cells and then added with a newly configured fixative (methanol:glacial acetic acid = 3:1), and the fixation was repeated three times to make the slides, which were placed in the oven for ageing. The slides were stained with Giemsa stain for 15 min, and the intermediate schizogony phases with high sorting index and well-dispersed chromosomes were selected for karyotyping through microscopic examination.

DNA extraction and library construction

Genomic DNA was extracted from muscle using SDS-based methods. The purity, concentrations and integrity of genomic DNA were measured using a Nanodrop spectrophotometer and Qubit 3.0 fluorometer. After the samples passed quality control, a BluePippin automated nucleic acid recovery instrument (Sage Sciences) was used to collect the target fragments (10–50 kb). AMPure XP bead purification, DNA damage repair, DNA end-repair and adaptor ligation (SQK-LSL109 Ligation Sequencing Kit) were performed to construct sequencing libraries. The Qubit fluorometer was used to quantify the sequencing-ready DNA libraries.

Data generation

Long-read sequencing of E. baileyi using Oxford Nanopore Technologies (ONT, 209 Gb) was performed to assemble contigs; a 51× coverage of short paired-end sequencing was performed on the Illumina platform for subsequent read correction and polishing. The assembled contigs were anchored to chromosomes using Hi-C data (315 Gb, Extended Data Table 1).

Standard operating procedures were followed to load the long-read sequencing libraries onto PromethION R9.4.1 flow cells for the long-read–SV dataset (Supplementary Table 10). The number of reads and N50 length for raw long-read data were calculated using NanoPlot v.1.38.0.

Short reads from 146 Eospalax individuals were utilized for SV frequency estimation (Supplementary Table 11), and 143 of the 146 were downloaded from public databases^3,27,54, including 55 E. baileyi. Of these, 40 belonged to E. baileyi1 and 15 belonged to E. baileyi2.

De novo genome assembly and annotation

Genome assembly.

The NextCorrect module of NextDenovo v.2.3.0 was used to obtain the consistency sequence using clean long reads from the ONT platform. NextGraph was used for the initial assembly of the genome. NextPolish was then used to subject the assembly to two rounds of corrections using Illumina reads and three rounds of self-polish corrections using ONT data. Mitochondrial sequences identified using blastn (blast v.2.2.29) were removed. BUSCO v.5.2.2 was used to assess the genome integrity.

The paired-end Hi-C reads were mapped to the assembled contigs using BWA-MEM v.0.7.17-r1188. Chromosomes were organized, ordered and oriented using automated 3D-DNA v.180922 and Juicer v.1.6 processes. Juicebox v.1.5.1 was used to correct the chromosomal-level genome.

Repetitive DNA annotation.

LTR_FINDER, Tandem Repeat Finder (TRF) v.4.07b, RepeatModeler v.4.07b, RepeatMasker v.1.331 and Repeat-ProteinMask v.4.1.0 were used to identify repetitive sequences. The overlapping transposable elements belonging to the same repeat class within the genome’s repetitive sequences were sorted and merged.

Gene annotation.

Combining ab initio predictions and homology and transcript evidence, we generated models for zokors’ protein-coding genes. AUGUSTUS v.3.3.1 and Genscan gene models based on repeat-masked genomes were annotated for ab initio prediction. The protein sequences of Mesocricetus auratus, Rattus norvegicus, Microtus ochrogaster, E. fontanierii and E. baileyi were used as templates in GeMoMa v.1.6.1. Exonerate v.2.4.0 (http://www.ebi.ac.uk/~guy/exonerate/) was utilized to align homologous protein sequences from the UniProt database to the E. baileyi genome. RNA-seq reads were mapped to genomes using HiSAT2 v.2.1.0 and re-assembled using StringTie v.2.2.1. Program Assemble Spliced Alignment v.2.3.3 was subsequently used to generate transcript predictions. TransDecoder v.2.0 (http://transdecoder.github.io) was then used to predict the genes. Evidence Modeler v.1.1.1 was used to integrate all predictions. Using BUSCO, the completeness of gene annotations was assessed.

Analysis of gene family expansions and contractions

We downloaded the genome and annotation files of E. fontanierii, Rhizomys pruinosus, Spalax galili, Rattus norvegicus, Mus musculus, Peromyscus maniculatus, Mesocricetus auratus and Jaculus jaculus from the database of NCBI. Together with the E. bailey assembly, nine species were used to examine lineage-specific changes in gene family size. Orthofinder v.2.5.4 was used to identify orthogroups among nine species based on the longest transcripts extracted with a custom Perl script. We generated a tree with divergence times based on the species tree and single-copy gene sequences using MCMCTree in PAML v.4.10.3. The species tree was calibrated using TIMETREE (http://www.timetree.org/) fossil times. The contracted and expanded gene families were identified using CAFÉ v.4.2.1 with orthogroup counts and a tree containing divergence time as input files. The likelihood approach modelled the evolution of gene family size using a stochastic birth and death process in CAFÉ v.4.2.1. We performed a multiple correction of P value (FDR < 0.05; Extended Data Table 1).

Identification of genes subject to selection

We identified PSGs in E. baileyi based on homologous single-copy genes from nine species. The codon sequences of each homologous single-copy gene were extracted for subsequent analysis using a Perl script (Epal2nal.pl, https://github.com/wonaya/ParaAT) based on the amino acid and CDS sequences. The selected genes were identified using the ‘aBSREL’ parameter of HyPhy (https://www.hyphy.org/), with an omega value greater than 1 indicating PSGs and an omega value less than 1 indicating negatively selected genes. P values were multiples corrected by FDR values (<0.05).

SV (<1 Mb) calling, description and verification

Illumina SV calling.

Quality control was conducted on the sequencing data using FastQC. The clean short reads were mapped against the E. fontanierii genome using BWA-MEM v.0.7.17-r1188. PCR duplicates were removed using Picard (https://broadinstitute.github.io/picard) with default parameters. Three SV-detecting tools (DELLY v.0.7.6, LUMPY v.0.2.13 and Manta v.1.6.0) with respective default parameters were used to call SVs in each individual. Only SVs with the ‘PASS’ and ‘PRECISE’ flags using all three approaches were kept and merged in svimmer (https://github.com/DecodeGenetics/svimmer). GraphTyper2 was used to refine the genotype results based on the alignment information in BAM files. Sites with a high ratio of missing data across all samples (--max missing 0.90) were filtered out using VCFtools v.0.1.16.

