Exploring the causal relationship between antihypertensive drugs and glioblastoma by combining drug target Mendelian randomization study, eQTL colocalization, and single‐cell RNA sequencing

Recent reports indicate a potential oncogenic role of antihypertensive drugs in common cancers. However, it remains uncertain whether this phenomenon influences the risk of glioblastoma multiforme (GBM). This study aimed to assess the potential causal effects of blood pressure (BP) and antihypertensive drugs on GBM. Genome‐wide association study (GWAS) summary statistics for systolic blood pressure (SBP), diastolic blood pressure (DBP), and GBM in Europeans were downloaded. To represent the effects of antihypertensive drugs, we utilized single nucleotide polymorphisms (SNPs) associated with SBP/DBP adjacent to the coding regions of different antihypertensive drugs as instrumental variables to model five antihypertensive drugs, including angiotensin‐converting enzyme inhibitors, angiotensin receptor blockers, calcium channel blockers, β‐receptor blockers (BBs), and thiazide diuretics. Positive control studies were performed using GWAS data in chronic heart failure. The primary method for causality estimation was the inverse‐variance‐weighted method. Mendelian randomization analysis showed that BBs with the β1‐adrenergic receptor (ADRB1) as a therapeutic target could significantly reduce the risk of GBM by mediating DBP (OR = 0.431, 95% CI: 0.267–0.697, p < .001) and that they could also significantly reduce the risk of GBM by mediating SBP (OR = 0.595, 95% CI: 0.422–0.837, p = .003). Sensitivity analysis and colocalization analysis reinforced the robustness of these findings. Finally, the low expression of the ADRB1 gene in malignant gliomas was found by GBM data from TCGA and single‐cell RNA sequencing, which most likely contributed to the poor prognosis of GBM patients. In summary, our study provides preliminary evidence of some causal relationship between ADRB1‐targeted BBs and glioblastoma development. However, more studies are needed to validate these findings and further reveal the complex relationship between BP and GBM.

which most likely contributed to the poor prognosis of GBM patients.In summary, our study provides preliminary evidence of some causal relationship between ADRB1-targeted BBs and glioblastoma development.However, more studies are needed to validate these findings and further reveal the complex relationship between BP and GBM.

| INTRODUCTION
Glioblastoma multiforme (GBM) stands as the most prevalent and deadly primary malignant tumor within the adult central nervous system, constituting around 50% of all gliomas. 1,2The complex microenvironment of tumor cells and intratumor heterogeneity pose great challenges for clinical treatment.Evolving treatments, including surgical resection, temozolomide chemotherapy, radiotherapy, and immunotherapy, can only maintain a median survival of approximately 15 months in patients with GBM. 3 Hence, there is an immediate requirement to discover novel therapeutic approaches and anti-tumor agents to address this challenge.
Hypertension is an important public health problem worldwide and is strongly associated with a high burden of cardiovascular disease (CVD) and premature death.According to the Global Burden of Disease Study, elevated blood pressure (BP) is the leading cause of death and disability-adjusted life years (Daly) worldwide, with 10.4 million deaths and 218 million Daly attributable to elevated systolic BP in 2017. 4Despite the widespread use of proven methods of lowering BP, namely lifestyle changes and medication, many patients still do not achieve guideline-recommended BP control. 5A brand new study shows that the same risk factors can lead to CVD in one person, cancer in another, or even two diseases in the same person. 6However, we know little about the relationship between hypertensive disease and cancer.In a large European cohort study, researchers followed subjects for 12 years and came to the startling conclusion that for every 10 mm Hg increase in BP, the incidence of cancer in men increased. 7Edlinger et al. 8 published a cohort study evaluating the relationship between metabolic syndrome and brain tumor risk.The study showed that elevated BP was associated with an increased risk of brain tumors, particularly meningiomas, while diastolic blood pressure (DBP) and triglycerides were associated with an increased risk of high-grade gliomas. 9It is important to note that there is information about the relationship between BP and cancer is still complex and the findings are not always consistent.Therefore, more research is needed to gain a deeper understanding of the relationship between hypertension and cancer and to prevent brain tumors by actively controlling BP and maintaining a healthy lifestyle.
To verify causality and to examine the impact of antihypertensive drugs on cancer treatment, a large meta-analysis of patients with a variety of tumors showed that renin-angiotensin-aldosterone system inhibitors were beneficial for all cancer endpoints including overall survival (OS), progression-free survival (PFS), and disease-free survival. 10Interestingly, one study showed that the use of BBs inhibited pancreatic cancer invasion and proliferation. 11A recent large study based on a UK population of 850 000 has demonstrated that the use of calcium channel blockers (CCBs) is not associated with an increased risk of cancer, regardless of the duration of drug use and type of cancer. 12sufficient scientific evidence exists to establish a direct association between gliomas, especially glioblastomas, and antihypertensive drugs.Recent systematic reviews and meta-analyses highlight the complexity of obtaining reliable evidence on this matter.Traditional pharmacoepidemiologic studies face challenges, including eternal time bias, selection bias, confounding, memory bias, and measurement error, impacting the accuracy and credibility of findings.These limitations underscore the need for cautious consideration in pharmacoepidemiologic research and advocate for more precise and comprehensive study designs and methods to accurately discern the relationship between medications and health outcomes.
Mendelian randomization (MR) is a method that leverages the natural random assignment relationships between genotypes and phenotypes, akin to randomization in experimental treatment and control groups. 13Rooted in Mendelian genetics, MR offers strengths in confounding control and causal inference.However, limitations include challenges in selecting genetic variants, the potential nonlinearity of exposure-outcome associations, and the existence of variants affecting multiple biological pathways. 14The efficacy failure and adverse effects in new drug development, particularly in randomized clinical trials, often result from insufficient genetic support.Drug targets MR, employing genetic instrumentation near or within the target gene, mimics potential drug effects, aiding predictions in drug development, and repurposing. 15This study utilized recently published genome-wide association study (GWAS) statistics to explore the causal relationship between five antihypertensive drugs and glioblastoma.The findings offer valuable insights for guiding antihypertensive medication use and informing clinicians in designing future glioma prevention strategies.

