Integrated multi-omics reveals population-specific mechanisms of cardiovascular disease in high-altitude environments

Weiyu Meng; Jiajun Yang; Xi Wang; Xiaobing Zhai; Gang Luo; Yuyang Sha; Hongxin Pan; Wei Xu; Henry HY Tong; Edmundo Patricio Lopes Lao; Ci-ren Zhong-ga; Song Cui; Xiantao Song; Kefeng Li

doi:10.1016/j.isci.2026.115109

. 2026 Feb 20;29(3):115109. doi: 10.1016/j.isci.2026.115109

Integrated multi-omics reveals population-specific mechanisms of cardiovascular disease in high-altitude environments

Weiyu Meng ^1,⁶, Jiajun Yang ^2,⁶, Xi Wang ^3,⁶, Xiaobing Zhai ¹, Gang Luo ¹, Yuyang Sha ¹, Hongxin Pan ¹, Wei Xu ¹, Henry HY Tong ¹, Edmundo Patricio Lopes Lao ⁴, Ci-ren Zhong-ga ⁵, Song Cui ^2,^∗, Xiantao Song ^2,^∗∗, Kefeng Li ^1,^7,^∗∗∗

PMCID: PMC12972740 PMID: 41816299

Summary

High-altitude populations exhibit lower cardiovascular disease incidence and mortality with significant differences between indigenous highlanders and migrants. However, related genetic adaptation and population-specific mechanisms remain underexplored. We conducted a comprehensive multi-omics investigation of three cardiovascular disease patient cohorts from distinct altitudes: indigenous high-altitude residents (IHA, n = 31), high-altitude migrants (HAM, n = 18), and low-altitude residents (LAR, n = 50). IHA cardiovascular patients exhibited distinctive genetic signatures with significant enrichment of UGT1A family gene variants. Lipidomic profiling identified 118 differential lipid species between IHA and HAM patients, with significant enrichment in sphingolipid metabolism pathways. Integration of genomic and lipidomic data identified 102 significant gene-metabolite associations in IHA patients, particularly between UGT1A variants and sphingomyelin species. Comparative analysis with LAR patients revealed both shared and population-specific metabolic signatures. Our findings offer unique perspectives on cardiovascular disease in high-altitude environments, revealing complex interactions between genetic adaptation, environmental exposure, and disease pathophysiology.

Subject areas: human physiology, omics

Graphical abstract

Highlights

•
Establish multicenter cardiovascular disease patient cohorts of distinct altitudes
•
Multi-omics analysis reveals population-specific cardiovascular mechanisms
•
UGT1A gene variants enriched in indigenous high-altitude cardiovascular patients
•
118 differential lipids identified with sphingolipid pathway enrichment

Human physiology; Omics

Introduction

Millions people worldwide reside at high altitudes, defined as regions exceeding 2,500 m above sea level, facing chronic exposure to hypobaric hypoxia. This environmental challenge fundamentally alters cardiovascular physiology and disease pathogenesis.¹ Previous studies have revealed distinctive cardiovascular disease patterns in high-altitude populations, with lower rates of cardiovascular disease (CVD) and reduced cardiovascular mortality compared to low-altitude residents.²^,³

The cardiovascular response to hypobaric hypoxia demonstrates significant population differences. Indigenous high-altitude (IHA) residents with multigenerational adaptation have resided at elevations above 2,500 m for thousands of years. This extended timescale has allowed for the development and selection of genetic adaptations that optimize cardiovascular function and metabolic mechanisms under chronic hypoxia.⁴^,⁵ In contrast, high-altitude adapted migrants (HAMs) exhibit acquired adaptive mechanisms including the reversible adjustments of physiological phenotypes. However, they often display higher rates of maladaptive cardiovascular responses.¹^,⁶^,⁷

Investigating cardiovascular disease patients experiencing identical hypobaric hypoxia but with different evolutionary backgrounds, such as IHA and HAM, provides a unique natural experiment. This approach illuminates how environmental exposure and genetic adaptation reshape disease pathophysiology and offers insights that cannot be captured from conventional sea-level studies.³ However, the relevant genetic variations and metabolic remodeling underlying these population-specific differences remain incompletely characterized. This represents a critical gap in the understanding of human adaptation to environmental extremes and the pathophysiology of high-altitude cardiovascular disease. As climate change and socioeconomic factors continue to influence population distributions globally, elucidating these adaptive mechanisms becomes increasingly important for high-altitude disease prevention and management.⁵

Metabolic remodeling, particularly in lipid metabolism, represents a critical yet understudied dimension of high-altitude physiology, especially in the context of cardiovascular disease. Previous research has documented differences in conventional lipid parameters between cardiovascular disease patients from high-altitude and low-altitude regions. These include higher high-density lipoprotein cholesterol levels and lower triglyceride concentrations in some highland populations.⁸^,⁹ This metabolic pattern potentially contributes to their reduced atherosclerotic burden under chronic hypoxic stress.¹⁰^,¹¹^,¹² However, these traditional lipid measurements provide only limited insight into the complex metabolic adaptations occurring at the molecular level.

Lipids serve numerous essential functions in cardiovascular physiology, including membrane structural integrity, signal transduction, energy storage, and inflammatory regulation.¹⁰ Both these functions may be modulated in response to hypobaric hypoxia. The advent of high-throughput metabolomic technologies has enabled more comprehensive characterization of lipid species, yet these approaches have been minimally applied to high-altitude research. The few existing metabolomic studies in this field have primarily focused on acute high-altitude exposure in low-altitude residents (LARs) with cardiovascular disease.⁸^,¹³ Comparative studies examining indigenous highlanders, migrants, and lowlanders with cardiovascular disease remain scarce. Consequently, the lipidome of high-altitude cardiovascular disease patients remains largely uncharacterized. This critical gap limits the understanding of how metabolic remodeling influences cardiovascular pathophysiology in high-altitude environments and obscures potentially valuable therapeutic targets.

In this study, we establish distinct altitude cardiovascular disease patient cohorts to understand the complex interplay between genetic adaptation, environmental exposure, and metabolic remodeling in high-altitude environments. We performed whole-exome sequencing in IHA and HAM to identify genetic variants that potentially confer differential cardiovascular disease susceptibility and mortality. Additionally, we performed untargeted lipidomic profiling across IHA, HAM, and LAR to characterize population-specific metabolic remodeling. Through integrated analysis of genetic variants and lipidomic data, we aimed to elucidate functional relationships between genetic adaptation and metabolic remodeling. Our findings provide mechanistic insights into high-altitude cardiovascular adaptation and disease pathophysiology, with potential implications for developing targeted therapeutic strategies for cardiovascular disease management in both high-altitude and sea-level populations exposed to hypoxic conditions.

Results

Overview of study population

Three independent cohorts with a total of 99 cardiovascular disease patients were included in this study (Table 1). In the high-altitude cohort, this study enrolled 49 cardiovascular disease patients residing at high-altitude environment in Tibet, comprising 31 IHA and 18 HAM. In the low-altitude cohort, another 50 cardiovascular disease patients residing at low-altitude in Beijing were also enrolled. More detailed description of IHA, HAM and LAR groups can be found in the materials and method part. The mean age of IHA was 61.8 ± 6.6 years, and the BMI was 23.2 ± 3.4 kg/m², while the mean age of HAM was 59.2 ± 7.4 years, and the BMI was 23.7 ± 4.1 kg/m².

Table 1.

Characteristics of three independent cohorts

Characteristics	High-altitude cohort		Low-altitude cohort	p value
Characteristics	IHA (n = 31)	HAM (n = 18)	LAR (n = 50)	p value
Age in years, mean (SD)	61.8 (6.6)	59.2 (7.4)	61.3 (9.5)	0.231
BMI (kg/m²), mean (SD)	23.2 (3.4)	23.7 (4.1)	22.8 (4.6)	<0.001
Altitude, mean (SD)	3,968.4 (7.5)	3,885.7 (6.2)	679.3 (7.9)	<0.001
Years at altitude, mean (SD)	59.7 (6.1)	23.4 (9.3)	60.5 (7.9)	0.017
Sex, n (%)	–	–	–	0.938
Male	18 (58.1%)	10 (55.6%)	27 (54.0%)	–
Female	13 (41.9%)	8 (44.4%)	23 (46.0%)	–
Ethnicity, n (%)	–	–	–	<0.001
Han	0 (0.0%)	18 (100.0%)	50 (100.0%)	–
Tibetan	31 (100.0%)	0 (0.0%)	0 (0.0%)	–
Current smoker, n (%)	–	–	–	0.049
Yes	13 (41.9%)	8 (44.5%)	10 (20.0%)	–
No	18 (58.1%)	10 (55.6%)	40 (80.0%)	–
Hypertension, n (%)	–	–	–	0.026
Yes	10 (32.3%)	11 (61.1%)	13 (26.0%)	–
No	21 (67.7%)	7 (38.9%)	37 (74.0%)	–
Marital status, n (%)	–	–	–	0.533
Single/separated	3 (32.3%)	2 (61.1%)	9 (18.0%)	–
Married	28 (67.7%)	16 (38.9%)	41 (82.0%)	–
Alcohol consumption, n (%)	–	–	–	0.427
Yes	17 (54.8%)	10 (55.6%)	21 (42.0%)	–
No	14 (45.2%)	8 (44.4%)	29 (58.0%)	–
Statin, n (%)	–	–	–	0.169
Yes	25 (80.6%)	11 (61.1%)	41 (82.0%)	–
No	6 (19.4%)	7 (38.9%)	9 (18.0%)	–
ACEI/ARB, n (%)	–	–	–	0.371
Yes	15 (48.4%)	11 (61.1%)	32 (64.0%)	–
No	16 (51.6%)	7 (38.9%)	18 (36.0%)	–
CCB, n (%)	–	–	–	0.721
Yes	19 (61.3%)	12 (66.7%)	35 (70.0%)	–
No	12 (38.7%)	6 (33.4%)	15 (30.0%)	–
βblocker, n (%)	–	–	–	0.117
Yes	11 (35.5%)	9 (50.0%)	12 (24.0%)	–
No	20 (19.4%)	9 (50.0%)	38 (76.0%)	–

Open in a new tab

HAMs, high-altitude adapted migrants; IHAs, indigenous high-altitude residents; LARs, low-altitude residents; ACEI, angiotensin-converting enzyme inhibitor; ARB, angiotensin receptor blocker; CCBs, calcium channel blockers.

Identification of high-frequency rare variants in HACVD patients

Whole-exome sequencing (WES) was performed in 49 high-altitude cardiovascular disease (HACVD) cases, including 31 IHA patients and 18 HAM patients. After quality control of the exome sequencing data, an average of 80,374,097 clean reads (12,056,114,693 base pairs) were obtained per sample (Table S1). The average GC content was 51.61%. An average of 100.0% of reads were aligned to the reference genome, with 87.47% mapping to unique positions within the reference genome. The average proportion of duplicate reads was 3.6%. After duplicate removal, the mean sequencing depth reached 106.74X (Table S2). We annotated an average of 32,456 synonymous variants and 25,370 missense variants in the coding regions. And we also annotated an average of 402 loss-of-function (LOF) variants resulted in the frameshift, 31 LOF variants resulted in the conversion of stop codons to non-stop codons, 59 LOF variants led to the loss of start codons, and 16,099 LOF variants in splice site regions affected transcript splicing or supply (Table S3).