Nanopore SV calling.

We mapped raw long reads to the E. fontanierii genome using NGM-LR v.0.2.7 with default parameters. SAMtools v.1.7 was used to sort and convert alignments into BAM format and calculate depth and coverage. SVs longer than 50 bp were first called in each sample using Sniffles v.1.0.12 with the option ‘-l 50’. After merging using surpyvor v.0.10.0, genotypes at each SV across all samples were identified again using Sniffles. The sequences of insertions and deletions were added and polished using Iris v.1.0.5. We merged SVs using surpyvor again, which was based on the comparison algorithm of SURVIVOR v.1.0.3. Complex SVs (that is, SVs with multiple types, overlapping SVs and SVs longer than 1 Mb) were considered as false negatives and removed. SVs with a missing rate greater than 0.15 across all samples were also removed for carrying insufficient genotype information.

Genomic region and repeat annotation for SVs.

The VCF format file with SV information was converted to bed format to get the genomic coordinates of each SV. The intersection between repetitive elements and SVs was then calculated using the intersect method in BEDtools v.2.29.2. The VCF format file of the long-read–SV dataset contained 17 individuals. The VCF format file of the SR-SV dataset had 146 individuals (55 E. baileyi, 25 E. cansus, 21 E. smithi, 18 E. rothschildi, 14 E. rufescens and 13 E. fontanierii).

Gene function of associated SVs.

Using a custom Python script, the frequencies of each SV were calculated among all species. SVs fixed in one species and not identified in others were defined as species-specific SVs. Notably, E. fontanierii was used as the reference genome; thus, E. fontanierii-specific SVs were identified as those SVs that were fixed in all other samples except those from E. fontanierii. We focused on species-specific and species-shared SV-associated genes in this study. Genes associated with these SVs were then extracted from the annotation files in each species. The functional enrichments of genes associated with species-specific SVs were performed using Metascape.

PCR validation of SVs.

Genomic DNA was extracted from the muscles of E. fontanierii and E. baileyi. PCR reactions were performed using a 2× Taq PCR Master mix (TIANGEN). The HSF1 gene deletion was confirmed via PCR genotyping using the following primer pairs: (F: CCTGCCCTCT-CACTGAGTTG, R: TCCGGGTCTTCTGTTCTGTC).

Large inversion (>1 Mb)

Identification and verification of large SVs.

The chromosome-level genome of E. baileyi was compared with that of E. fontanierii using the nucmer model of MUMmer v.4.0.0 with default parameters. Alignments shorter than 100 bp were filtered using the delta-filter, and their coordinates were obtained using the show-coords module. SyRI v.1.5.6 was used to identify large inversions (>1 Mb).

We aligned the contig-level assemblies of E. baileyi to the E. fontanierii genome using nucmer with default parameters. Inversion breakpoints were identifiable if (1) a contig contains the entire inversion, with the inversion sequence aligned in reverse to the reference genome, and (2) a contig includes only a partial region of the inversion, with a sequence from one end of the contig to the inversion breakpoint aligned in the same direction as the reference genome and a sequence from the other end of the contig to the inversion breakpoint aligned in reverse to the reference genome. The breakpoints of each inversion were identified based on the alignments at the contig level.

Genomes from two additional individuals of E. baileyi were de novo assembled, and one public scaffold-level genome³ was used to verify the SV on chromosome 1. First, we used NextDenovo to de novo assemble the ONT sequencing data to obtain contig-level genomes. Then, we used the ‘correct’ module of Ragtag v.2.0.1 to scaffold the contigs using the chromosome-level E. baileyi genome as reference. Then, the same method was used to identify the presence of SV on chromosome 1 in both genomes.

Predicting telomere locations.

We used tolkit v.0.2.2 (https://github.com/tolkit/telomeric-identifier) to identify telomeres in the E. fontanierii genome.

Inverted repeat and segmental duplications at inversion breakpoints.

We identified segmental duplications near inversion breakpoints using SEDEF with the soft-masked repeat sequence as input. Segmental duplications were then screened for length ≥1 kb and soft-masked repeat sequences <70%. We calculated the segmental duplication density around 500 kb of each inversion breakpoint. To compare the density at the breakpoints with the density in random regions of the whole genome, we randomly selected 1,000 sites from the genome and calculated the segmental duplication density in the region 500 kb upstream and downstream from each site. Inverted repeats were identified in three ways: (1) segmental duplications (>1 kb) located near two breakpoint regions of an SV that were inverted and highly similar were regarded as inverted repeats; (2) IRF v.3.05 was used to identify inverted repeats longer than 500 bp; (3) short (>100 bp) inverted repeats were obtained by aligning each chromosome to itself.

Potential effect of large inversions.

We checked whether any inversion breakpoints were located within the genes. We calculated the distance between each inversion breakpoint and the nearest transcriptional start site. LOF SNPs were utilized to represent genetic load. SnpEff v.5.1 was used to identify LOF SNPs (stop gained, start lost, splice acceptor variant and splice donor variant). We calculated the number of LOF SNPs in each inverted region of E. fontanierii and E. baileyi, in addition to the overall number of LOF SNPs found in their genomes.

Construction of homologous bin pairs.

To identify homologous bins between E. fontanierii and E. baileyi, we used LAST v.885 to align the query E. baileyi sequence to the target E. fontanierii sequence. According to published pipelines⁵⁵, the MAF file obtained from LAST was converted into a bin pairs file. Using the University of California, Santa Cruz genome browser, the MAF file was converted to PSL format using mafToPsl and then converted to a chain file using axtChain (-linearGap = medium). The liftOver tool (-minMatch = 0.85) was used to construct homologous bin pairs of varying sizes.

Comparison of 3D chromatin architecture.

We compared the 3D chromatin architecture in different regions (within the chromosome rearrangement region and in other regions) of E. baileyi and E. fontanierii, including the A/B compartments, TADs and significant interaction bins.

Principal component analysis was performed using the reciprocal matrix at 150 kb resolution. The A and B compartments were separated along PC1, with positive and negative values for A and B compartments, respectively. The gene density was higher in the A compartment than in the B compartment. The homologous bins of the two species were aligned, and the counterpart A/B compartments in each genome were identified. The homologous bins were identified as switched regions if the two bins showed different compartments, whereas the remaining bins were regarded as stable regions. The ratio of the two region types was counted (A to B: A in E. fontanierii but B in E. baileyi; B to A: B in E. fontanierii but A in E. baileyi; stable: bins in both genomes contained the same compartment type).