| Selection of genetic instruments
This study aimed to assess the causal association between BP (systolic blood pressure [SBP] and DBP) and glioblastoma mediated by five antihypertensive drugs (angiotensin-converting enzyme inhibitors [ACEIs], angiotensin receptor blockers [ARBs], β-receptor blockers [BBs], CCBs, and thiazide diuretics [TDs]).The GWAS summary-level statistics for systolic and DBP were sourced from a recent extensive GWAS meta-analysis involving 757 601 Europeans. 16The study populations were drawn from two major cohorts, namely the UK Biobank and the International Consortium for BP, and these GWAS summary statistics are accessible through the IEU Open GWAS Project database (ieu-b-38/ieu-b-39). 17 To simulate the effects of five antihypertensive drugs, instrumental variables (IVs) targeting gene loci associated with BP reduction were utilized.The selection of IVs for different antihypertensive drugs followed these steps: first, the Durg-Bank online database provided 20 target genes for the five drugs-ACEIs, ARBs, BBs, CCBs, and TDs (Table S1).Subsequently, the chromosomal locations of these target genes were obtained from the NCBI Gene database.Finally, a tool was introduced to proxy exposure to antihypertensive drugs by selecting SNPs associated with SBP/DBP at a genome-wide significance level ( p < 5 Â 10 À8 ) within a 100 kb window of the target genes for each drug.These BPassociated SNPs exclusively included common SNPs (minor allele frequency >1%).We allowed the SNPs used as IVs to be in low-weak linkage disequilibrium with each other (r2 < 0.3) to obtain maximum validity of the instrument.

GWAS summary statistics for glioblastoma (excluding all cancers)
were extracted from the FinnGen consortium.GWAS summary statistics for GBM included 91 cases and 174 006 controls.To validate the reliability of the selected target genes and IVs, a positive control study was conducted.Since antihypertensive drugs are mainly used to treat and prevent CVDs such as hypertension, we used chronic heart failure (HF) as a positive control outcome.Similarly, GWAS summary statistics for HF were obtained from the IEU Open GWAS Project database (ebi-a-GCST90018806), which includes 14 262 cases and 471 898 controls. 18

| GWAS-eQTL-integrated colocalization analysis
Initially, we acquired expression quantitative trait loci (eQTLs) data for the target genes from the eQTLGen Consortium (https://www.eqtlgen.org/)and subsequently conducted GWAS-eQTL colocalization using a Bayesian approach. 19The method assesses whether GWAS and eQTL associations best fit a model in which the association is caused by a single shared variable (summarized by posterior probabilities).There are five hypotheses for the colocalization analysis, namely: H0, SNPs within the colocalized region are not related to either trait.H1, SNPs within the colocalized region are related to the first trait but not to the second.H2, SNPs within the colocalized region are related to the second trait but not to the first.H3, SNPs within the colocalized region are related to both traits but not to the same locus.H4, SNPs within the colocalized region are related to both traits and are at the same locus.H4 (PP.H4) has a posteriori probability of at least 50%, suggesting that colocalized domains are possible. 20