To understand the mutation distributions of genes across our cohort, we analyzed the characteristics of rare damaging variants that contributed to the burden for each patient. The exome-wide rare variant burden (ERVB) displayed considerable inter-individual variability, with 8.2% patients exhibiting substantially higher mutation loads than others (Figure 1A). Then, we implemented a filtration strategy to identify high-frequency rare variants in HACVD patients. Their variant distributions were notable enriched in our HACVD population relative to global reference populations. A total of 142 high-frequency rare variants in HACVD patients were identified, including 114 damaging loss-of-function (LOF-D) variants and 28 damaging missense (Mis-D) variants (Figure 1B). They appeared with over 10% carrier frequencies in our 49 HACVD cases, while rare in general populations (MAF <1%). The resulting heatmap (Figure 1A) was constructed to visualize the distributions of top 20 genes with the highest Mis-D variant ratios and top 20 genes with the highest LOF-D variant ratios.

Genetic characterization of high-frequency rare variants in high-altitude cardiovascular disease patients

(A) Hierarchical clustering and heatmap visualize distributions of top 20 genes with the highest Mis-D variant ratios and top 20 genes with the highest LOF-D variant ratios across 49 high-altitude cardiovascular disease patients. The upper histogram displays the exome-wide rare variant burden (ERVB) for each patient. Unsupervised hierarchical clustering was performed using the Euclidean distance metric and Ward’s minimum variance method to minimize the total within-cluster variance.

(B) Multiple-trait Manhattan plot representing the REVEL distribution of missense variants in 49 high-altitude cardiovascular disease patients. The two dotted lines indicate the common thresholds.

(C) Gene interaction network analysis of genes with high-frequency rare variants in high-altitude cardiovascular disease patients. Nodes represent genes and are colored according mutation type. Edges represent different types of functional relationships. Gray triangles indicate additional indirectly connected nodes. HAM, high-altitude adapted migrants; IHA, indigenous high-altitude residents; Mis-D, damaging missense variants; LOF-D, damaging loss-of-function variants.

To better explored how these gene associate with HACVD, we conducted gene-gene interaction network analysis using the genes with the top 20 Mis-D/LOF-D variant ratios (Figure 1C). It revealed multiple connections, including shared protein domains (43.75%), physical interactions (32.14%), and co-expression (23.21%) among HACVD susceptibility genes. Notably, the densely connected module predominantly comprising UGT1A family genes, exhibited extensive interconnectivity, suggesting potential coordinated functional roles in high-altitude cardiovascular disease.

Comparative analysis of indigenous high-altitude residents and high-altitude adapted migrants

Differential genetic profiles between IHA and HAM cardiovascular disease patients

To investigate potential population-specific genetic adaptations in high-altitude cardiovascular disease (HACVD), we compared the mutational profiles of 142 high-frequency rare variants in HACVD patients between IHA (n = 31) and HAM (n = 18). To justify our sample sizes and assess the statistical power of genetic comparisons, we conducted both a priori power analyses for differences between IHA and HAM population. As shown in Figure 2A and Table S4, while most genes exhibited comparable mutation rates between both groups, several genes showed substantial population-specific enrichment. Further Fisher’s exact test identified 18 genes with significantly different mutation rates between IHA and HAM populations (p < 0.05). To assess the robustness of the detected genetic differences between IHA and HAM populations, we conducted comprehensive post-hoc power analyses for all 18 genes showing significant mutation rate differences (Table S4). The 16 genes significantly enriched in IHA patients demonstrated consistently large effect sizes. In contrast, the two genes enriched in HAM patients exhibited smaller effect sizes and consequently lower achieved power. Among these, 16 genes showed a significant enrichment of mutations in IHA cardiovascular disease patients when compared with HAM patients (Figure 2B; Table S5). These IHA-enriched genes included multiple members of the UGT1A family (UGT1A1, UGT1A3, UGT1A4, UGT1A5, UGT1A6, UGT1A7, UGT1A9, UGT1A10, and UGT1A8), as well as ASB15, TOGARAM1, TTLL8, GPR142, HRNR, SCIN, and NUP43. In contrast, only two genes, including ZDHHC11B and FAM135A, demonstrated significantly higher mutation rates in HAM cardiovascular disease patients.

Comparative analysis of indigenous high-altitude residents (IHAs) and high-altitude adapted migrants (HAMs)

(A) Scatterplot showing the distribution of mutation frequencies for 142 genes with high-frequency rare variants of high-altitude cardiovascular disease patients in IHA and HAM populations. Each dot represents a gene, with orange dots indicating genes with enrichment of mutations in IHA and blue dots representing genes with enrichment of mutations in HAM. Selected genes with the highest mutation rates are labeled.

(B) 18 genes with significantly different mutation frequencies between IHA and HAM groups were showed. Dark blue indicated higher frequency in IHA patients and light blue indicated higher frequency in HAM patients.

(C) Functional enrichment analysis of the 16 genes with significantly higher mutation rates in IHA cardiovascular disease patients. The upper (blue bars) shows biological process enrichment. The lower (orange bars) displays tissue-specific enrichment. Gene symbols in parentheses list the specific genes contributing to each enriched category.

(D) Orthogonal partial least squares-discriminant analysis (OPLS-DA) scores plot demonstrating clear separation between IHA (green circles) and HAM (red circles) groups.

(E) Variable importance in projection (VIP) scores from the OPLS-DA model. Lipid species are arranged in descending order of their VIP scores. Lipids with VIP scores >2.0 are considered major contributors to the group separation.

(F) The volcano plot represents the distribution of these differential metabolites, with significantly upregulated metabolites in IHA patients shown in orange and downregulated metabolites shown in blue. Points with VIP values ≥ 1 are represented as circles, while those with VIP values <1 are shown as triangles. The horizontal dashed line indicates the significance threshold (q value = 0.05), while the vertical dashed lines represent the fold change thresholds.

(G) Distribution of differential lipid species across major lipid categories. The stacked bar chart illustrates the distribution of upregulated and downregulated lipids in each category.

To understand the potential functional implications of these IHA-enriched variants, we performed comprehensive enrichment analysis of the 16 genes with significantly higher mutation rates in IHA patients. As shown in Figure 2C and Table S6, enrichment analysis at the biological process level revealed significant associations with negative regulation of cellular glucuronidation (p = 7.34 × 10⁻²⁸), tamoxifen metabolism (p = 1.59 × 10⁻⁷), and cell projection assembly (p = 1.60 × 10⁻³). The strong enrichment in glucuronidation-related processes, primarily driven by the UGT1A family genes, suggests potential high-altitude cardiovascular adaptation in indigenous populations. Tissue-specific enrichment analysis further revealed significant associations with heart (p = 1.58 × 10⁻¹⁰), liver (p = 3.16 × 10⁻⁸), and kidney (p = 3.98 × 10⁻⁶) tissues. Additionally, the significant enrichment in liver and kidney tissues suggests that adaptation mechanisms may involve multiple organ systems beyond the cardiovascular system, particularly those involved in metabolic processes.

Lipidomic profiles reveal distinct metabolic adaptations between population groups

Building upon the observed genetic differences between IHA and HAM group, we sought to determine whether these genetic variations might be reflected in distinct metabolic phenotypes. We performed comprehensive lipidomic profiling of plasma samples from all 49 high-altitude cardiovascular disease patients (31 IHA and 18 HAM). A total of 1,025 lipids was detected in all the plasma samples. After quality control filter, 922 lipids with excellent analytical reproducibility (median inter-day coefficient variation [CV] of 10.1%) were used for further analysis (Table S7). Categories included 672 glycerophospholipids, 105 glycerolipids, 97 sphingolipids, 32 saccharolipids, 11 sterollipids, four fatty acyls, and one prenollipids.

Orthogonal partial least squares discriminant analysis (OPLS-DA) was employed to distinguish the metabolic profiles between IHA and HAM patients. The reliability of the OPLS-DA model was validated using the leave-one-out cross-validation (LOOCV). Q² value in the first component (Component 1) was over 0.5 considered as a reliable model. As shown in Figure 2D, clear separation was demonstrated between the two population groups, indicating significant metabolic differences between indigenous residents and adapted migrants with cardiovascular disease at high altitude. To identify the specific metabolites contributing significantly to the observed group separation, we calculated Variable Importance in Projection (VIP) scores from the OPLS-DA model (Figure 2E; Table S8). Numerous lipid species demonstrated high VIP scores, indicating their substantial contribution to the model.

Based on stringent filtering criteria (VIP score ≥ 1; fold change [FC] > 1.2 or <0.83, and the false discovery rate [FDR] < 0.05), we identified 118 differentially abundant lipid species between IHA and HAM cardiovascular disease patients (Table S8). The volcano plot in Figure 2F visually represents the distribution of these differential metabolites. Among the 118 differentially abundant metabolites, 84 (71.2%) were significantly upregulated in IHA patients compared to HAM patients, while 34 (28.8%) were significantly downregulated (Figure 2G; Table S8). Classification of these differential metabolites by lipid categories revealed distinct patterns of regulation across major lipid classes (Table S9). Glycerophospholipids constituted the largest proportion of both upregulated and downregulated lipid species. Sphingolipids represented a substantial proportion of upregulated lipids, and Glycerolipids accounted for a smaller fraction of the differential lipid profile. Notably, saccharolipids were exclusively downregulated in IHA patients, and none upregulated.

In-depth analysis of differential lipids reveals population-specific metabolic signatures and metabolic remodeling

To further characterize the metabolic differences between IHA and HAM populations, we conducted comprehensive analyses of the 118 differential lipid species identified in our initial screening. Hierarchical clustering analysis of the differential lipids revealed distinct expression patterns between the two population groups (Figure 3A). The heatmap visualization demonstrated clear clustering of patients according to their population group (IHA versus HAM), with well-defined blocks of coordinately regulated lipid species.

Comprehensive analysis of differential lipids and genetic-metabolic integration in high-altitude cardiovascular disease

(A) Hierarchical clustering heatmap of 118 differential lipid species between indigenous high-altitude residents (IHAs) and high-altitude adapted migrants (HAMs). Red and blue coloration indicates upregulation and downregulation, respectively. The dendrogram demonstrates clear population-specific clustering of lipid profiles, with distinct blocks of coordinately regulated lipids.

(B) Circular correlation network of differential lipids, displaying strong positive correlations (R ≥ 0.85, p < 0.05) among lipid species. Nodes are color-coded by lipid class. The extensive interconnections demonstrate coordinated regulation across lipid classes.

(C) Pathway enrichment analysis of differential lipids. The bubble plot displays significantly enriched pathways. Sphingolipid metabolism emerged as the most significantly enriched pathway, followed by PI3K-AKT-mTOR signaling pathway.