We used cworld-dekker (https://github.com/dekkerlab/cworld-dekker/releases) to identify TAD boundaries based on the 40 kb resolution matrix and obtained an insulation score. For a given TAD in one genome, we identified all bins that showed overlap with this TAD. Counterpart TADs between the two species with more than 70% overlap were defined as shared TADs, and all other TADs were classified as species-specific TADs.

Finally, the significance of genome-wide interactions between each pair of loci (at 20 kb resolution) was determined using FitHic v.1.1.3. Interactions with P ≤ 0.01, q ≤ 0.01,and >9 reads supporting the interactions were significant interactions. We limited our analysis to significant interaction regions within the same chromosomes. The bins of significant interaction regions were determined according to their coordinates. We then tallied the number of shared and species-specific bins between the two genomes and calculated the distance between the interaction bins.

RNA-seq analysis

The transcriptomes of the hearts, livers and lungs of E. baileyi, E. cansus and E. rothschildi from common garden experiments were sequenced. We collected the zokors on the way from our university to our field station located at Guoluo, Qinghai Province, with an altitude of ~3800 m. All the male adults were in the same environment there for at least 1 week. We anaesthetized the zokor with chloral hydrate, collected tissue and froze it in liquid nitrogen and stored it in a −80 °C refrigerator. We extracted RNA using RNAprep pure Tissue Kit (TIAN-GEN). NanoPhotometer (IMPLEN) was used to detect RNA purity. The RNA Nano 6000 Assay Kit of the Bioanalyzer 2100 system (Agilent Technologies) was used to assess RNA integrity and concentration. Then 1–2 µg of RNA per sample was used to construct the cDNA library. Qubit3.0 was used for preliminary quantification. Agilent Bioanalyzer 2100 was used to detect the library insert size. StepOnePlus Real-Time PCR System was used to accurately quantitate the insert size (library valid concentration >10 nM). E. baileyi was used as the high-altitude species, whereas E. cansus and E. rothschildi were the low-altitude species. The raw reads were filtered using Trim Galore to remove low-quality reads, and the filtered data were assessed using FastQC. HISAT2 was used to align clean reads to the genome. We calculated the number of transcripts of each gene mapped against the genome using Stringtie v.2.0.4.

DESeq v.2 was used to identify DEGs (adjusted P ≤ 0.05). We used the result of multiple corrections (Benjamini–Hochberg model). We corrected RNA-seq data using phylogenetic comparative methods, such as phylogenetic generalized least squares. We increased statistical power and controlled the FDR using mashr (https://github.com/stephenslab/mashr; Extended Data Table 2). Metascape was used for functional enrichments.

Dual-luciferase assay

HEK293T cells were cultured using an RPM11640 medium supplemented with penicillin, streptomycin and horse serum. The synthesized sequences of HIF1A (Ensembl: Chr13: 44667409–44669626), EGLN1 (Ensembl: Chr1: 1889355–1890797) and HSF1 (Ensembl: Chr9: 81333952–81335801) were inserted into pGL13-Promoter (Promega). We prepared and transformed competent cells and then extracted plasmid DNA. Both the plasmid with the target gene (pGL13) and the control plasmid (pRL-TK) were transfected into HEK293T cells using Lipofectamine 2000 (Thermo Fisher Scientific) and OPTI-MEM medium (GIBCO, Invitrogen). Then 24 h after transfection, cellular lysates were collected. The ratio of FLUC and RLUC values (that is, luciferase activity) was calculated per well. All data were presented as means with standard deviations, and statistical differences between groups were examined using one-way analysis of variance and Tukey’s test, with a P value less than 0.05 indicating significance, and we used a two-sided t-test.

Prediction of SFTPD protein structure

We used RoseTTAFold with default parameters to predict the protein structure of SFTPD of E. fontanierii and E. baileyi. PyMOL (https://pymol.org) was used to visualize the predicted results.

3D model reconstruction of genomes

The ‘LorDG’ module of Genomeflow (https://github.com/jianlin-cheng/GenomeFlow) was used to build 3D chromosomes and genome models. Then we viewed the 3D structure using PyMOL.

Ethics statement

The Ethics Committee of the College of Ecology, Lanzhou University (ethical approval form no. EAF2022007) authorized all experimental methods and sample collection protocols. China’s Ministry of Science and Technology developed the ‘Guidelines on the Ethical Treatment of Experimental Animals’, which governed all sample operations.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Extended Data

Extended Data Fig. 3 | — The dark blue line at the top of each panel represents the genome of *E. fontanierii*, while the dark orange line represents the genome of *E. baileyi*. The orange block in each panel denotes the inversion position. Each panel (a-m) was one of 32 pseudochromosomes.

Extended Data Fig. 4 | — (a) Contigs from *de novo* genome assemblies (‘query’, y-axis) were aligned to the *E. fontanierii* reference genome (‘reference’, x-axis). Contigs (blue) and those identifying inversion breakpoints (red) are shown. Predicted inversion boundaries are highlighted (purple box), showing predicted inversion (arrow) above. (b) The Synteny comparison between the genomes of *E. fontanierii* and *E. baileyi* showed that part of the contig of *E. baileyi* was reversed and the other contig was positive.

Extended Data Fig. 5 | — The fixed-SV related differentially expressed genes in the lung (a) liver (b), and heart (c) of *E. baileyi*. Negative binomial distribution model and generalized linear model were used to identify DEGs, and we multiple-corrected using Benjamini-Hochberg model (padj ≤ 0.05). (d) GO enrichment terms of fixed SV-related genes typically in *E. baileyi*. (e) Enrichment of genes associated with *E. baileyi*-specific SVs, *E. smithi*-specific SVs and the shared SVs between the two species (Hypergeometric test, p < 0.05; Benjamini-Hochberg model, q < 0.05).