| Data analysis
First, we coordinated exposure-related drug-targeted IVs with the outcome dataset, and then performed traditional MR analysis using MR Egger, weighted median, inverse variance weighted (IVW), simple mode, and weighted mode, of which the IVW method was the most dominant.
The IVW method was used to determine the presence of heterogeneity using the Cochran Q test, with p > .05indicating the absence of heterogeneity.We used MR-Egger regression to assess the potential horizontal pleiotropy of SNPs as IVs.An overall p < .05MR-Egger analysis indicated the presence of horizontal pleiotropy.The assumptions of MR require that SNPs are not directly related to the outcome.Therefore, the online website PhenoScanner (http://www.phenoscanner.medschl.cam.ac.uk//) was used for the exclusion of SNPs associated with tumors.Sensitivity analyses were performed again after the removal of outliers by the MR-PRESSO test.Finally, a leave-one-out sensitivity analysis was performed, and MR results were considered stable if the removal of any IVs did not result in a significant change in the results.All results are shown as ORs and 95% confidence intervals (CIs), with differences considered statistically significant at p < .05.All analyses in this MR study were performed using the R software packages "TwoSampleMR," "MRPRESSO," "Coloc," and "locuscompare" package. 21

| Source of scRNA and bulk RNA sequencing data
The Tumor Immune Single-cell Hub (TISCH) is a single-cell RNA sequencing (scRNA-seq) database (http://tisch.comp-genomics.org/)dedicated to the tumor microenvironment (TME).TISCH automatically parses and organizes scRNA-seq data sourced from GEO or Array Express databases, specifically focusing on tumor scRNA-seq datasets.
All datasets undergo a standardized workflow that includes quality control, batch effect removal, cell clustering, differential expression analysis, and multi-level cell type annotation.Each dataset featured in TISCH provides relevant study information, such as species, treatment details, patient and cell numbers, technology platform, stage of the study, and associated research studies.The comprehensive nature of TISCH ensures that the data are uniformly processed, allowing for a more reliable and standardized analysis across different datasets within the TME. 22Gene expression profiling data and related clinical information for 168 glioblastoma samples from The Cancer Genome Atlas Program (TCGA) were then obtained by download from the UCSC Xena website (https://xena.ucsc.edu/).The Wilcoxon rank sum test was used to identify β1-adrenergic receptor (ADRB1)-associated pathologic features.

| Identification of IVs
The overall research idea of this study is shown in Figure 1.The target genes of various types of antihypertensive drugs such as ACEIs, ARBs, BBs, CCBs, and TDs are shown in Table S1, and the chromosomal locations of these target genes are shown in Table S2.In addition, we examined SNPs associated with different antihypertensive drugs and corresponding target genes.GWAS summary statistics were performed for each exposure and outcome after the exclusion of illdefined SNPs and SNPs that were not available in the results.If there were less than five SNPs available for the corresponding exposure and outcome, the corresponding exposure was also removed by us.The results showed that among the drugs mediating SBP, BBs targeting ADRB1 had 12 SNPs, CCBs targeting CACNA1D had 10 SNPs, CCBs targeting CACNB2 had 42 SNPs, and TDs targeting SLC12A2 had 6 SNPs.Among the drugs mediating DBP, BBs targeting ADRB1 had 14 SNPs, CCBs targeting CACNA1D had 9 SNPs, and CCBs targeting CACNB2 had 43 SNPs.In all integrated data, no abnormal aberrant IVs were detected by MR-PRESSO.The final IVs used for MR analysis are shown in Tables S3 and S4.
To assess the reliability of the IVs used in this study, we analyzed the causal effect of antihypertensive drug-mediated SBP/DBP on the positive control outcome (HF) (Figure S1).The results showed that BBs targeting ADRB1 and CCBs targeting CACNA1D/CACNB2 significantly reduced the risk of HF (OR <1, p < .05),suggesting that their IVs were reliable in subsequent MR analyses regarding GBM.In contrast, TDs targeting SLC12A2 increased the risk of HF but the results were not reliable (OR > 1, p > .05).