(D) Receiver operating characteristic (ROC) curves based on lipid biomarkers. The graph displays the top-performing individual lipids and lipid combinations.

(E) Circular visualization of gene-metabolite associations within the IHA population. The inner circle represents the 18 differentially mutated genes between IHA and HAM populations, while the outer circle displays their significantly associated lipid species. The color of the outer circle conveys the strength of association measured through coefficients. The visualization reveals distinct gene-lipid association patterns, including coordinated regulation of sphingomyelins by UGT1A gene family members and diverse lipid associations of ASB15. HAM, high-altitude adapted migrants; IHA, indigenous high-altitude residents; PC, phosphatidylcholine; PI, phosphatidylinositol; SM, sphingomyelin; PE, phosphatidylethanolamine; PA, phosphatidic acid; DG, diglyceride.

To explore potential relationships between the differential lipids, we performed correlation analysis across all 118 metabolites. Stringent filtering criteria (p < 0.05 and correlation coefficient r ≥ 0.85) were applied to identify strongly correlated lipid pairs, which were then used to construct a correlation network (Figure 3B). The correlation network revealed exclusively positive correlations among the differential lipids, suggesting synchronized regulation of these metabolites. The network visualization demonstrated extensive interconnections between glycerophospholipids, which formed the largest and most densely connected component of the network. Sphingolipids also displayed strong positive correlations both within their class and with glycerophospholipids, indicating coordinated regulation across lipid categories. Glycerolipids and saccharolipids showed fewer but nonetheless significant positive correlations. The existence of such strong correlations across numerous lipid pairs suggests that the observed metabolic differences between IHA and HAM populations likely reflect systematic alterations in lipid metabolism pathways rather than isolated changes in individual lipid species.

To identify the biological pathways associated with the differential lipids, we performed pathway enrichment analysis (Figure 3C). The analysis revealed significant enrichment in several key pathways, with “sphingolipid metabolism, integrated pathway” emerging as the most significantly enriched pathway. This finding is particularly notable given the substantial proportion of sphingolipids among upregulated lipids in IHA patients, suggesting that alterations in sphingolipid metabolism may play a central role in high-altitude cardiovascular adaptation. The “PI3K-AKT-mTOR signaling pathway” was also highly enriched, indicating potential involvement of this key regulatory pathway in mediating cardiovascular responses to high-altitude environments. Additional enriched pathways included “T cell antigen receptor pathway,” “T cell receptor and co-stimulatory signaling,” and “insulin signaling,” suggesting that the lipid alterations may influence multiple aspects of cellular signaling and immune function in high-altitude cardiovascular disease.

To assess the discriminatory power of the differential lipids for distinguishing between population-specific metabolic signatures of high-altitude cardiovascular disease, we employed support vector machine (SVM) algorithms to develop classification models based on lipid profiles. The permutation test (n = 1,000 times) was used to confirm the robustness of ROC analysis. As shown in Figure 3D and Table S10, the ROC curve analysis demonstrated that several individual lipids exhibited strong ability to discriminate between IHA and HAM patients with cardiovascular disease. The top-performing individual lipids included SM(d42:6) with an AUC of 0.874 (95% CI, 0.745–0.996, permutation p = 0.001), PC (36:4e) with an AUC of 0.865 (95% CI, 0.767–0.965, permutation p = 0.004), PC (33:0p/rep) with an AUC of 0.859 (95% CI, 0.655–0.979, permutation p < 0.001), SM(d44:6) with an AUC of 0.858 (95% CI, 0.692–0.972, permutation p < 0.001), and SM(d42:2) with an AUC of 0.857 (95% CI, 0.623–0.986, permutation p = 0.002). Notably, three of the five top-performing individual lipids were sphingomyelins (SMs), aligning with our pathway enrichment findings that highlighted the importance of sphingolipid metabolism in distinguishing between the two population groups.

Moreover, the combination of multiple lipid species substantially improved the discriminatory performance, with the combination of PC (36:2) and SM(d44:6) achieving an AUC of 0.930, and the combination of PC (36:4e) and SM(d44:6) reaching an AUC of 0.942.

Integration of genetic variant and lipid profiles reveals population-specific signatures and adaptation mechanisms associated with cardiovascular disease in high-altitude environments

To establish functional connections between the identified genetic variants and metabolic alterations, we performed comprehensive gene-metabolite association analyses to elucidate the molecular mechanisms through which population-specific genetic variants might influence lipid metabolism in high-altitude cardiovascular disease.

We first conducted gene-metabolite association analyses across the entire cohort (n = 49), using the 18 genes showing significant mutational differences and the 118 differential lipids between IHA and HAM populations. Using a stringent statistical threshold, we identified 70 significant gene-metabolite associations (Table S11). To further explore population-specific genotype-phenotype relationships, we performed stratified analyses. In the analysis of indigenous high-altitude residents (IHA, n = 31) with cardiovascular disease, 102 significant gene-metabolite associations were identified (Figure 3E; Table S12). The substantially higher number of significant associations identified in the IHA-specific analysis compared to the whole-cohort analysis (102 vs. 70) suggests that gene-metabolite relationships may be more distinct within the indigenous population.

Remarkably, multiple SM species were influenced by the UGT1A family, which represent the most prominent group of IHA-specific genetic variants. This observation suggests a coordinated influence of UGT1A variants on sphingolipid metabolism in IHA, and provides direct evidence linking the indigenous-specific genetic adaptations in glucuronidation pathways to the difference in sphingolipid metabolism. Concurrently, multiple phosphatidylcholine (PC) species were significantly influenced by ASB15. Among the identified genes, ASB15 exhibited the most extensive metabolic influence with 25 significant associations, affecting diverse lipid species including PC, phosphatidylinositol (PI), SM, and phosphatidylethanolamine (PE). GPR142 demonstrated similarly broad influence, affecting 11 different lipid species spanning multiple lipid categories, including triglycerides (TGs), SM, PI, PE, phosphatidic acid (PA), diglyceride (DG), and PC. Significantly, specific lipid species were influenced by multiple genes from distinct functional categories. For example, PI (36:3) was influenced by GPR142, SCIN, and FAM135A. Similarly, DG (16:0/18:2) was influenced by GPR142, NUP43, ZDHHC11B, and FAM135A. This co-influence on specific lipid species suggests potential functional nodes in lipid metabolism that may be particularly relevant to high-altitude cardiovascular adaptation.

Comparative analysis of high-altitude and low-altitude cardiovascular disease patients

To further elucidate the influence of the hypoxic environment of high altitude on cardiovascular disease metabolic profiles, we enrolled an additional cohort of 50 cardiovascular disease patients residing at low-altitude. Sample size determination for lipidomic analyses was conducted a priori using power calculations based on anticipated metabolic differences between altitude and population group. This three-way comparison design allowed us to dissect the contributions of genetic adaptation and environmental exposure to metabolic phenotypes in cardiovascular disease. IHA represent populations with potential genetic adaptations acquired over multigenerational exposure; HAM represent individuals with recent environmental exposure but without long-term genetic adaptations; and LAR represent the reference population without either genetic adaptations or environmental exposure to high altitude.

After performing batch effect correction between high-altitude and low-altitude batches using the ComBat method (Figure S1) to harmonize datasets from different experimental batches, we identified 37 common lipid species detected across all cohorts. Using these common metabolites, we conducted three separate differential analyses to characterize altitude-specific and population-specific metabolic signatures in cardiovascular disease. The first comparison between all high-altitude patients (IHA + HAM, n = 49) and LAR (n = 50) identified seven differential lipids, all of which were significantly upregulated in high-altitude patients (Figure 4A; Table S13). The consistent upregulation of these lipids in high-altitude patients regardless of genetic adaptation background suggests that these metabolic alterations may represent direct responses to the hypoxic environment of high altitude. The second comparison between IHA (n = 31) and LAR (n = 50) revealed six differential lipid species (Figure 4A; Table S13), all upregulated in IHA patients. The third comparison between HAM (n = 18) and LAR (n = 50) identified eight differential lipid species (Figure 4A; Table S13), all upregulated in HAM patients. To evaluate the statistical robustness of identified differential lipids across the three-way comparison, we conducted comprehensive post-hoc power analyses.

Comparative lipidomic analysis between high-altitude and low-altitude cardiovascular disease patients

(A) Radial plot displaying differential lipids identified across three group comparisons: high-altitude patients (IHA + HAM) versus low-altitude residents (LARs) in blue, indigenous high-altitude residents (IHAs) versus LAR in pink, and high-altitude adapted migrants (HAMs) versus LAR in teal. Circle size corresponds to variable importance in projection (VIP) score, with only metabolites having VIP ≥ 1 displayed. Asterisks (∗) indicate statistically significant differences (p < 0.05).

(B) Venn diagram illustrating the overlap of differential lipids among the three comparison groups.

(C) OPLS-DA score plots demonstrating separation between high-altitude and low-altitude cardiovascular disease patients. Left, combined high-altitude patients (IHA + HAM, green) versus LAR (orange); middle, IHA (blue) versus LAR (orange); right, HAM (purple) versus LAR (orange).

(D) Receiver operating characteristic (ROC) curves for the differential lipids in discriminating between cardiovascular disease patients from different altitude backgrounds. Left, ROC curves for top differential lipids in IHA + HAM vs. LAR comparison; middle, ROC curves for top differential lipids in IHA vs. LAR comparison; right, ROC curves for top differential lipids in HAM vs. LAR comparison; HAM, high-altitude adapted migrants; IHA, indigenous high-altitude residents; LAR, low-altitude residents.

The Venn diagram analysis (Figure 4B) revealed intricate patterns of overlap between the three comparisons. Three lipid species were commonly differential across all three comparisons: Cer (d18:1/26:2), PC (16:0/18:2), and PE (36:1). These universally altered lipids likely represent core metabolic responses to high-altitude environments regardless of genetic background or duration of exposure. Conversely, each comparison also yielded unique differential metabolites.

The OPLS-DA analysis demonstrated clear separation between high-altitude and low-altitude patients across all three comparisons (Figure 4C). The reliability of the OPLS-DA model was validated using the LOOCV. Q² value in the component 1 was over 0.5 considered as a reliable model. The combined high-altitude versus low-altitude comparison (left) showed moderate separation. More distinct separations were observed in the IHA versus LAR comparison (middle) and the HAM versus LAR comparison (right).

To assess the discriminatory power of the identified differential lipids, we performed ROC curve analysis (Figure 4D). The permutation test (n = 1,000 times) was used to confirm the robustness of ROC analysis. In the combined high-altitude versus low-altitude comparison (left), PC (16:0/18:2) demonstrated the highest discriminatory power (AUC = 0.656, 95% CI, 0.558–0.774, permutation p = 0.029). In the IHA versus LAR comparison (middle), PE (36:2) exhibited the strongest performance (AUC = 0.706, 95% CI, 0.591–0.812, permutation p = 0.042). In the HAM versus LAR comparison (right), PC (16:0/22:6) showed the highest discriminatory power (AUC = 0.746, 95% CI, 0.605–0.908, permutation p = 0.026).