Extended Data Fig. 6 | — (a) The frequency of inversion (2,212 bp) in *HIF1A* in *Eospalax*. (b) The nucleotide sequence and amino acid sequence of *SFTPD* in *E. baileyi* reveal a loss of one exon and 55 amino acids. De novo prediction of protein structure for *E. fontanierii* and *E. baileyi*. (c) A 1,989 bp intronic deletion in *UVRAG* was validated by coverage of reads. (d) Map and coverage of reads of deletion in *XRCC4*.

Extended Data Fig. 7 | — (a) Compartment A/B, gene density and GC content of chromosome 1 in *E. fontanierii* and *E. baileyi*. A/B compartments are determined by the PC1 value, the red bar indicates compartment A and the blue bar indicates compartment B. (b) Topologically associating domains (TAD) insulation score of chromosome 1 in *E. fontanierii* and *E. baileyi*. (c) Venn diagram showing species-specific and shared TADs in SV regions and background regions. (d) Two *E. baileyi*-specific TADs associated hypoxia adaptation. (e) The number of shared/species-specific significant interaction regions between *E. fontanierii* and *E. baileyi* in the rearrangement region.

Extended Data Fig. 8 | — (a) The overlap of inversions (< 1 Mb) with different types of repeats. Different colors represent different repeat sequence types. The X-axis of each figure represents the percentage of the length of the inverted sequence overlapped with different types of repeats as a percentage of the length of the inverted sequence. The Y-axis represents the number of inversions. (b) The type and proportion of repeated sequences in inversion (< 1 Mb). The x-coordinate represents the ratio of LINE1 to each inversion sequence. (c) Examples of inversions with inverted repeats at both breakpoints on chromosomes 1, 2, 3, 6, 8, 9, 11, 17, 21, 23, 24, 27 and 31. Dotplot depicts the inverted repeats near each inversion breakpoint. Locations of breakpoints are denoted by orange arrows, and only alignments longer than 100 bp and within 1 Mb of the breakpoints are displayed. Inverted repeats located within 1 Mb of both breakpoints are depicted (red) and highlighted (purple box). (d) The pie shows the number of inversions (> 1 Mb) that have at least one pair of inverted repeats at both inversion breakpoints. The histogram shows the length of inverted repeats near breakpoints. (e) Segmental duplication (SD) percentage of randomly selected 1 Mb regions across the genome and inversion (>1 Mb) regions.

Extended Data Table. 1 |.

Genome assembly data

Sequencing Platform	Data Size	Sequencing Depth
Oxford Nanopore Technologies, ONT	209.0 Gb	84×
High-through Chromosome Conformation Capture, Hi-C	314.53 Gb	139×
Illumina	255.1 Gb	51×

Open in a new tab

Extended Data Table. 2 |.

Dual-luciferase assay of EGLN1, HIF1A, and HSF1

	EGLN1				HIF1A		HSF1
	pGL3-promoter+Luc	pGL3-promoter+ZH-EGLN1+Luc	pGL3-promoter+GY-EGLN1+Luc	pGL3-promoter+Luc	pGL3-promoter+ZH-HIF1A+Luc	pGL3-promoter+GY-HIF1A+Luc	pGL3-promoter+Luc	pGL3-promoter+ZH-HSF1+Luc	pGL3-promoter+GY-HSF1+Luc
Luc	48,059	81,987	130,256	50,468	99,549	140,249	44,768	112,628	173,763
	46,509	65,681	128,695	58,783	87,829	135,135	51,810	87,089	147,634
	56,059	98,063	125,254	59,134	94,763	149,959	55,337	104,957	150,417
Average	50,209	81,910.33333	128,068.3333	56,128.33333	94,047	141,781	50,638.33333	101,558	157,271.3333
SD	5,125.182924	16,191.13613	2,559.205801	4905.13306	5,892.715164	7,529.808231	5,381.035433	13,104.38976	14,349.82837
RLuc	67,107	68,754	64,518	71,279	70,420	68,874	64,869	62,900	66,264
	57,873	54,668	56,741	75,354	74,283	72,729	68,988	64,160	67,168
	60,176	65,452	66,808	69,098	67,994	72,581	66,022	65,603	69,318
Average	61,718.66667	62,958	62,689	71,910.33333	70,899	71,394.66667	66,626.33333	64,221	67,583.33333
SD	4,806.407633	7,366.741206	5,276.841195	3,175.424434	3,171.744157	2,184.215267	2,124.959843	1,352.53207	1,568.791042
Luc/RLuc	0.7162	1.1925	2.0189	0.708	1.4136	2.0363	0.6901	1.7906	2.6223
	0.8036	1.2015	2.2681	0.7801	1.1824	1.8581	0.751	1.3574	2.198
	0.9316	1.4982	1.8748	0.8558	1.3937	2.0661	0.8382	1.5999	2.17
Average	0.817133333	1.2974	2.053933333	0.7813	1.3299	1.986833333	0.759766667	1.582633333	2.3301
SD	0.108335836	0.173956115	0.19897669	0.073907307	0.12812568	0.112477613	0.074438185	0.217115553	0.253439598
Relative	0.8765	1.4594	2.4707	0.9062	1.8093	2.6063	0.9083	2.3568	3.4515
luciferase	0.9834	1.4704	2.7757	0.9985	1.5134	2.3782	0.9885	1.7866	2.893
activity	1.1401	1.8335	2.2944	1.0954	1.7838	2.6444	1.1032	2.1058	2.8561
Average	1	1.587766667	2.5136	1.000033333	1.702166667	2.542966667	1	2.083066667	3.066866667
SD	0.132581711	0.21288237	0.243500986	0.094609319	0.163973179	0.14395813	0.097957593	0.285778959	0.333612805

Open in a new tab

Supplementary Material

supplementary material

NIHMS2090629-supplement-supplementary_material.pdf^{(2MB, pdf)}

The online version contains supplementary material available at https://doi.org/10.1038/s41559-023-02275-7.

Acknowledgements

We thank members of the Hoekstra laboratory for commenting on the paper. We thank X. Luo from the Kunming Institute of Zoology. This project was supported by the National Natural Science Foundation of China (grants 32271691 and 32071487 to K.L.), the National Key Research and Development Programs (grant 2021YFD1200901 to K.L.), the Fundamental Research Funds for Central Universities, LZU (grants lzujbky-2021-ey17 to K.L. and lzujbky-2022-it01 to Y.W.), the Science Fund for Creative Research Groups of Gansu Province (grant 21JR7RA533 to K.L.), Lanzhou University’s ‘Double First-Class’ Guided Project-Team Building Funding-Research Startup Fee for K. Li, a grant from State Key Laboratory of Herbage Improvement and Grassland Agro-ecosystems (Lanzhou University) (grant SKLGAE-202001, -202009 and -202010 to K.L.) and the Key Basic Research Project of Qinghai Provincial Department of Science and Technology (grant 2022-ZJ-733 to Q.X.). We received support for computational work from the Big Data Computing Platform for Western Ecological Environment and Regional Development and Supercomputing Center of Lanzhou University.