| Causal effects of SBP/DBP on glioblastoma mediated by antihypertensive drugs
Subsequently, we further assessed the causal effect of SBP/DBP on GBM mediated through the target genes of BBs and CCBs, and the results are shown in Figure 2. The IVW approach revealed that BBs targeting ADRB1 significantly reduced the risk of GBM by mediating DBP (OR = 0.431, 95% CI: 0.267-0.697,p < .001),And it also significantly reduced the risk of GBM by mediating SBP (OR = 0.595, 95% CI: 0.422-0.837,p = .003).In addition, all three other MR methods showed similar results, although the p-values were not significant.We then proceeded to analyze the effect of CCBs on GBM risk.According to the IVW approach, CCBs targeting CACNB2 significantly increased the risk of GBM by mediating DBP (OR = 1.597, 95% CI: 1.168-2.183,p = .003),and they also significantly increased the risk of GBM by mediating SBP (OR = 1.321, 95% CI: 1.101-1.584,p = .003).Both MR Egger and Weight median also support this finding.Interestingly, by mediating DBP, CCBs targeting CACNA1D significantly increased the risk of GBM (OR = 2.034, 95% CI: 1.289-3.208,p = .002),yet their mediated SBP was not causally associated with the risk of GBM ( p > .05).In conclusion, our study tentatively suggests that BBs targeting ADRB1 and CCBs targeting CACNB2 are likely to have some causal association with GBM.

| Sensitivity analysis
Cochran Q and MR Egger regression equations were used to assess the levels of heterogeneity and horizontal pleiotropy (Tables S5 and   S6).In the IVW-MR and MR-egger test methods, the Cochran Q test did not reveal any heterogeneity between the reported results (p > .05).However, the MR-egger test showed that SBP mediated by CCBs targeting CACNA1D was affected by horizontal pleiotropy (p < .05),whereas the overall horizontal pleiotropy of the intercept terms for the other exposures was not significant (p > .05).Such F I G U R E 1 Framework and process for Mendelian randomization (MR) studies of drug targets.To validate the existence of a causal relationship, the following conditions must be met: (1) the instrumental variable is not associated with confounders (dashed line), ( 2) the instrumental variable is associated with exposure (solid line), and (3) the instrumental variable is not directly associated with the outcome (dashed line).ADRB1, β1-adrenergic receptor; DBP, diastolic blood pressure; eQTLs, expression quantitative trait loci; GBM, glioblastoma multiforme; GWAS, genome-wide association study; HF, heart failure; SBP, systolic blood pressure; SNPs, single nucleotide polymorphisms; TCGA, The Cancer Genome Atlas; TISCH, Tumor Immune Single-cell Hub.results violated the core assumptions of MR, so we excluded CCBs in subsequent further analyses.Finally, leave-one-out tests showed that the MR results were stable, with no significant differences in the results after excluding any of the SNPs associated with BBs targeting ADRB1 (Figure 3).

| Colocalization analysis
We performed colocalization analyses at the SNP level to assess evidence for common causal variants between the ADRB1 gene and GBM.We used Bayesian colocalization analysis to rank SNPs located F I G U R E 2 Effects of antihypertensive drugs targeting different genes on glioblastoma.MR, Mendelian randomization.
F I G U R E 3 Sensitivity analysis of BBs targeting ADRB1 as a therapeutic target on glioblastoma.If the combined effect of the remaining single nucleotide polymorphisms (SNPs) after removal of one SNP was consistent with the main effect, a leave-one-out approach was used to assess the over-representation of individual SNPs on the Mendelian randomization (MR) analysis.±1000 kb between the ADRB1 gene and GBM risk.The colocalization analysis showed that SNPs associated with ADRB1 expression and GBM risk had a posteriori probability of sharing a causal variant in the ADRB1 locus of 3.2% (PH1 = 0.00%, PH2 = 62.8%,PH3 = 34%, PH4 = 3.2%).This suggests that there is a low likelihood that there is a common causal variation between the two.Finally, we used the R  package "locuscompare" to show the linkage disequilibrium of SNPs and the distribution of the lead SNP (rs4917675) in both data (Figure 4).