Notably, ceramides featured prominently among the differential lipids in all three comparisons, with different ceramide species exhibiting population-specific alterations. Ceramides are important components of sphingolipid metabolism. Given the established roles of ceramides in cardiovascular pathophysiology, these findings suggest that sphingolipid metabolism may represent a key pathway in altitude-specific cardiovascular adaptations and pathologies.

Discussion

Our comprehensive multi-omics investigation of high-altitude cardiovascular disease patients has uncovered distinct genetic variants and metabolic signatures associated with high-altitude environments. The comparison between three cardiovascular disease patient populations, including IHA, HAM, and LAR, revealed population-specific adaptations at both genetic and metabolic levels. Genomic analysis identified significant enrichment of UGT1A gene family variants in IHA cardiovascular disease patients. Lipidomic profiling complemented these genetic findings by demonstrating substantial metabolic remodeling in high-altitude cardiovascular disease patients. We identified 118 differential lipid species between IHA and HAM patients despite their shared environmental exposure. This finding points to genetically influenced metabolic adaptations. The significant enrichment of sphingolipid metabolism pathways in these differential profiles suggests that membrane composition and sphingolipid signaling may be critical components of cardiovascular adaptation at high altitude. Integration of genomic and lipidomic data established 102 significant gene-metabolite associations, particularly between UGT1A variants and SM species. These associations provide functional connections between genetic variants and metabolic phenotypes. Comparison with cardiovascular disease patients from low-altitude regions further revealed altitude-specific metabolic signature. These signatures were characterized by upregulation of specific lipid species, particularly ceramides, in high-altitude patients regardless of genetic background. This pattern indicates common pathways of metabolic response to hypobaric hypoxia in cardiovascular disease. These findings establish a framework for understanding how genetic adaptation, environmental exposure, and metabolic remodeling interact to shape cardiovascular disease pathophysiology in high-altitude environments, potentially informing population-specific approaches to risk stratification and therapeutic targeting in these unique patient populations.

The significant enrichment of UGT1A family gene variants specifically in indigenous high-altitude cardiovascular disease patients represents a genetic adaptation. This finding extends our understanding of high-altitude adaptation beyond classical hypoxia response pathways and offers potential insights into unique disease protection or susceptibility mechanisms. Previous studies have focused primarily on genes involved in the HIF pathway (such as EPAS1 and EGLN1) in high-altitude populations.⁴^,⁵^,¹⁴ In contrast, our findings raise the possibility that glucuronidation processes and bilirubin metabolism may constitute additional adaptive mechanisms. These mechanisms may have evolved through multigenerational genetic adaptation and could potentially influence cardiovascular disease pathophysiology. The UGT1A gene family encodes UDP-glucuronosyltransferases that catalyze the conjugation of glucuronic acid to bilirubin and other endogenous compounds.¹⁵ The predominance of UGT1A variants in IHA cardiovascular disease patients, but not in HAM patients with the same disease phenotypes and environmental exposure, is particularly noteworthy.

This pattern suggests that genetic adaptations in these pathways may confer unique cardiovascular advantages specific to indigenous high-altitude populations. High-altitude environments are characterized by hypoxia-reoxygenation cycles and elevated UV radiation.¹^,⁴ In this context, these adaptations may potentially modulate disease manifestation and progression through altered bilirubin metabolism and enhanced antioxidant capacity. Such mechanisms could counteract hypoxia-induced oxidative stress.¹⁶ Bilirubin exhibits potent antioxidant properties that are approximately 20 times greater than glutathione and significantly more effective than α-tocopherol in preventing lipid peroxidation.¹⁷ The adaptations in bilirubin metabolism could theoretically provide substantial cardiovascular benefits that modify disease progression or presentation.¹⁷ This balance could explain differences in disease manifestation between genetically adapted and non-adapted populations exposed to the same environmental stressors. Several studies have demonstrated relationships between serum bilirubin levels and cardiovascular disease risk in various populations.¹⁸^,¹⁹ Additionally, experimental evidence suggests that bilirubin influences vascular tone through NADPH oxidase and modulation of endothelial nitric oxide synthase activity.¹⁹ These mechanisms may directly impact cardiovascular disease pathophysiology.

Beyond bilirubin metabolism, UGT1A enzymes influence cardiovascular function through glucuronidation of other endogenous compounds including steroid hormones, thyroid hormones, and vasoactive compounds, offering alternative mechanistic explanations for the observed genetic enrichment.¹⁶ Glucuronidation typically reduces the biological activity of these substrates. This suggests that variants in UGT1A genes could modulate cardiovascular signaling pathways by altering the bioavailability of key regulatory molecules.¹⁶

The distinct lipidomic profiles identified in high-altitude cardiovascular disease patients reflect substantial metabolic remodeling. This remodeling likely represents both longer-term adaptive responses and pathological processes specific to cardiovascular disease under chronic hypoxia. Our analysis revealed significant differences in IHA cardiovascular disease patients across multiple lipid classes. These bioactive lipids regulate diverse cellular processes, including apoptosis, inflammation, and vascular tone.⁹ The prominence of sphingolipids, especially ceramides and SMs, is particularly noteworthy in our study of cardiovascular disease patients at high altitude. Ceramides act as signaling molecules that regulate critical cardiovascular processes including vascular tone, endothelial function, and cardiomyocyte survival.²⁰ Under hypoxic conditions, ceramides can exert adaptive effects by triggering protective preconditioning responses. These responses occur through activation of protein kinase C and other cardioprotective signaling pathways, potentially influencing disease outcomes differently in adapted populations.²¹ As major components of membrane microdomains or “lipid rafts,” SMs influence membrane fluidity, receptor trafficking, and signal transduction. These properties directly impact cardiovascular function and disease processes under hypoxic stress.²² Furthermore, SMs serve as reservoirs for ceramide generation through sphingomyelinase activity, providing a mechanism for rapid signaling responses to environmental challenges that may differ between indigenous and migrant patients.²³

The identification of both shared and population-specific metabolic signatures between IHA and HAM cardiovascular disease patients compared to LAR cardiovascular disease patients provides insight into how disease processes may be differently modified by short-term versus long-term adaptation to high altitude. Common alterations likely represent immediate metabolic responses to hypoxic stress in the context of cardiovascular disease. In contrast, population-specific signatures may reflect longer-term adaptations influenced by genetic background that modify disease progression. The unique upregulation of SMs in IHA patients suggests that specific sphingolipid species may be particularly important in indigenous adaptation to cardiovascular stress. This importance may operate through effects on membrane fluidity and signaling platform organization that influence cellular responses to chronic hypoxia and disease stimuli.²²^,²⁴ Similarly, the HAM-specific upregulation of phosphatidic acids, PI and PE may represent compensatory metabolic adjustments in the absence of genetic adaptations, potentially explaining differences in disease manifestation between populations.²⁴ These patterns of metabolic remodeling in cardiovascular disease patients align with the concept of “metabolic flexibility” in environmental adaptation. Alterations in lipid metabolism may provide rapid and reversible mechanisms to maintain cardiovascular function under challenging conditions, complementing slower genetic adaptations that evolve over generations.²⁵

The extensive network of gene-metabolite associations identified in our study of high-altitude cardiovascular disease patients provides mechanistic insights into how genetic adaptations may translate into functional metabolic phenotypes that influence disease manifestation. Particularly striking are the consistent correlations between UGT1A family variants and SM species in indigenous cardiovascular disease patients. These correlations suggest that population-specific genetic adaptations may influence lipid homeostasis and thereby modify cardiovascular disease processes at high altitude. This relationship between UGT1A variants and sphingolipid metabolism may be mediated through several potential mechanisms in the context of cardiovascular disease. UGT1A variants could influence sphingolipid levels by altering the glucuronidation of lipid precursors or regulatory molecules, directly affecting membrane composition in cardiovascular tissues.⁹^,¹⁶ An alternative or complementary hypothesis is that UGT1A variations in IHA patients impair bilirubin glucuronidation, leading to UCB accumulation that, under chronic hypoxic and cardiovascular stress, generates a pro-oxidant environment (Figure S2). This redox imbalance activates neutral sphingomyelinase (nSMase) while simultaneously upregulating SM synthase activity, shifting the equilibrium toward SM production in cardiomyocytes and vascular smooth muscle cells.²⁶ Elevated reactive oxygen species also enhance SM biosynthesis by modulating the activity of serine palmitoyltransferase.²⁷ In cardiovascular tissue, SM accumulation in plasma membranes and lipoproteins promotes endothelial dysfunction, impairs nitric oxide signaling, and contributes to atherosclerotic plaque formation.²⁸ Furthermore, SM-enriched membrane microdomains disrupt insulin receptor signaling and calcium homeostasis in cardiomyocytes, compromising contractile function.²⁹ This population-specific genetic-metabolic signature likely reflects distinct evolutionary pressures experienced by indigenous populations over thousands of years of high-altitude residence. These associations might provide a framework for personalized cardiovascular risk assessment and treatment selection based on integrated genetic and metabolic profiling.

The identification of population-specific genetic and metabolic signatures in cardiovascular disease patients at high altitude has significant implications for clinical practice and therapeutic development.⁷ Our findings suggest that a “one-size-fits-all” approach to cardiovascular disease management may be inadequate for high-altitude populations, particularly for indigenous residents with distinct genetic adaptations that modify disease processes. The enrichment of UGT1A variants specifically in indigenous high-altitude cardiovascular disease patients indicates that bilirubin metabolism and antioxidant pathways may represent important therapeutic targets for this specific population. Similarly, the prominent alterations in sphingolipid metabolism observed across high-altitude cardiovascular disease patients suggest that targeting ceramide and SM pathways may offer therapeutic strategies specific to high-altitude cardiovascular disease.²⁰^,²³ Sphingolipid-modulating agents, which are currently under investigation for various cardiovascular indications. Such targeted approaches could address the unique aspects of cardiovascular disease manifestation in high-altitude environments more effectively than conventional treatments.

Moreover, our results highlight the importance of considering environmental context and population history in both cardiovascular research and clinical practice. This is particularly relevant for indigenous high-altitude populations who migrate to lower altitudes or for low-altitude populations who relocate to high-altitude regions. In these scenarios, mismatches between genetic adaptations and environmental conditions may influence disease susceptibility and progression. The development of population-specific and environment-aware clinical guidelines for cardiovascular disease represents an important step toward more effective disease management in diverse populations with unique adaptation profiles, potentially improving both prevention strategies and treatment outcomes.

Limitations of the study

The absence of objective physiological adaptation markers (e.g., hemoglobin concentration and oxygen saturation) in our HAM cohort limits our ability to verify the completeness of individual-level acclimatization. Another limitation of our study is the absence of direct bilirubin measurements. Future studies with prospective enrollment, more detailed information (e.g., dietary information and bilirubin measurements), and serial physiological monitoring would further strengthen the interpretation of genetic and environmental contributions. Validation of the top lipid markers or gene-lipid associations in an independent set or by targeted assays represent another important future direction.

