UK Biobank Genome-wide Association Study Research Articles

Unsupervised deep representation learning enables phenotype discovery for genetic association studies of brain imaging

Understanding the genetic architecture of brain structure is challenging, partly due to difficulties in designing robust, non-biased descriptors of brain morphology. Until recently, brain measures for genome-wide association studies (GWAS) consisted of traditionally expert-defined or software-derived image-derived phenotypes (IDPs) that are often based on theoretical preconceptions or computed from limited amounts of data. Here, we present an approach to derive brain imaging phenotypes using unsupervised deep representation learning. We train a 3-D convolutional autoencoder model with reconstruction loss on 6130 UK Biobank (UKBB) participants’ T1 or T2-FLAIR (T2) brain MRIs to create a 128-dimensional representation known as Unsupervised Deep learning derived Imaging Phenotypes (UDIPs). GWAS of these UDIPs in held-out UKBB subjects (n = 22,880 discovery and n = 12,359/11,265 replication cohorts for T1/T2) identified 9457 significant SNPs organized into 97 independent genetic loci of which 60 loci were replicated. Twenty-six loci were not reported in earlier T1 and T2 IDP-based UK Biobank GWAS. We developed a perturbation-based decoder interpretation approach to show that these loci are associated with UDIPs mapped to multiple relevant brain regions. Our results established unsupervised deep learning can derive robust, unbiased, heritable, and interpretable brain imaging phenotypes.

Targeting postsynaptic glutamate receptor scaffolding proteins PSD-95 and PICK1 for obesity treatment.

Human genome-wide association studies (GWAS) suggest a functional role for central glutamate receptor signaling and plasticity in body weight regulation. Here, we use UK Biobank GWAS summary statistics of body mass index (BMI) and body fat percentage (BF%) to identify genes encoding proteins known to interact with postsynaptic α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) and N-methyl-d-aspartate (NMDA) receptors. Loci in/near discs large homolog 4 (DLG4) and protein interacting with C kinase 1 (PICK1) reached genome-wide significance (P < 5 × 10-8) for BF% and/or BMI. To further evaluate the functional role of postsynaptic density protein-95 (PSD-95; gene name: DLG4) and PICK1 in energy homeostasis, we used dimeric PSD-95/disc large/ZO-1 (PDZ) domain-targeting peptides of PSD-95 and PICK1 to demonstrate that pharmacological inhibition of PSD-95 and PICK1 induces prolonged weight-lowering effects in obese mice. Collectively, these data demonstrate that the glutamate receptor scaffolding proteins, PICK1 and PSD-95, are genetically linked to obesity and that pharmacological targeting of their PDZ domains represents a promising therapeutic avenue for sustained weight loss.

Relationship Between Sleep Duration and Cognitive Function in Older Adults: Analysis of NHANES and UK Biobank GWAS Data.

To explore the relationship between sleep duration and cognitive functions in older adults using NHANES, a national US population study dataset, and to explore the causal association with Mendelian randomization (MR) using the UK Biobank. First, an observational study was conducted with the NHANES database with participants ≥60 years. Sleep duration was measured with accelerometers for 7 consecutive days. Participants were divided into habitual short sleep (<7h) and long sleep (>9h) groups. Cognitive functions were measured with the CERAD Word Learning sub-set, Animal Fluency, and Digit Symbol Substitution test (DSST). Multivariate regression models were used to explore relationships between sleep duration and cognitive functions. Second, bidirectional MR was conducted with data for self-reported sleep duration, which came from a genome-wide association study (GWAS) comprising 446,118 adults from the UK Biobank, and general cognitive performance, which was obtained from a recent GWAS study (N = 257,841). Inverse-variance weighted (IVW) was used as the primary estimation of the outcome. In the observational study, 2687 participants were included. Sleep duration was associated with cognitive functions in a non-linear way. Habitual long sleep (>9°h) was associated with lower scores on DSST (OR = 0.01, p = .003) in the fully-adjusted model. The association between habitual short sleep and cognitive functions was insignificant. For the MR, genetically predicted lower general cognitive performance was causally associated with a higher prevalence of habitual short sleep (OR = 0.97, p = 5.1 × 10-7) and long sleep (OR = 0.97, p = 8.87 × 10-16). Short and long sleep duration might be both causally associated with worse outcomes of cognitive functions in older adults, highlighting the importance of maintaining sleep health.

Anxiety disorder and cardiovascular disease: a two-sample Mendelian randomization study.

Cardiovascular disease is the leading cause of death worldwide. Anxiety disorders are common psychiatric conditions associated with cardiovascular outcomes. This two-sample Mendelian randomization (MR) study investigated the causal relationship between anxiety disorders and coronary heart disease (CHD), myocardial infarction (MI), heart failure (HF), and atrial fibrillation (AF). Single nucleotide polymorphisms (SNPs) associated with anxiety disorders (16730 cases; 101021 controls) were obtained from the UK Biobank genome-wide association study (GWAS). Cardiovascular outcome data were derived from the FinnGen study (CHD: 21012 cases and 197780 controls; MI: 12801 cases and 187840 controls; HF: 23397 cases and 194811 controls; and AF: 22068 cases and 116926 controls). Inverse variance weighted (IVW), MR-Egger, weighted median, simple mode, and weighted mode analyses examined causality. IVW analysis demonstrated significant causal relationships between anxiety disorders and increased risk of CHD [odds ratio (OR): 4.496; 95% confidence interval (CI): 1.777-11.378; P=0.002], MI (OR: 5.042; 95% CI: 1.451-17.518; P=0.011), and HF (OR: 3.255; 95% CI: 1.461-7.252; P=0.004). No relationship was observed with AF (OR: 1.775; 95% CI: 0.612-5.146; P=0.29). Other methods showed non-significant associations. Two-way analysis indicated no reverse causality. Anxiety disorders were causally associated with greater risk of CHD, MI, and HF but not AF among individuals of European descent. Further research on mediating mechanisms and in diverse populations is warranted.