Footnotes

Code availability

The code used for the analyses is available from GitHub (https://github.com/anxuan-web/Structural-variants-in-Eospalax).

Competing interests

The authors declare no competing interests.

Additional information

Extended data is available for this paper at https://doi.org/10.1038/s41559-023-02275-7.

Peer review information

Nature Ecology & Evolution thanks Erica Heinrich and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. Peer reviewer reports are available.

Data availability

DNA sequence data are available from NCBI SRA under BioProject accession PRJNA947478. The RNA sequence data reported in this paper are deposited in the Genome Sequence Archive in National Genomics Data Center, China National Center for Bioinformation/Beijing Institute of Genomics, Chinese Academy of Sciences (PRJCA019941: CRA012698), which are publicly accessible at https://ngdc.cncb.ac.cn/gsa. The chromosomal level genome assembly and annotation reported in this paper are deposited in the China National GeneBank DataBase (https://db.cngb.org) with BioProject ID CNP0005002 and uploaded in figshare (https://doi.org/10.6084/m9.figshare.24085119). Source data are provided with this paper.

References

1.Chiou KL et al. Genomic signatures of high-altitude adaptation and chromosomal polymorphism in geladas. Nat. Ecol. Evol 6, 630–643 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
2.Cai Z et al. Adaptive transcriptome profiling of subterranean Zokor, Myospalax baileyi, to high-altitude stresses in Tibet. Sci. Rep 8, 4671 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
3.Zhang T et al. Phenotypic and genomic adaptations to the extremely high elevation in plateau zokor (Myospalax baileyi). Mol. Ecol 30, 5765–5779 (2021). [DOI] [PubMed] [Google Scholar]
4.Simonson TS et al. Genetic evidence for high-altitude adaptation in Tibet. Science 329, 72–75 (2010). [DOI] [PubMed] [Google Scholar]
5.Yi X et al. Sequencing of 50 human exomes reveals adaptation to high altitude. Science 329, 75–78 (2010). [DOI] [PMC free article] [PubMed] [Google Scholar]
6.Peng Y et al. Genetic variations in Tibetan populations and high-altitude adaptation at the Himalayas. Mol. Biol. Evol 28, 1075–1081 (2011). [DOI] [PubMed] [Google Scholar]
7.Peng Y et al. Down-regulation of EPAS1 transcription and genetic adaptation of Tibetans to high-altitude hypoxia. Mol. Biol. Evol 34, 818–830 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
8.Huerta-Sánchez E et al. Genetic signatures reveal high-altitude adaptation in a set of Ethiopian populations. Mol. Biol. Evol 30, 1877–1888 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
9.Schweizer RM et al. Physiological and genomic evidence that selection on the transcription factor Epas1 has altered cardiovascular function in high-altitude deer mice. PLoS Genet 15, e1008420 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
10.Schweizer RM et al. Broad concordance in the spatial distribution of adaptive and neutral genetic variation across an elevational gradient in deer mice. Mol. Biol. Evol 38, 4286–4300 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
11.Scott GR et al. Adaptive modifications of muscle phenotype in high-altitude deer mice are associated with evolved changes in gene regulation. Mol. Biol. Evol 32, 1962–1976 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
12.Storz JF High-altitude adaptation: mechanistic insights from integrated genomics and physiology. Mol. Biol. Evol 38, 2677–2691 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
13.Storz JF & Cheviron ZA Physiological genomics of adaptation to high-altitude hypoxia. Annu. Rev. Anim. Biosci 9, 149–171 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
14.McClelland GB & Scott GR Evolved mechanisms of aerobic performance and hypoxia resistance in high-altitude natives. Annu. Rev. Physiol 81, 561–583 (2019). [DOI] [PubMed] [Google Scholar]
15.Storz JF & Scott GR Life ascending: mechanism and process in physiological adaptation to high-altitude hypoxia. Annu. Rev. Ecol., Evol. Syst 50, 503 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
16.Storz JF et al. Evolution of physiological performance capacities and environmental adaptation: insights from high-elevation deer mice (Peromyscus maniculatus). J. Mammal 100, 910–922 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
17.Stewart NB & Rogers RL Chromosomal rearrangements as a source of new gene formation in Drosophila yakuba. PLoS Genet 15, e1008314 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
18.Shi J et al. Structural variants involved in high-altitude adaptation detected using single-molecule long-read sequencing. Nat. Commun 14, 8282 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
19.Quan C et al. Characterization of structural variation in Tibetans reveals new evidence of high-altitude adaptation and introgression. Genome Biol 22, 159 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
20.Almarri MA et al. Population structure, stratification, and introgression of human structural variation. Cell 182, 189–199. e115 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
21.Mérot C et al. A roadmap for understanding the evolutionary significance of structural genomic variation. Trends Ecol. Evol 35, 561–572 (2020). [DOI] [PubMed] [Google Scholar]
22.Villoutreix R et al. Large-scale mutation in the evolution of a gene complex for cryptic coloration. Science 369, 460–466 (2020). [DOI] [PubMed] [Google Scholar]
23.Yan Z et al. Evolution of a supergene that regulates a trans-species social polymorphism. Nat. Ecol. Evol 4, 240–249 (2020). [DOI] [PubMed] [Google Scholar]
24.Hager ER et al. A chromosomal inversion contributes to divergence in multiple traits between deer mouse ecotypes. Science 377, 399–405 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
25.Harringmeyer OS & Hoekstra HE Chromosomal inversion polymorphisms shape the genomic landscape of deer mice. Nat. Ecol. Evol 6, 1–15 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
26.Shams I et al. Oxygen and carbon dioxide fluctuations in burrows of subterranean blind mole rats indicate tolerance to hypoxic-hypercapnic stresses. Comp. Biochem. Physiol. A 142, 376–382 (2005). [DOI] [PubMed] [Google Scholar]
27.Liu X et al. Genomic insights into zokors’ phylogeny and speciation in China. Proc. Natl Acad. Sci. USA 119, e2121819119 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