| Bioinformatics analysis
Finally, we validated ADRB1 expression by bulk RNA-seq and scRNAseq.In the TCGA cohort, ADRB1 expression was relatively lower in GBM (Figure 5A).Based on TCGA, GBM can be categorized into four subtypes, that is, classical (TCGA-CL), mesenchymal (TCGA-ME), preneurological (TCGA-PN), and neurological (TCGA-NE), in which the PN subtype has relatively the best prognosis, but the ME subtype is more aggressive and has a poorer prognosis. 23We found that ADRB1 expression was relatively lower in the ME subtype (Figure 5B).A total of eight cell types including oligodendrocytic precursor cell-like (OPC-like) malignant cells, astrocyte-like (AC-like) malignant cells, neural progenitor cell-like (NPC-like) malignant cells, mesenchymal-like (MES-like) malignant cells, malignant cells, oligodendrocytes, CD8 T cells, and myeloid cells were annotated in the GSE131928 data (Figure 5C,D).Where ADRB1 overall had low expression in GBM and its expression was relatively lowest in MES-like malignant cells (Figure 5E,F).

| DISCUSSION
We used GWAS pooled data and eQTL colocalization analyses in MR analyses of two samples to infer the potential pathogenic impact of antihypertensive drugs on glioblastoma.Our study preliminarily suggests that ADRB1-targeted BBs are likely to have a causal relationship with the risk of GBM.By mediating DBP, ADRB1-targeted BBs significantly reduced the risk of GBM (OR = 0. 431, 95% CI: 0.267-0.697,p < .001),and by mediating the level of DBP, BBs also significantly reduced the risk of GBM (OR = 0.595, 95% CI: 0.422-0.837,p = .003).The combined analyses of systolic and DBP suggested a potential protective effect of β-receptor blockers (BBs) against GBM.
The consistency across the results of the five methods underscores the accuracy of our study.Various sensitivity analyses were conducted to affirm the reliability of our findings.Nevertheless, the MR-Egger test indicated that systolic BP mediated by CCBs targeting CACNA1D was influenced by horizontal pleiotropy ( p < .05).As a result, we excluded CCBs as an exposure in our subsequent analyses and conclusions.Of course, we also performed leave-one-out sensitivity tests, which showed that the relationship between causal BBs and GBM was not affected by a single SNP.Finally, colocalization analyses further support the possibility, albeit with a relatively low probability (PH4 = 3.2%), that there may be common causal variants in the loci associated with BBs and GBMs targeting ADRB1.
β-adrenergic receptor blockers, commonly prescribed for HF treatment, function to alleviate cardiac load and enhance cardiac function by inhibiting β-adrenergic receptors' activity.The ADRB1 gene encodes the ADRB1, a pivotal target for all BBs.Variations in the ADRB1 gene have been identified as predictors of responses to antihypertensive drugs. 24Extensive evidence suggests that specific ADRB1 genotypes are linked to cardiovascular event risks, which could be mitigated by BB treatment, implying varying benefits based on genotypes. 25ADRB1, identified as a target gene for β-adrenergic receptor blockers in the DrugBank database, was employed in this MR study as a tool to substitute the exposure of BBs.Given their role as ADRB1 antagonists, BBs may contribute to a reduced risk of glioblastoma.
The β-adrenergic receptor (β-AR) signaling pathway is a pivotal player in tumor progression and metastasis, impacting various facets of tumor behavior, including growth, migration, invasiveness, apoptosis, and angiogenesis. 26These effects are orchestrated through the activation of cancer cell β-AR, subsequently initiating downstream cellular cyclic AMP-protein kinase A signaling pathways.The induction of catecholamines is intricately linked to the activation of genes associated with metastasis, inflammation, cell proliferation pathways, and the upregulation of pro-angiogenic factors.However, the relationship between β-AR blockade and cancer is complex.Several epidemiological studies have reported a reduced risk of cancer development in patients using β-adrenergic receptor blockers.For instance, one study observed higher survival rates among colorectal cancer patients utilizing β-adrenergic receptor blockers. 27In addition, studies have found that the use of β-adrenergic receptor blockers can inhibit tumor growth and metastasis.For example, a prospective cohort study of 839 individuals monitored for 10 years showed that β-adrenergic receptor blockers may lead to a reduction in cancer risk. 28However, there is some controversy regarding the effect of β-adrenergic receptor blockers on cancer risk.0][31] Therefore, more research is needed to clarify the effect of β-adrenergic receptor blockers on cancer risk.