Resource availability

Lead contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Kefeng Li (kefengl@mpu.edu.mo).

Materials availability

This study did not generate new unique reagents.

Data and code availability

•
Lipidomics data have been deposited at Metabolomics Workbench (https://www.metabolomicsworkbench.org) as study ID: ST004571, and are publicly available as of the date of publication. The WES data reported in this study cannot be deposited in a public repository, because local law prohibits depositing raw genomic data in public repositories. In addition, summary statistics describing these data are accessible in Supplementary tables.
•
This article does not report original codes.
•
Any additional information required to reanalyze the data reported in this article is available from the lead contact upon request.

Acknowledgments

We want to particularly acknowledge all the study participants, and the clinical support from Lhasa People’s Hospital and Beijing Anzhen Hospital for collecting samples. This project was funded by The Science and Technology Development Funds (FDCT) of Macao (0033/2023/RIB2), Capital’s Funds for Health Improvement and Research (no.2022-1-2061), Beijing Nova Program (20220484222), Key Project of Lhasa Science and Technology Bureau (no. LSKJ202103), Natural Science Foundation of Tibet Autonomous Region (no. XZ2021ZR-ZY29 [Z], and ZRK X2021000200), and the startup fund from Lhasa People’s Hospital (SYKY2021009) with the submission approval code of (fca.0386.2733.c). The sponsor had no role in the design and conduct of the study; the collection, management, analysis, and interpretation of the data; the preparation, review, or approval of the manuscript; or the decision to submit the manuscript for publication.

Author contributions

W.M. performed the data analysis and drafted the manuscript. J.Y. and X.W. supplied clinical material and contributed data to the study. X.Z., G.L., Y.S., H.P., and W.X. devised methodology and conducted data analysis. W.M., H.T., E.P.L.L., C.-r.Z.-g., and K.L. conceptualized the study. W.M., S.C., X.S., and K.L. reviewed and edited the manuscript.

Declaration of interests

The authors declare no conflict of interest.

STAR★Methods

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Biological samples

Human blood samples	Beijing Anzhen Hospital	N/A
Human blood samples	Lhasa People’s Hospital	N/A

Deposited data

Lipidomics dataset	this paper	Metabolomics Workbench³⁰ Study ID: ST004571

Software and algorithms

BWA(v0.7.17)	Li H. et al.	http://bio-bwa.sourceforge.net/
GATK(v4.4.0.0)	Broad Institute	https://gatk.broadinstitute.org/hc/enus/articles/
LipidSearch (v4.1)	Thermo Fisher Scientific	https://www.thermofisher.cn/
DNBSEQ	BGI	https://www.mgi-tech.com/DNBSEQ-Technology/
GraphPad Prism(version 8.0.2)	GraphPad Software	https://www.graphpad.com/

Open in a new tab

Experimental model and study participant details

Study population and sample collection

The IHA group comprises individuals of Tibetan ancestry whose families have continuously inhabited regions above 2,500m for at least three generations, representing populations with millennia of genetic adaptation to high-altitude environments. These individuals possess distinctive genetic signatures reflecting ancestral selection for environment resilience. The HAM group comprises individuals of Han Chinese ancestry who permanently relocated to high-altitude regions (>2,500m) at least five years prior to study enrollment. This population represents a model of environmentally induced physiological adaptation without substantial genetic selection for altitude, as they descend from lowland Han Chinese ancestors with no historical high-altitude residence. The minimum five-year adaptation period was established to ensure completion of major physiological acclimatization processes while allowing investigation of environmentally-induced adaptations distinct from the genetic adaptations observed in indigenous highlanders. The LAR group consists of individuals of Han Chinese ancestry who have continuously resided at elevations below 1,000m throughout their lifetime, with no familial history of high-altitude habitation for at least three generations. This population serves as a reference group without either genetic adaptations or environmental exposure to high-altitude hypoxia. Genetic ancestry was confirmed through self-reported family history to ensure homogeneity of ethnic background within this group.

In our study, total 49 cardiovascular disease patients in Lhasa People’s Hospital were enrolled, comprising 31 IHA patients and 18 HAM patients. An additional cohort of 50 cardiovascular disease patients in Beijing Anzhen Hospital, who residing at low-altitude, were also enrolled. The enrollment criteria included confirmed diagnosis of cardiovascular disease based on clinical, imaging, and laboratory evaluations according to established international guidelines. Participants with severe comorbidities, including malignancies, autoimmune disorders, or severe infections, were excluded. More detailed inclusion and exclusion criteria can be found in the Supporting Information. Our study protocol adhered to the principles outlined in the World Medical Association Declaration of Helsinki for medical research involving human subjects. It was reviewed and approved by the Institutional Review Boards (IRBs) of Beijing Anzhen Hospital (IRB#: AZHEC2016-0516, AZHEC2017-0813, and AZHEC2022-0724) and Lhasa People’s Hospital (IRB#: 2018-XN-03). Verbal and written informed consent were obtained from all participants prior to study enrollment. After an overnight fast, blood samples were collected from each patient. Theses samples were processed according to standardized protocols and stored at -80°C until analysis.

Method details

Whole-exome sequencing

Whole-exome sequencing and quality control

Genomic DNA was extracted from blood samples and subsequently fragmented to generate fragments predominantly within the 200-300 bp range. The DNA fragments underwent end-repair, and the library underwent linear amplification through LM-PCR (ligation-mediated PCR). The amplified library was then hybridized with RNA probes targeting exonic regions, followed by washing steps to remove non-targeted fragments. The captured library was further amplified and subjected to sequencing on an Illumina platform. Raw sequencing data were processed using DNBSEQ base calling software and converted to FASTQ format files. These files contained paired-end reads with associated quality scores and represented the raw dataset for subsequent analysis.

Quality control metrics, including read length distribution, and base quality distribution, were assessed to ensure data integrity before proceeding to downstream analyses. Briefly, we utilized SOAPnuke (v2.2.6) for initial data filtering, which removed reads containing adapter sequences and low-quality reads where the percentage of bases with quality score <10 exceeded 50% of the read length. These filtering steps produced clean data that were subsequently evaluated for sequencing quality metrics including read count, data volume, and quality score distribution.

To account for potential population stratification and cryptic relatedness that could confound genetic association analyses, we performed comprehensive ancestry and kinship assessment. We extracted high-quality common variants from the whole-exome sequencing data that passed stringent quality control filters: genotype call rate > 95%, a liberal Hardy-Weinberg equilibrium P < 1×10^-6 among unrelated samples, and sequencing depth > 10×. Then we performed linkage disequilibrium (LD)-based pruning using PLINK v1.9. To exclude any individuals related at second degree or closer and potential sample duplications, we estimated pairwise identity-by-descent (IBD) sharing using PLINK's --genome function. Principle components analysis (PCA) was performed by PLINK v1.9 using the high-quality independent autosomal variants. The first 10 principal components were adjusted as covariates across genetic and association analysis.

Variant calling and annotation

Clean sequencing data were aligned to the human reference genome (GRCh38) using Burrows-Wheeler Aligner (BWA v0.7.17) with the BWA-MEM algorithm. The resulting SAM files were processed using Samtools (v1.3.1) to generate sorted BAM format files, which served as the foundation for subsequent variant detection.

Due to PCR amplification during library preparation, duplicate reads were present in the sequencing data and could potentially skew variant detection. These duplicate reads were marked using GATK MarkDuplicates tool (v4.4.0.0) to ensure accurate variant calling. Following duplicate marking, base quality score recalibration (BQSR) was performed using GATK BaseRecalibrator and ApplyBQSR to correct errors in base quality scores. This recalibration utilized known variant sites from dbSNP and 1000 Genomes Project as reference points for quality score adjustment.

Variant calling was conducted using GATK HaplotypeCaller (v4.4.0.0) to accurately identify both SNPs and indels. The raw variant calls were stored in VCF format for subsequent filtering and annotation. To obtain high-confidence variants, we implemented a hard-filtering approach that applied specific filtering parameters separately to SNPs and indels. Variants passing these filters were marked in the final VCF output.

The filtered variants were annotated using a comprehensive annotation approach that integrated information from multiple databases including The Genome Aggregation Database (GnomAD), 1000 Genomes Project database (1000G), NHLBIGO Exome Sequencing Project (ESP6500), Online Mendelian Inheritancein Man (OMIM), The Human Gene Mutation Database (HGMD), Sorting Intolerant from Tolerant (SIFT), Polymorphism Phenotyping v2 (PolyPhen2) and so on. The filtered variants were classified based on predicted functional annotations into missense variants and loss-of-function (LOF) variants (including frameshift insertions/deletions, stop gain/ loss mutations, splice donor/ acceptor, and start loss mutations).

Meanwhile, the deleteriousness score was computationally predicted by these databases based on their effects on protein structure and function. Variants were classified as probably damaging loss-of-function (LOF-D) variants and damaging missense (Mis-D) variants if the variants were predicted by both these two tools as “high damaging” or “moderately damaging”. To offer a more clinical perspective on variants, we further filtered Mis-D variants using the Rare Exome Variant Ensemble Learner (REVEL), applying a stringent threshold of REVEL > 0.773, which corresponds to the top 5% most deleterious mutations according to benchmarking studies.

Rare variant burden analysis

We filtered for rare variants by selecting those with minor allele frequency (MAF) < 1% in large population databases including the 1000G, GnomAD, and ESP6500. This threshold was selected to exclude common polymorphisms while retaining variants that could be specifically enriched in high-altitude cardiovascular disease patients.

To quantify the total load of rare variations in each patient, the exome-wide rare variant burden (ERVB) for each patient was calculated based on the weighted rare variants in each sequenced gene. For each rare variant, we assigned weights based on its predicted functional significance.

Gene enrichment analysis

To gain functional insights into the biological processes potentially impacted by the identified genes, we conducted comprehensive pathway and functional enrichment analyses. The statistical significance of enrichment was assessed using a hypergeometric test with Benjamini-Hochberg correction for multiple testing. Terms with the adjusted p-values after FDR < 0.01, the number of genes from our dataset present in the pathway > 3, and the enrichment factor > 1.5 were considered significantly enriched and grouped into clusters based on their membership similarities.³¹ Gene set was used for pathway enrichment and biological process analysis against the background databases included Gene Ontology (GO) biological processes, Reactome pathways, and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways databases.³²^,³³ DisGeNET database was used to identify the related diseases of the specific gene patterns of clusters.³⁴ PaGenBase databases was used to identify the tissue-specific of clusters.³⁵

Gene interaction network analysis

The network of the 124 related genes was constructed with GeneMANIA.³⁶ Multiple types of gene-gene associations were incorporated in this network integration approach, including co-expression, physical interactions, and shared protein domains indicating potential functional similarity. This analysis identified both direct interactions among input genes and indirect connections through intermediary genes not present in our initial list but functionally related to multiple input genes. The resulting network was visualized using Cytoscape (version 3.8.2).