28.Storz JF & Bautista NM Altitude acclimatization, hemoglobin-oxygen affinity, and circulatory oxygen transport in hypoxia. Mol. Asp. Med 84, 101052 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
29.Partha R et al. Subterranean mammals show convergent regression in ocular genes and enhancers, along with adaptation to tunneling. eLife 6, e25884 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
30.Davies KTJ et al. Limited evidence for parallel molecular adaptations associated with the subterranean niche in mammals: a comparative study of three superorders. Mol. Biol. Evol 35, 2544–2559 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
31.Chang L et al. Akr1c1 connects autophagy and oxidative stress by interacting with sqstm1 in a catalytic-independent manner. Acta Pharmacol. Sin 43, 703–711 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
32.Bochud M et al. Association of CYP3A5 genotypes with blood pressure and renal function in African families. J. Hypertens 24, 923–929 (2006). [DOI] [PubMed] [Google Scholar]
33.Martinez C et al. Expression of paclitaxel-inactivating CYP3A activity in human colorectal cancer: implications for drug therapy. Br. J. Cancer 87, 681–686 (2002). [DOI] [PMC free article] [PubMed] [Google Scholar]
34.Jain IH et al. Genetic screen for cell fitness in high or low oxygen highlights mitochondrial and lipid metabolism. Cell 181, 716–727. e711 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
35.Brutsaert TD et al. Association of EGLN1 gene with high aerobic capacity of Peruvian Quechua at high altitude. Proc. Natl Acad. Sci. USA 116, 24006–24011 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
36.Chien T et al. GSK3β negatively regulates TRAX, a scaffold protein implicated in mental disorders, for NHEJ-mediated DNA repair in neurons. Mol. Psychiatry 23, 2375–2390 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
37.Ostertag EM & Kazazian HH Twin priming: a proposed mechanism for the creation of inversions in L1 retrotransposition. Genome Res 11, 2059–2065 (2001). [DOI] [PMC free article] [PubMed] [Google Scholar]
38.Yang J et al. Genetic signatures of high-altitude adaptation in Tibetans. Proc. Natl Acad. Sci. USA 114, 4189–4194 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
39.Hu L et al. Arctic introgression and chromatin regulation facilitated rapid Qinghai-Tibet Plateau colonization by an avian predator. Nat. Commun 13, 6413 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
40.Davies KT et al. Family wide molecular adaptations to underground life in African mole-rats revealed by phylogenomic analysis. Mol. Biol. Evol 32, 3089–3107 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
41.Hsia CC et al. Enhanced alveolar growth and remodeling in guinea pigs raised at high altitude. Respir. Physiol. Neurobiol 147, 105–115 (2005). [DOI] [PubMed] [Google Scholar]
42.Llapur CJ et al. Increased lung volume in infants and toddlers at high compared to low altitude. Pediatr. Pulmonol 48, 1224–1230 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
43.Li J et al. A new homotetramer hemoglobin in the pulmonary surfactant of plateau zokors (Myospalax Baileyi). Front. Genet 13, 824049 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
44.Cimmino F et al. HIF-1 transcription activity: HIF1A driven response in normoxia and in hypoxia. BMC Med. Genet 20, 37 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
45.Aggarwal S et al. EGLN1 involvement in high-altitude adaptation revealed through genetic analysis of extreme constitution types defined in Ayurveda. Proc. Natl Acad. Sci. USA 107, 18961–18966 (2010). [DOI] [PMC free article] [PubMed] [Google Scholar]
46.Parathath S et al. Hypoxia is present in murine atherosclerotic plaques and has multiple adverse effects on macrophage lipid metabolism. Circ. Res 109, 1141–1152 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]
47.Mendillo ML et al. HSF1 drives a transcriptional program distinct from heat shock to support highly malignant human cancers. Cell 150, 549–562 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]
48.Santagata S et al. High levels of nuclear heat-shock factor 1 (HSF1) are associated with poor prognosis in breast cancer. Proc. Natl Acad. Sci. USA 108, 18378–18383 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]
49.Gottschling DE et al. Position effect at S. cerevisiae telomeres: reversible repression of Pol II transcription. Cell 63, 751–762 (1990). [DOI] [PubMed] [Google Scholar]
50.Nimmo ER et al. Telomere-associated chromosome breakage in fission yeast results in variegated expression of adjacent genes. EMBO J 13, 3801–3811 (1994). [DOI] [PMC free article] [PubMed] [Google Scholar]
51.Wang J et al. A heterochromatin domain forms gradually at a new telomere and is dynamic at stable telomeres. Mol. Cell. Biol 38, e00393–17 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
52.Samonte RV & Eichler EE Segmental duplications and the evolution of the primate genome. Nat. Rev. Genet 3, 65–72 (2002). [DOI] [PubMed] [Google Scholar]
53.Porubsky D et al. Recurrent inversion polymorphisms in humans associate with genetic instability and genomic disorders. Cell 185, 1986–2005. e1926 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
54.Zhang T et al. Phylogenetic relationships of the zokor genus Eospalax (Mammalia, Rodentia, Spalacidae) inferred from whole-genome analyses, with description of a new species endemic to Hengduan Mountains. Zool. Res 43, 331–342 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
55.Kielbasa SM et al. Adaptive seeds tame genomic sequence comparison. Genome Res 21, 487–493 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

supplementary material

NIHMS2090629-supplement-supplementary_material.pdf^{(2MB, pdf)}

Data Availability Statement

[R1] 1.Chiou KL et al. Genomic signatures of high-altitude adaptation and chromosomal polymorphism in geladas. Nat. Ecol. Evol 6, 630–643 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R2] 2.Cai Z et al. Adaptive transcriptome profiling of subterranean Zokor, Myospalax baileyi, to high-altitude stresses in Tibet. Sci. Rep 8, 4671 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R3] 3.Zhang T et al. Phenotypic and genomic adaptations to the extremely high elevation in plateau zokor (Myospalax baileyi). Mol. Ecol 30, 5765–5779 (2021). [DOI] [PubMed] [Google Scholar]

[R4] 4.Simonson TS et al. Genetic evidence for high-altitude adaptation in Tibet. Science 329, 72–75 (2010). [DOI] [PubMed] [Google Scholar]