In several recent studies, researchers analyzed gene expression and found that the expression of ADRB1 was significantly increased in CD8+ T cells in the depleted state and was closely related to other markers of T cell depletion. 32The high activity of ADRB1 resulted in the inhibition of T cell receptor signaling and activation of the cAMP signaling pathway, which together contribute to CD8+ T cells entering a state of depletion, which is manifested by decreased cellular function and impairment of antigenic response.Tumors were shown to be significantly smaller in melanoma mice treated with β-blockers in combination with anti-PD-1 and anti-CTLA-4 immunotherapy. 33lation of β1-adrenergic signaling limited the progression of T-cells to the exhausted state and improved effector function when combined with immune checkpoint blockade in melanoma.This study provides new ideas for future immunotherapy, and pharmacological interventions targeting ADRB1 may be considered to improve the efficacy of immunotherapy, particularly in immunocompetent tumors.
There is relatively limited information on the effects of BBs in gliomas.Pavlova et al. 34 conducted an experiment in which they injected rat glioma cells into the brain and administered a β-AR antagonist to half of the rats from the date of implantation, while the control rats were given saline.The results showed that the group of rats receiving the β-AR antagonist survived significantly longer.Confocal imaging showed that β-AR blockade reduced glioma cell migration by 20% and blood-brain barrier disruption was also significantly reduced. 35other study conducted a retrospective cohort study involving 218 patients to investigate the effect of BBs on recurrent GBM.They compared patients who received BBs with patients who did not receive AR-modifying drugs, and both groups were treated with bevacizumab.Unfortunately, no association was shown between the use of BBs and OS or PFS. 36Glioblastoma has been described as a "desert" for immunotherapy because of its strongly immunosuppressive tumor environment, resulting in a limited response to conventional immunotherapy.In recent years, many scientists have attempted to treat this tumor with immune-enhancing therapeutic agents, but have failed to achieve the expected results.In our study, we initially found that the expression activity of ADRB1 in GBM is relatively low through bioinformatics, which is promising to explain why BBs have a positive effect on gliomas under laboratory conditions, but not in clinical observation.
Numerous studies have explored the association between antihypertensive drugs and cancer, encompassing randomized controlled trials, fundamental laboratory research, and epidemiological investigations.Nevertheless, no prior investigations have specifically examined the correlation between antihypertensive drugs and the risk of developing GBM.Our study represents a groundbreaking initiative, utilizing a two-sample MR approach, to establish a potential causal link between antihypertensive drugs and a reduced risk of GBM.However, it is crucial to acknowledge certain limitations inherent in our study.First, since our study mainly included individuals of European origin, the generalization of the results may be somewhat limited.We mention this in our results, but the consistency or variability of the results across other groups needs to be explored in more depth in the discussion.Future studies could consider validation in a wider population.In addition to this, MR-PRESSO did not detect abnormal IVs, but these hypothesized limitations still need to be treated with caution.In addition, unmeasured confounders could potentially interfere with the results, and therefore the results need to be interpreted in the context of these potential influences.The present study focused on BBs and CCBs; whether similar trends or other causal relationships exist for the effects of other antihypertensive drugs still requires more in-depth study.Further study of the effects of different classes of antihypertensive drugs may help us gain a more comprehensive understanding of the complexity of this field.MR studies focused on drug targets may not comprehensively reflect the actual impacts of drugs, as they are unable to consider confounding factors like drug dosage and metabolism, mechanism of action, individual variations, and the duration of drug exposure.Finally, colocalization analyses provide some support for causality, but there is some uncertainty.Our results suggest the existence of possible cocausal variants within the ADRB1 locus, but the posterior probability is not 100%.This uncertainty needs to be fully recognized in the conclusions, and future studies may need to further validate these results.

F I G U R E 4
The left plot represents the distribution of single nucleotide polymorphisms (SNPs) in genome-wide association study (GWAS) and expression quantitative trait loci (eQTL) Àlog10(p), with smaller pvalues above the Y-axis.The two plots on the right are each subtabulated to represent the distribution of eQTL and GWAS (the horizontal coordinates are the loci of the SNPs, and the vertical coordinates represent the Àlog10(p) value of the SNP in that GWAS/eQTL data).The rsid, which is labeled, is the value that adds up to the smallest p-value in the two data, that is, the lead SNP.

F I G U R E 5
Bioinformatics analysis of β1-adrenergic receptor (ADRB1).(A) ADRB1 expression in the Cancer Genome Atlas cohort.(B) ADRB1 expression in glioblastoma multiforme of different molecular subtypes.(C, D) Distribution and percentage of different cell types in single-cell sequencing data.(E, F) ADRB1 expression in different cell types.CL, classical; ME, mesenchymal; NE, neurological; PN, preneurological.