Lipidomics

Lipid extraction

Lipid extraction was performed following the standardized protocol.³⁷^,³⁸^,³⁹ Briefly, 100 μL of each sample underwent extraction with 300 μL of pre-cooled isopropanol (IPA). They were subjected to vortex for 1 minute, and the extraction mixtures were then maintained at -20°C overnight. Following centrifugation at 4,000g for 30 minutes at 4°C, the supernatant fractions were carefully transferred to fresh 96-well plates and diluted in a 1:10 ratio with an IPA/acetonitrile (ACN)/H₂O solvent system (2:1:1, v:v:v). Additionally, pooled plasma samples were generated as quality control (QC) sample by combining 10 μL from each extraction mixture.

Lipidomics analysis

Lipidomic profiling was performed using a Waters 2777C ultra-performance liquid chromatography (UPLC) (Waters, USA) connected to a Q Exactive HF mass spectrometer (Thermo Fisher Scientific, USA). The samples were randomly ordered, and 10 QC samples were initially injected to condition the column. One QC sample was injected and analyzed every 10 samples to investigate the repeatability of the data. Raw data files were processed using LipidSearch software (v4.1, Thermo Fisher Scientific, USA) for peak detection, alignment, integration, and metabolite annotation.

Lipidomics data analysis

Orthogonal partial least squares discriminant analysis (OPLS-DA) was analyzed using the MetaboAnalystR package in R. Then we implemented comprehensive covariate adjustment strategies across lipidomic differential analysis to avoid the potential confounding effects of baseline clinical and demographic differences. We employed multivariable linear regression models with log₂-transformed metabolite as dependent variables and group membership as the primary independent variable, adjusting for baseline population factors as covariates. Statistical significance was determined based on the adjusted p-values after FDR correction using the Benjamini-Hochberg method. Differential metabolites between IHA and HAM groups were identified based on the following criteria: (1) VIP score >= 1 from the OPLS-DA model, (2) FC > 1.2 or < 0.83, and (3) FDR adjusted p-value (q-value) < 0.05. Fold changes were calculated as the ratio of mean metabolite intensities between two groups, with values > 1 indicating up-regulation and values < 1 indicating down-regulation in IHA patients relative to HAM patients. Differential metabolites in the three-way comparison analysis of high-altitude with low-altitude were identified based on the following criteria: (1) VIP score >= 1 from the OPLS-DA model, (2) FDR adjusted p-value (q-value) < 0.05. Statistical significance was assessed using Student's t-test with Benjamini-Hochberg correction for multiple testing. Identified differential metabolite path set enrichment was analyzed using the MetaboAnalystR package in R.

Correlation analysis was performed across differential lipids. Correlation coefficients (R) and corresponding p-values were calculated for each lipid pair. A stringent filtering criterion was applied (R >= 0.85 and p < 0.05) to identify highly correlated lipid pairs.

Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways database was used to perform pathway enrichment analysis. Fisher's exact test was used to verify the P-value of the calculated pathway enrichment. Benjamini-Hochberg method was employed to control the FDR.

Receiver Operating Characteristic (ROC) curve analysis was performed using the MetaboAnalystR package in R. Support Vector Machine (SVM) algorithms were employed to develop classification models based on lipid profiles. SVM models were trained with 10-fold cross-validation to optimize model parameters and prevent overfitting. Two approaches were implemented: (1) single-lipid models to assess the discriminatory power of individual lipid species and (2) multi-lipid models to evaluate the performance of lipid combinations. The 95% confidence intervals for AUC values were calculated using 100 times bootstrap resampling. Sensitivity, specificity, positive predictive value, and negative predictive value were calculated at the optimal classification threshold. The1000 times permutation tests were used to confirm the robustness.

Association analysis

We assessed the associations between the identified mutations and the metabolic data by employing linear regression. Genetic principal components and baseline population factors were incorporated as covariates in linear regression models. Associations with FDR-corrected p-values < 0.05 were considered statistically significant.

Quantification and statistical analysis

Differences between categorical variables were compared by the Fisher's exact test or chi-square test. Non-parametric variables were evaluated by the Mann‒Whitney test. Hazard ratios (95 % confidence intervals) were calculated by univariate and multivariate Cox regression analyses. A p-value < 0.05 was considered as statistically significant.

Published: February 20, 2026

Footnotes

Supplemental information can be found online at https://doi.org/10.1016/j.isci.2026.115109.

Contributor Information

Song Cui, Email: cuisongdoctor@163.com.

Xiantao Song, Email: song0929@mail.ccmu.edu.cn.

Kefeng Li, Email: kefengl@mpu.edu.mo.