[R5] 5.Yi X et al. Sequencing of 50 human exomes reveals adaptation to high altitude. Science 329, 75–78 (2010). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R6] 6.Peng Y et al. Genetic variations in Tibetan populations and high-altitude adaptation at the Himalayas. Mol. Biol. Evol 28, 1075–1081 (2011). [DOI] [PubMed] [Google Scholar]

[R7] 7.Peng Y et al. Down-regulation of EPAS1 transcription and genetic adaptation of Tibetans to high-altitude hypoxia. Mol. Biol. Evol 34, 818–830 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R8] 8.Huerta-Sánchez E et al. Genetic signatures reveal high-altitude adaptation in a set of Ethiopian populations. Mol. Biol. Evol 30, 1877–1888 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R9] 9.Schweizer RM et al. Physiological and genomic evidence that selection on the transcription factor Epas1 has altered cardiovascular function in high-altitude deer mice. PLoS Genet 15, e1008420 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R10] 10.Schweizer RM et al. Broad concordance in the spatial distribution of adaptive and neutral genetic variation across an elevational gradient in deer mice. Mol. Biol. Evol 38, 4286–4300 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R11] 11.Scott GR et al. Adaptive modifications of muscle phenotype in high-altitude deer mice are associated with evolved changes in gene regulation. Mol. Biol. Evol 32, 1962–1976 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R12] 12.Storz JF High-altitude adaptation: mechanistic insights from integrated genomics and physiology. Mol. Biol. Evol 38, 2677–2691 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R13] 13.Storz JF & Cheviron ZA Physiological genomics of adaptation to high-altitude hypoxia. Annu. Rev. Anim. Biosci 9, 149–171 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R14] 14.McClelland GB & Scott GR Evolved mechanisms of aerobic performance and hypoxia resistance in high-altitude natives. Annu. Rev. Physiol 81, 561–583 (2019). [DOI] [PubMed] [Google Scholar]

[R15] 15.Storz JF & Scott GR Life ascending: mechanism and process in physiological adaptation to high-altitude hypoxia. Annu. Rev. Ecol., Evol. Syst 50, 503 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R16] 16.Storz JF et al. Evolution of physiological performance capacities and environmental adaptation: insights from high-elevation deer mice (Peromyscus maniculatus). J. Mammal 100, 910–922 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R17] 17.Stewart NB & Rogers RL Chromosomal rearrangements as a source of new gene formation in Drosophila yakuba. PLoS Genet 15, e1008314 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R18] 18.Shi J et al. Structural variants involved in high-altitude adaptation detected using single-molecule long-read sequencing. Nat. Commun 14, 8282 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R19] 19.Quan C et al. Characterization of structural variation in Tibetans reveals new evidence of high-altitude adaptation and introgression. Genome Biol 22, 159 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R20] 20.Almarri MA et al. Population structure, stratification, and introgression of human structural variation. Cell 182, 189–199. e115 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R21] 21.Mérot C et al. A roadmap for understanding the evolutionary significance of structural genomic variation. Trends Ecol. Evol 35, 561–572 (2020). [DOI] [PubMed] [Google Scholar]

[R22] 22.Villoutreix R et al. Large-scale mutation in the evolution of a gene complex for cryptic coloration. Science 369, 460–466 (2020). [DOI] [PubMed] [Google Scholar]

[R23] 23.Yan Z et al. Evolution of a supergene that regulates a trans-species social polymorphism. Nat. Ecol. Evol 4, 240–249 (2020). [DOI] [PubMed] [Google Scholar]

[R24] 24.Hager ER et al. A chromosomal inversion contributes to divergence in multiple traits between deer mouse ecotypes. Science 377, 399–405 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R25] 25.Harringmeyer OS & Hoekstra HE Chromosomal inversion polymorphisms shape the genomic landscape of deer mice. Nat. Ecol. Evol 6, 1–15 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R26] 26.Shams I et al. Oxygen and carbon dioxide fluctuations in burrows of subterranean blind mole rats indicate tolerance to hypoxic-hypercapnic stresses. Comp. Biochem. Physiol. A 142, 376–382 (2005). [DOI] [PubMed] [Google Scholar]

[R27] 27.Liu X et al. Genomic insights into zokors’ phylogeny and speciation in China. Proc. Natl Acad. Sci. USA 119, e2121819119 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R28] 28.Storz JF & Bautista NM Altitude acclimatization, hemoglobin-oxygen affinity, and circulatory oxygen transport in hypoxia. Mol. Asp. Med 84, 101052 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R29] 29.Partha R et al. Subterranean mammals show convergent regression in ocular genes and enhancers, along with adaptation to tunneling. eLife 6, e25884 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R30] 30.Davies KTJ et al. Limited evidence for parallel molecular adaptations associated with the subterranean niche in mammals: a comparative study of three superorders. Mol. Biol. Evol 35, 2544–2559 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R31] 31.Chang L et al. Akr1c1 connects autophagy and oxidative stress by interacting with sqstm1 in a catalytic-independent manner. Acta Pharmacol. Sin 43, 703–711 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R32] 32.Bochud M et al. Association of CYP3A5 genotypes with blood pressure and renal function in African families. J. Hypertens 24, 923–929 (2006). [DOI] [PubMed] [Google Scholar]

[R33] 33.Martinez C et al. Expression of paclitaxel-inactivating CYP3A activity in human colorectal cancer: implications for drug therapy. Br. J. Cancer 87, 681–686 (2002). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R34] 34.Jain IH et al. Genetic screen for cell fitness in high or low oxygen highlights mitochondrial and lipid metabolism. Cell 181, 716–727. e711 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R35] 35.Brutsaert TD et al. Association of EGLN1 gene with high aerobic capacity of Peruvian Quechua at high altitude. Proc. Natl Acad. Sci. USA 116, 24006–24011 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R36] 36.Chien T et al. GSK3β negatively regulates TRAX, a scaffold protein implicated in mental disorders, for NHEJ-mediated DNA repair in neurons. Mol. Psychiatry 23, 2375–2390 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R37] 37.Ostertag EM & Kazazian HH Twin priming: a proposed mechanism for the creation of inversions in L1 retrotransposition. Genome Res 11, 2059–2065 (2001). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R38] 38.Yang J et al. Genetic signatures of high-altitude adaptation in Tibetans. Proc. Natl Acad. Sci. USA 114, 4189–4194 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R39] 39.Hu L et al. Arctic introgression and chromatin regulation facilitated rapid Qinghai-Tibet Plateau colonization by an avian predator. Nat. Commun 13, 6413 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R40] 40.Davies KT et al. Family wide molecular adaptations to underground life in African mole-rats revealed by phylogenomic analysis. Mol. Biol. Evol 32, 3089–3107 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R41] 41.Hsia CC et al. Enhanced alveolar growth and remodeling in guinea pigs raised at high altitude. Respir. Physiol. Neurobiol 147, 105–115 (2005). [DOI] [PubMed] [Google Scholar]