Supplemental information

Document S1. Figures S1 and S2 and Tables S1–S13

mmc1.pdf^{(996.6KB, pdf)}

References

1.He Y., Zheng W., Guo Y., Yue T., Cui C., Wu T., Zhang H., Zhang H., Liu K., Yang Z., et al. Deep phenotyping of 11,880 highlanders reveals novel adaptive traits in native Tibetans. iScience. 2023;26 doi: 10.1016/j.isci.2023.107677. [DOI] [PMC free article] [PubMed] [Google Scholar]
2.Burtscher M. Effects of living at higher altitudes on mortality: a narrative review. Aging Dis. 2014;5:274–280. doi: 10.14336/ad.2014.0500274. [DOI] [PMC free article] [PubMed] [Google Scholar]
3.Riggs D.W., Yeager R.A., Bhatnagar A. Defining the Human Envirome: An Omics Approach for Assessing the Environmental Risk of Cardiovascular Disease. Circ. Res. 2018;122:1259–1275. doi: 10.1161/circresaha.117.311230. [DOI] [PMC free article] [PubMed] [Google Scholar]
4.Bai J., Li L., Li Y., Zhang L. Genetic and immune changes in Tibetan high-altitude populations contribute to biological adaptation to hypoxia. Environ. Health Prev. Med. 2022;27:39. doi: 10.1265/ehpm.22-00040. [DOI] [PMC free article] [PubMed] [Google Scholar]
5.Storz J.F. High-Altitude Adaptation: Mechanistic Insights from Integrated Genomics and Physiology. Mol. Biol. Evol. 2021;38:2677–2691. doi: 10.1093/molbev/msab064. [DOI] [PMC free article] [PubMed] [Google Scholar]
6.Yan S.Y., Ma L.H., Yang W.X. Altitude and prognosis after PCI: A propensity score-matched analysis. Heliyon. 2024;10 doi: 10.1016/j.heliyon.2024.e33577. [DOI] [PMC free article] [PubMed] [Google Scholar]
7.Richalet J.P., Hermand E., Lhuissier F.J. Cardiovascular physiology and pathophysiology at high altitude. Nat. Rev. Cardiol. 2024;21:75–88. doi: 10.1038/s41569-023-00924-9. [DOI] [PubMed] [Google Scholar]
8.Mallet R.T., Burtscher J., Richalet J.P., Millet G.P., Burtscher M. Impact of High Altitude on Cardiovascular Health: Current Perspectives. Vasc. Health Risk Manag. 2021;17:317–335. doi: 10.2147/vhrm.S294121. [DOI] [PMC free article] [PubMed] [Google Scholar]
9.Soppert J., Lehrke M., Marx N., Jankowski J., Noels H. Lipoproteins and lipids in cardiovascular disease: from mechanistic insights to therapeutic targeting. Adv. Drug Deliv. Rev. 2020;159:4–33. doi: 10.1016/j.addr.2020.07.019. [DOI] [PubMed] [Google Scholar]
10.Midha A.D., Zhou Y., Queliconi B.B., Barrios A.M., Haribowo A.G., Chew B.T.L., Fong C.O.Y., Blecha J.E., VanBrocklin H., Seo Y., Jain I.H. Organ-specific fuel rewiring in acute and chronic hypoxia redistributes glucose and fatty acid metabolism. Cell Metab. 2023;35:504–516.e5. doi: 10.1016/j.cmet.2023.02.007. [DOI] [PMC free article] [PubMed] [Google Scholar]
11.Meng W., Pan H., Sha Y., Zhai X., Xing A., Lingampelly S.S., Sripathi S.R., Wang Y., Li K. Metabolic Connectome and Its Role in the Prediction, Diagnosis, and Treatment of Complex Diseases. Metabolites. 2024;14 doi: 10.3390/metabo14020093. [DOI] [PMC free article] [PubMed] [Google Scholar]
12.Pan H., Sha Y., Zhai X., Luo G., Xu W., Meng W., Li K. Bootstrap inference and machine learning reveal core differential plasma metabolic connectome signatures in major depressive disorder. J. Affect. Disord. 2025;378:281–292. doi: 10.1016/j.jad.2025.02.109. [DOI] [PubMed] [Google Scholar]
13.Gao J., Zhao M., Cheng X., Yue X., Hao F., Wang H., Duan L., Han C., Zhu L. Metabolomic analysis of human plasma sample after exposed to high altitude and return to sea level. PLoS One. 2023;18 doi: 10.1371/journal.pone.0282301. [DOI] [PMC free article] [PubMed] [Google Scholar]
14.Knutson A.K., Williams A.L., Boisvert W.A., Shohet R.V. HIF in the heart: development, metabolism, ischemia, and atherosclerosis. J. Clin. Investig. 2021;131 doi: 10.1172/jci137557. [DOI] [PMC free article] [PubMed] [Google Scholar]
15.Yang N., Sun R., Liao X., Aa J., Wang G. UDP-glucuronosyltransferases (UGTs) and their related metabolic cross-talk with internal homeostasis: A systematic review of UGT isoforms for precision medicine. Pharmacol. Res. 2017;121:169–183. doi: 10.1016/j.phrs.2017.05.001. [DOI] [PubMed] [Google Scholar]
16.Jarrar Y., Lee S.J. The Functionality of UDP-Glucuronosyltransferase Genetic Variants and their Association with Drug Responses and Human Diseases. J. Personalized Med. 2021;11 doi: 10.3390/jpm11060554. [DOI] [PMC free article] [PubMed] [Google Scholar]
17.Vitek L., Hinds T.D., Jr., Stec D.E., Tiribelli C. The physiology of bilirubin: health and disease equilibrium. Trends Mol. Med. 2023;29:315–328. doi: 10.1016/j.molmed.2023.01.007. [DOI] [PMC free article] [PubMed] [Google Scholar]
18.Suh S., Cho Y.R., Park M.K., Kim D.K., Cho N.H., Lee M.K. Relationship between serum bilirubin levels and cardiovascular disease. PLoS One. 2018;13 doi: 10.1371/journal.pone.0193041. [DOI] [PMC free article] [PubMed] [Google Scholar]
19.Jain V., Ghosh R.K., Bandyopadhyay D., Kondapaneni M., Mondal S., Hajra A., Aronow W.S., Lavie C.J. Serum Bilirubin and Coronary Artery Disease: Intricate Relationship, Pathophysiology, and Recent Evidence. Curr. Probl. Cardiol. 2021;46 doi: 10.1016/j.cpcardiol.2019.06.003. [DOI] [PubMed] [Google Scholar]
20.Zietzer A., Düsing P., Reese L., Nickenig G., Jansen F. Ceramide Metabolism in Cardiovascular Disease: A Network With High Therapeutic Potential. Arterioscler. Thromb. Vasc. Biol. 2022;42:1220–1228. doi: 10.1161/atvbaha.122.318048. [DOI] [PubMed] [Google Scholar]
21.Xia Q.S., Lu F.E., Wu F., Huang Z.Y., Dong H., Xu L.J., Gong J. New role for ceramide in hypoxia and insulin resistance. World J. Gastroenterol. 2020;26:2177–2186. doi: 10.3748/wjg.v26.i18.2177. [DOI] [PMC free article] [PubMed] [Google Scholar]
22.Piccoli M., Cirillo F., Ghiroldi A., Rota P., Coviello S., Tarantino A., La Rocca P., Lavota I., Creo P., Signorelli P., et al. Sphingolipids and Atherosclerosis: The Dual Role of Ceramide and Sphingosine-1-Phosphate. Antioxidants. 2023;12 doi: 10.3390/antiox12010143. [DOI] [PMC free article] [PubMed] [Google Scholar]
23.Foran D., Antoniades C., Akoumianakis I. Emerging Roles for Sphingolipids in Cardiometabolic Disease: A Rational Therapeutic Target? Nutrients. 2024;16 doi: 10.3390/nu16193296. [DOI] [PMC free article] [PubMed] [Google Scholar]
24.Bhargava S., de la Puente-Secades S., Schurgers L., Jankowski J. Lipids and lipoproteins in cardiovascular diseases: a classification. Trends Endocrinol. Metabol. 2022;33:409–423. doi: 10.1016/j.tem.2022.02.001. [DOI] [PubMed] [Google Scholar]
25.Goodpaster B.H., Sparks L.M. Metabolic Flexibility in Health and Disease. Cell Metab. 2017;25:1027–1036. doi: 10.1016/j.cmet.2017.04.015. [DOI] [PMC free article] [PubMed] [Google Scholar]
26.Laaksonen R., Ekroos K., Sysi-Aho M., Hilvo M., Vihervaara T., Kauhanen D., Suoniemi M., Hurme R., März W., Scharnagl H., et al. Plasma ceramides predict cardiovascular death in patients with stable coronary artery disease and acute coronary syndromes beyond LDL-cholesterol. Eur. Heart J. 2016;37:1967–1976. doi: 10.1093/eurheartj/ehw148. [DOI] [PMC free article] [PubMed] [Google Scholar]
27.Hernández-Corbacho M.J., Salama M.F., Canals D., Senkal C.E., Obeid L.M. Sphingolipids in mitochondria. Biochim. Biophys. Acta Mol. Cell Biol. Lipids. 2017;1862:56–68. doi: 10.1016/j.bbalip.2016.09.019. [DOI] [PMC free article] [PubMed] [Google Scholar]
28.Choi S., Snider A.J. Sphingolipids in High Fat Diet and Obesity-Related Diseases. Mediat. Inflamm. 2015;2015 doi: 10.1155/2015/520618. [DOI] [PMC free article] [PubMed] [Google Scholar]
29.Chaurasia B., Summers S.A. Ceramides in Metabolism: Key Lipotoxic Players. Annu. Rev. Physiol. 2021;83:303–330. doi: 10.1146/annurev-physiol-031620-093815. [DOI] [PMC free article] [PubMed] [Google Scholar]
30.Sud M., Fahy E., Cotter D., Azam K., Vadivelu I., Burant C., Edison A., Fiehn O., Higashi R., Nair K.S., et al. Metabolomics Workbench: An international repository for metabolomics data and metadata, metabolite standards, protocols, tutorials and training, and analysis tools. Nucleic Acids Res. 2016;44:D463–D470. doi: 10.1093/nar/gkv1042. [DOI] [PMC free article] [PubMed] [Google Scholar]
31.Zhou Y., Zhou B., Pache L., Chang M., Khodabakhshi A.H., Tanaseichuk O., Benner C., Chanda S.K. Metascape provides a biologist-oriented resource for the analysis of systems-level datasets. Nat. Commun. 2019;10:1523. doi: 10.1038/s41467-019-09234-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
32.Gene Ontology Consortium, Aleksander S.A., Balhoff J., Carbon S., Cherry J.M., Drabkin H.J., Ebert D., Feuermann M., Gaudet P., Harris N.L., et al. The Gene Ontology knowledgebase in 2023. Genetics. 2023;224 doi: 10.1093/genetics/iyad031. [DOI] [PMC free article] [PubMed] [Google Scholar]
33.Kanehisa M., Furumichi M., Sato Y., Matsuura Y., Ishiguro-Watanabe M. KEGG: biological systems database as a model of the real world. Nucleic Acids Res. 2025;53:D672–D677. doi: 10.1093/nar/gkae909. [DOI] [PMC free article] [PubMed] [Google Scholar]
34.Piñero J., Ramírez-Anguita J.M., Saüch-Pitarch J., Ronzano F., Centeno E., Sanz F., Furlong L.I. The DisGeNET knowledge platform for disease genomics: 2019 update. Nucleic Acids Res. 2020;48:D845–D855. doi: 10.1093/nar/gkz1021. [DOI] [PMC free article] [PubMed] [Google Scholar]
35.Pan J.B., Hu S.C., Shi D., Cai M.C., Li Y.B., Zou Q., Ji Z.L. PaGenBase: a pattern gene database for the global and dynamic understanding of gene function. PLoS One. 2013;8 doi: 10.1371/journal.pone.0080747. [DOI] [PMC free article] [PubMed] [Google Scholar]
36.Warde-Farley D., Donaldson S.L., Comes O., Zuberi K., Badrawi R., Chao P., Franz M., Grouios C., Kazi F., Lopes C.T., et al. The GeneMANIA prediction server: biological network integration for gene prioritization and predicting gene function. Nucleic Acids Res. 2010;38:W214–W220. doi: 10.1093/nar/gkq537. [DOI] [PMC free article] [PubMed] [Google Scholar]
37.Sarafian M.H., Gaudin M., Lewis M.R., Martin F.P., Holmes E., Nicholson J.K., Dumas M.E. Objective set of criteria for optimization of sample preparation procedures for ultra-high throughput untargeted blood plasma lipid profiling by ultra performance liquid chromatography-mass spectrometry. Anal. Chem. 2014;86:5766–5774. doi: 10.1021/ac500317c. [DOI] [PubMed] [Google Scholar]
38.Zhong H., Fang C., Fan Y., Lu Y., Wen B., Ren H., Hou G., Yang F., Xie H., Jie Z., et al. Lipidomic profiling reveals distinct differences in plasma lipid composition in healthy, prediabetic, and type 2 diabetic individuals. GigaScience. 2017;6:1–12. doi: 10.1093/gigascience/gix036. [DOI] [PMC free article] [PubMed] [Google Scholar]
39.Barranco-Altirriba M., Alonso N., Weber R.J.M., Lloyd G.R., Hernandez M., Yanes O., Capellades J., Jankevics A., Winder C., Falguera M., et al. Lipidome characterisation and sex-specific differences in type 1 and type 2 diabetes mellitus. Cardiovasc. Diabetol. 2024;23:109. doi: 10.1186/s12933-024-02202-5. [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

Document S1. Figures S1 and S2 and Tables S1–S13

mmc1.pdf^{(996.6KB, pdf)}

Data Availability Statement

•
Lipidomics data have been deposited at Metabolomics Workbench (https://www.metabolomicsworkbench.org) as study ID: ST004571, and are publicly available as of the date of publication. The WES data reported in this study cannot be deposited in a public repository, because local law prohibits depositing raw genomic data in public repositories. In addition, summary statistics describing these data are accessible in Supplementary tables.
•
This article does not report original codes.
•
Any additional information required to reanalyze the data reported in this article is available from the lead contact upon request.

[bib1] 1.He Y., Zheng W., Guo Y., Yue T., Cui C., Wu T., Zhang H., Zhang H., Liu K., Yang Z., et al. Deep phenotyping of 11,880 highlanders reveals novel adaptive traits in native Tibetans. iScience. 2023;26 doi: 10.1016/j.isci.2023.107677. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib2] 2.Burtscher M. Effects of living at higher altitudes on mortality: a narrative review. Aging Dis. 2014;5:274–280. doi: 10.14336/ad.2014.0500274. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib3] 3.Riggs D.W., Yeager R.A., Bhatnagar A. Defining the Human Envirome: An Omics Approach for Assessing the Environmental Risk of Cardiovascular Disease. Circ. Res. 2018;122:1259–1275. doi: 10.1161/circresaha.117.311230. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib4] 4.Bai J., Li L., Li Y., Zhang L. Genetic and immune changes in Tibetan high-altitude populations contribute to biological adaptation to hypoxia. Environ. Health Prev. Med. 2022;27:39. doi: 10.1265/ehpm.22-00040. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib5] 5.Storz J.F. High-Altitude Adaptation: Mechanistic Insights from Integrated Genomics and Physiology. Mol. Biol. Evol. 2021;38:2677–2691. doi: 10.1093/molbev/msab064. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib6] 6.Yan S.Y., Ma L.H., Yang W.X. Altitude and prognosis after PCI: A propensity score-matched analysis. Heliyon. 2024;10 doi: 10.1016/j.heliyon.2024.e33577. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib7] 7.Richalet J.P., Hermand E., Lhuissier F.J. Cardiovascular physiology and pathophysiology at high altitude. Nat. Rev. Cardiol. 2024;21:75–88. doi: 10.1038/s41569-023-00924-9. [DOI] [PubMed] [Google Scholar]

[bib8] 8.Mallet R.T., Burtscher J., Richalet J.P., Millet G.P., Burtscher M. Impact of High Altitude on Cardiovascular Health: Current Perspectives. Vasc. Health Risk Manag. 2021;17:317–335. doi: 10.2147/vhrm.S294121. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib9] 9.Soppert J., Lehrke M., Marx N., Jankowski J., Noels H. Lipoproteins and lipids in cardiovascular disease: from mechanistic insights to therapeutic targeting. Adv. Drug Deliv. Rev. 2020;159:4–33. doi: 10.1016/j.addr.2020.07.019. [DOI] [PubMed] [Google Scholar]

[bib10] 10.Midha A.D., Zhou Y., Queliconi B.B., Barrios A.M., Haribowo A.G., Chew B.T.L., Fong C.O.Y., Blecha J.E., VanBrocklin H., Seo Y., Jain I.H. Organ-specific fuel rewiring in acute and chronic hypoxia redistributes glucose and fatty acid metabolism. Cell Metab. 2023;35:504–516.e5. doi: 10.1016/j.cmet.2023.02.007. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib11] 11.Meng W., Pan H., Sha Y., Zhai X., Xing A., Lingampelly S.S., Sripathi S.R., Wang Y., Li K. Metabolic Connectome and Its Role in the Prediction, Diagnosis, and Treatment of Complex Diseases. Metabolites. 2024;14 doi: 10.3390/metabo14020093. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib12] 12.Pan H., Sha Y., Zhai X., Luo G., Xu W., Meng W., Li K. Bootstrap inference and machine learning reveal core differential plasma metabolic connectome signatures in major depressive disorder. J. Affect. Disord. 2025;378:281–292. doi: 10.1016/j.jad.2025.02.109. [DOI] [PubMed] [Google Scholar]