[R42] 42.Llapur CJ et al. Increased lung volume in infants and toddlers at high compared to low altitude. Pediatr. Pulmonol 48, 1224–1230 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R43] 43.Li J et al. A new homotetramer hemoglobin in the pulmonary surfactant of plateau zokors (Myospalax Baileyi). Front. Genet 13, 824049 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R44] 44.Cimmino F et al. HIF-1 transcription activity: HIF1A driven response in normoxia and in hypoxia. BMC Med. Genet 20, 37 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R45] 45.Aggarwal S et al. EGLN1 involvement in high-altitude adaptation revealed through genetic analysis of extreme constitution types defined in Ayurveda. Proc. Natl Acad. Sci. USA 107, 18961–18966 (2010). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R46] 46.Parathath S et al. Hypoxia is present in murine atherosclerotic plaques and has multiple adverse effects on macrophage lipid metabolism. Circ. Res 109, 1141–1152 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R47] 47.Mendillo ML et al. HSF1 drives a transcriptional program distinct from heat shock to support highly malignant human cancers. Cell 150, 549–562 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R48] 48.Santagata S et al. High levels of nuclear heat-shock factor 1 (HSF1) are associated with poor prognosis in breast cancer. Proc. Natl Acad. Sci. USA 108, 18378–18383 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R49] 49.Gottschling DE et al. Position effect at S. cerevisiae telomeres: reversible repression of Pol II transcription. Cell 63, 751–762 (1990). [DOI] [PubMed] [Google Scholar]

[R50] 50.Nimmo ER et al. Telomere-associated chromosome breakage in fission yeast results in variegated expression of adjacent genes. EMBO J 13, 3801–3811 (1994). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R51] 51.Wang J et al. A heterochromatin domain forms gradually at a new telomere and is dynamic at stable telomeres. Mol. Cell. Biol 38, e00393–17 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R52] 52.Samonte RV & Eichler EE Segmental duplications and the evolution of the primate genome. Nat. Rev. Genet 3, 65–72 (2002). [DOI] [PubMed] [Google Scholar]

[R53] 53.Porubsky D et al. Recurrent inversion polymorphisms in humans associate with genetic instability and genomic disorders. Cell 185, 1986–2005. e1926 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R54] 54.Zhang T et al. Phylogenetic relationships of the zokor genus Eospalax (Mammalia, Rodentia, Spalacidae) inferred from whole-genome analyses, with description of a new species endemic to Hengduan Mountains. Zool. Res 43, 331–342 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]

[R55] 55.Kielbasa SM et al. Adaptive seeds tame genomic sequence comparison. Genome Res 21, 487–493 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]

PERMALINK

Genomic structural variation is associated with hypoxia adaptation in high-altitude zokors

Xuan An

Leyan Mao

Yinjia Wang

Qinqin Xu

Xi Liu

Shangzhe Zhang

Zhenglei Qiao

Bowen Li

Fang Li

Zhuoran Kuang

Na Wan

Xiaolong Liang

Qijiao Duan

Zhilong Feng

Xiaojie Yang

Sanyuan Liu

Eviatar Nevo

Jianquan Liu

Jay F Storz

Kexin Li

Abstract

Fig. 1 |. Phylogenetic relationships, elevational distributions and patterns of phenotypic variation among Eospalax.

Results

Phenotypic variation among zokors native to different elevations

Genome assembly, scaffolding and annotation

Table 1 |.

Genome evolution and possible mechanisms of hypoxia adaptation

SV identification, validation and distribution

Distribution of <1 Mb SVs in Eospalax genomes.

Fig. 2 |. Identification and distribution of large inversions (>1 Mb) and SVs (<1 Mb, INS, DEL, INV, TRA and DUP).

Large inversions (>1 Mb) in the E. baileyi genome.

Possible functional effects of SVs

Differentially expressed genes associated with SVs (<1 Mb).

Functional effects of SVs (<1 Mb).

Fig. 3 |. Intronic SVs upregulated gene expression and facilitated hypoxia adaptation.

Limited deleterious effects of large inversions.

Fig. 4 |. Possible functional effects of large inversions (>1 Mb) and rearrangement.

Rearrangement on chromosome 1 in Eospalax.

Mechanisms of inversion formation

Mechanism of short inversion (<1 Mb) formation by LINE-1.

Mechanism of large inversion (>1 Mb) formation by NAHR.

Discussion

Methods

Sample information

Phenotypic data of Eospalax

Fibroblast culture and karyotype analysis

Skin fibroblast culture.

Cellular chromosome karyotype analysis.

DNA extraction and library construction

Data generation

De novo genome assembly and annotation

Genome assembly.

Repetitive DNA annotation.

Gene annotation.

Analysis of gene family expansions and contractions

Identification of genes subject to selection

SV (<1 Mb) calling, description and verification

Illumina SV calling.

Nanopore SV calling.

Genomic region and repeat annotation for SVs.

Gene function of associated SVs.

PCR validation of SVs.

Large inversion (>1 Mb)

Identification and verification of large SVs.

Predicting telomere locations.

Inverted repeat and segmental duplications at inversion breakpoints.

Potential effect of large inversions.

Construction of homologous bin pairs.

Comparison of 3D chromatin architecture.

RNA-seq analysis

Dual-luciferase assay

Prediction of SFTPD protein structure

3D model reconstruction of genomes

Ethics statement

Reporting summary

Extended Data

Extended Data Fig. 1 |. Genome assembly and annotation of E. baileyi (plateau zokor v3.0).

Extended Data Fig. 2 |. SV (<1 Mb) dataset summary (LR-SV and SR-SV).