[bib13] 13.Gao J., Zhao M., Cheng X., Yue X., Hao F., Wang H., Duan L., Han C., Zhu L. Metabolomic analysis of human plasma sample after exposed to high altitude and return to sea level. PLoS One. 2023;18 doi: 10.1371/journal.pone.0282301. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib14] 14.Knutson A.K., Williams A.L., Boisvert W.A., Shohet R.V. HIF in the heart: development, metabolism, ischemia, and atherosclerosis. J. Clin. Investig. 2021;131 doi: 10.1172/jci137557. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib15] 15.Yang N., Sun R., Liao X., Aa J., Wang G. UDP-glucuronosyltransferases (UGTs) and their related metabolic cross-talk with internal homeostasis: A systematic review of UGT isoforms for precision medicine. Pharmacol. Res. 2017;121:169–183. doi: 10.1016/j.phrs.2017.05.001. [DOI] [PubMed] [Google Scholar]

[bib16] 16.Jarrar Y., Lee S.J. The Functionality of UDP-Glucuronosyltransferase Genetic Variants and their Association with Drug Responses and Human Diseases. J. Personalized Med. 2021;11 doi: 10.3390/jpm11060554. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib17] 17.Vitek L., Hinds T.D., Jr., Stec D.E., Tiribelli C. The physiology of bilirubin: health and disease equilibrium. Trends Mol. Med. 2023;29:315–328. doi: 10.1016/j.molmed.2023.01.007. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib18] 18.Suh S., Cho Y.R., Park M.K., Kim D.K., Cho N.H., Lee M.K. Relationship between serum bilirubin levels and cardiovascular disease. PLoS One. 2018;13 doi: 10.1371/journal.pone.0193041. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib19] 19.Jain V., Ghosh R.K., Bandyopadhyay D., Kondapaneni M., Mondal S., Hajra A., Aronow W.S., Lavie C.J. Serum Bilirubin and Coronary Artery Disease: Intricate Relationship, Pathophysiology, and Recent Evidence. Curr. Probl. Cardiol. 2021;46 doi: 10.1016/j.cpcardiol.2019.06.003. [DOI] [PubMed] [Google Scholar]

[bib20] 20.Zietzer A., Düsing P., Reese L., Nickenig G., Jansen F. Ceramide Metabolism in Cardiovascular Disease: A Network With High Therapeutic Potential. Arterioscler. Thromb. Vasc. Biol. 2022;42:1220–1228. doi: 10.1161/atvbaha.122.318048. [DOI] [PubMed] [Google Scholar]

[bib21] 21.Xia Q.S., Lu F.E., Wu F., Huang Z.Y., Dong H., Xu L.J., Gong J. New role for ceramide in hypoxia and insulin resistance. World J. Gastroenterol. 2020;26:2177–2186. doi: 10.3748/wjg.v26.i18.2177. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib22] 22.Piccoli M., Cirillo F., Ghiroldi A., Rota P., Coviello S., Tarantino A., La Rocca P., Lavota I., Creo P., Signorelli P., et al. Sphingolipids and Atherosclerosis: The Dual Role of Ceramide and Sphingosine-1-Phosphate. Antioxidants. 2023;12 doi: 10.3390/antiox12010143. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib23] 23.Foran D., Antoniades C., Akoumianakis I. Emerging Roles for Sphingolipids in Cardiometabolic Disease: A Rational Therapeutic Target? Nutrients. 2024;16 doi: 10.3390/nu16193296. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib24] 24.Bhargava S., de la Puente-Secades S., Schurgers L., Jankowski J. Lipids and lipoproteins in cardiovascular diseases: a classification. Trends Endocrinol. Metabol. 2022;33:409–423. doi: 10.1016/j.tem.2022.02.001. [DOI] [PubMed] [Google Scholar]

[bib25] 25.Goodpaster B.H., Sparks L.M. Metabolic Flexibility in Health and Disease. Cell Metab. 2017;25:1027–1036. doi: 10.1016/j.cmet.2017.04.015. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib26] 26.Laaksonen R., Ekroos K., Sysi-Aho M., Hilvo M., Vihervaara T., Kauhanen D., Suoniemi M., Hurme R., März W., Scharnagl H., et al. Plasma ceramides predict cardiovascular death in patients with stable coronary artery disease and acute coronary syndromes beyond LDL-cholesterol. Eur. Heart J. 2016;37:1967–1976. doi: 10.1093/eurheartj/ehw148. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib27] 27.Hernández-Corbacho M.J., Salama M.F., Canals D., Senkal C.E., Obeid L.M. Sphingolipids in mitochondria. Biochim. Biophys. Acta Mol. Cell Biol. Lipids. 2017;1862:56–68. doi: 10.1016/j.bbalip.2016.09.019. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib28] 28.Choi S., Snider A.J. Sphingolipids in High Fat Diet and Obesity-Related Diseases. Mediat. Inflamm. 2015;2015 doi: 10.1155/2015/520618. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib29] 29.Chaurasia B., Summers S.A. Ceramides in Metabolism: Key Lipotoxic Players. Annu. Rev. Physiol. 2021;83:303–330. doi: 10.1146/annurev-physiol-031620-093815. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib30] 30.Sud M., Fahy E., Cotter D., Azam K., Vadivelu I., Burant C., Edison A., Fiehn O., Higashi R., Nair K.S., et al. Metabolomics Workbench: An international repository for metabolomics data and metadata, metabolite standards, protocols, tutorials and training, and analysis tools. Nucleic Acids Res. 2016;44:D463–D470. doi: 10.1093/nar/gkv1042. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib31] 31.Zhou Y., Zhou B., Pache L., Chang M., Khodabakhshi A.H., Tanaseichuk O., Benner C., Chanda S.K. Metascape provides a biologist-oriented resource for the analysis of systems-level datasets. Nat. Commun. 2019;10:1523. doi: 10.1038/s41467-019-09234-6. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib32] 32.Gene Ontology Consortium, Aleksander S.A., Balhoff J., Carbon S., Cherry J.M., Drabkin H.J., Ebert D., Feuermann M., Gaudet P., Harris N.L., et al. The Gene Ontology knowledgebase in 2023. Genetics. 2023;224 doi: 10.1093/genetics/iyad031. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib33] 33.Kanehisa M., Furumichi M., Sato Y., Matsuura Y., Ishiguro-Watanabe M. KEGG: biological systems database as a model of the real world. Nucleic Acids Res. 2025;53:D672–D677. doi: 10.1093/nar/gkae909. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib34] 34.Piñero J., Ramírez-Anguita J.M., Saüch-Pitarch J., Ronzano F., Centeno E., Sanz F., Furlong L.I. The DisGeNET knowledge platform for disease genomics: 2019 update. Nucleic Acids Res. 2020;48:D845–D855. doi: 10.1093/nar/gkz1021. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib35] 35.Pan J.B., Hu S.C., Shi D., Cai M.C., Li Y.B., Zou Q., Ji Z.L. PaGenBase: a pattern gene database for the global and dynamic understanding of gene function. PLoS One. 2013;8 doi: 10.1371/journal.pone.0080747. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib36] 36.Warde-Farley D., Donaldson S.L., Comes O., Zuberi K., Badrawi R., Chao P., Franz M., Grouios C., Kazi F., Lopes C.T., et al. The GeneMANIA prediction server: biological network integration for gene prioritization and predicting gene function. Nucleic Acids Res. 2010;38:W214–W220. doi: 10.1093/nar/gkq537. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib37] 37.Sarafian M.H., Gaudin M., Lewis M.R., Martin F.P., Holmes E., Nicholson J.K., Dumas M.E. Objective set of criteria for optimization of sample preparation procedures for ultra-high throughput untargeted blood plasma lipid profiling by ultra performance liquid chromatography-mass spectrometry. Anal. Chem. 2014;86:5766–5774. doi: 10.1021/ac500317c. [DOI] [PubMed] [Google Scholar]

[bib38] 38.Zhong H., Fang C., Fan Y., Lu Y., Wen B., Ren H., Hou G., Yang F., Xie H., Jie Z., et al. Lipidomic profiling reveals distinct differences in plasma lipid composition in healthy, prediabetic, and type 2 diabetic individuals. GigaScience. 2017;6:1–12. doi: 10.1093/gigascience/gix036. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib39] 39.Barranco-Altirriba M., Alonso N., Weber R.J.M., Lloyd G.R., Hernandez M., Yanes O., Capellades J., Jankevics A., Winder C., Falguera M., et al. Lipidome characterisation and sex-specific differences in type 1 and type 2 diabetes mellitus. Cardiovasc. Diabetol. 2024;23:109. doi: 10.1186/s12933-024-02202-5. [DOI] [PMC free article] [PubMed] [Google Scholar]

PERMALINK

Integrated multi-omics reveals population-specific mechanisms of cardiovascular disease in high-altitude environments

Weiyu Meng

Jiajun Yang

Xi Wang

Xiaobing Zhai

Gang Luo

Yuyang Sha

Hongxin Pan

Wei Xu

Henry HY Tong

Edmundo Patricio Lopes Lao

Ci-ren Zhong-ga

Song Cui

Xiantao Song

Kefeng Li

Summary

Graphical abstract

Highlights

Introduction

Results

Overview of study population

Table 1.

Identification of high-frequency rare variants in HACVD patients

Figure 1.

Comparative analysis of indigenous high-altitude residents and high-altitude adapted migrants

Differential genetic profiles between IHA and HAM cardiovascular disease patients

Figure 2.

Lipidomic profiles reveal distinct metabolic adaptations between population groups

In-depth analysis of differential lipids reveals population-specific metabolic signatures and metabolic remodeling

Figure 3.

Integration of genetic variant and lipid profiles reveals population-specific signatures and adaptation mechanisms associated with cardiovascular disease in high-altitude environments

Comparative analysis of high-altitude and low-altitude cardiovascular disease patients

Figure 4.

Discussion

Limitations of the study

Resource availability

Lead contact

Materials availability

Data and code availability

Acknowledgments

Author contributions

Declaration of interests

STAR★Methods

Key resources table

Experimental model and study participant details

Study population and sample collection

Method details

Whole-exome sequencing

Whole-exome sequencing and quality control

Variant calling and annotation

Rare variant burden analysis

Gene enrichment analysis

Gene interaction network analysis

Lipidomics

Lipid extraction

Lipidomics analysis

Lipidomics data analysis

Association analysis

Quantification and statistical analysis

Footnotes

Contributor Information

Supplemental information

References

Associated Data

Supplementary Materials

Data Availability Statement

ACTIONS

PERMALINK

RESOURCES

Similar articles

Cited by other articles

Links to NCBI Databases