Principal and independent genomic components of brain structure and function.

The highly polygenic and pleiotropic nature of behavioural traits, psychiatric disorders and structural and functional brain phenotypes complicate mechanistic interpretation of related genome-wide association study (GWAS) signals, thereby obscuring underlying causal biological processes. We propose genomic principal and independent component analysis (PCA, ICA) to decompose a large set of univariate GWAS statistics of multimodal brain traits into more interpretable latent genomic components. Here we introduce and evaluate this novel methods various analytic parameters and reproducibility across independent samples. Two UK Biobank GWAS summary statistic releases of 2240 imaging-derived phenotypes (IDPs) were retrieved. Genome-wide beta-values and their corresponding standard-error scaled z-values were decomposed using genomic PCA/ICA. We evaluated variance explained at multiple dimensions up to 200. We tested the inter-sample reproducibility of output of dimensions 5, 10, 25 and 50. Reproducibility statistics of the respective univariate GWAS served as benchmarks. Reproducibility of 10-dimensional PCs and ICs showed the best trade-off between model complexity and robustness and variance explained (PCs: |rz - max| = 0.33, |rraw - max| = 0.30; ICs: |rz - max| = 0.23, |rraw - max| = 0.19). Genomic PC and IC reproducibility improved substantially relative to mean univariate GWAS reproducibility up to dimension 10. Genomic components clustered along neuroimaging modalities. Our results indicate that genomic PCA and ICA decompose genetic effects on IDPs from GWAS statistics with high reproducibility by taking advantage of the inherent pleiotropic patterns. These findings encourage further applications of genomic PCA and ICA as fully data-driven methods to effectively reduce the dimensionality, enhance the signal to noise ratio and improve interpretability of high-dimensional multitrait genome-wide analyses.

Uncovering myocardial infarction genetic signatures using GWAS exploration in Saudi and European cohorts

Genome-wide association studies (GWAS) have yielded significant insights into the genetic architecture of myocardial infarction (MI), although studies in non-European populations are still lacking. Saudi Arabian cohorts offer an opportunity to discover novel genetic variants impacting disease risk due to a high rate of consanguinity. Genome-wide genotyping (GWG), imputation and GWAS followed by meta-analysis were performed based on two independent Saudi Arabian studies comprising 3950 MI patients and 2324 non-MI controls. Meta-analyses were then performed with these two Saudi MI studies and the CardioGRAMplusC4D and UK BioBank GWAS as controls. Meta-analyses of the two Saudi MI studies resulted in 17 SNPs with genome-wide significance. Meta-analyses of all 4 studies revealed 66 loci with genome-wide significance levels of p < 5 × 10–8. All of these variants, except rs2764203, have previously been reported as MI-associated loci or to have high linkage disequilibrium with known loci. One SNP association in Shisa family member 5 (SHISA5) (rs11707229) was evident at a much higher frequency in the Saudi MI populations (> 12% MAF). In conclusion, our results replicated many MI associations, whereas in Saudi-only GWAS (meta-analyses), several new loci were implicated that require future validation and functional analyses.

Changes in iron load in specific brain areas lead to neurodegenerative diseases of the central nervous system

The causes of neurodegenerative diseases remain largely elusive, increasing their personal and societal impacts. To reveal the causal effects of iron load on Parkinson's disease (PD), Alzheimer's disease (AD), amyotrophic lateral sclerosis and multiple sclerosis, we used Mendelian randomisation and brain imaging data from a UK Biobank genome-wide association study of 39,691 brain imaging samples (predominantly of European origin). Using susceptibility-weighted images, which reflect iron load, we analysed genetically significant brain regions. Inverse variance weighting was used as the main estimate, while MR Egger and weighted median were used to detect heterogeneity and pleiotropy. Nine clear associations were obtained. For AD and PD, an increased iron load was causative: the right pallidum for AD and the right caudate, left caudate and right accumbens for PD. However, a reduced iron load was identified in the right and left caudate for multiple sclerosis, the bilateral hippocampus for mixed vascular dementia and the left thalamus and bilateral accumbens for subcortical vascular dementia. Thus, changes in iron load in different brain regions have causal effects on neurodegenerative diseases. Our results are crucial for understanding the pathogenesis and investigating the treatment of these diseases.

Investigating the Impact of Sun/UV Protection and Ease of Skin Tanning on the Risk of Pseudoexfoliation Glaucoma: A Mendelian Randomization Study.

To investigate the potential causal associations between the use of sun/ultraviolet (UV) protection and ease of skin tanning and the risk of pseudoexfoliation glaucoma (PXG) in European populations. Single nucleotide polymorphisms (SNPs) associated with the use of sun/UV protection and ease of skin tanning were selected from the UK Biobank genome-wide association study database consisting of 498,751 European participants. SNPs of PXG were obtained from the FinnGen study including 3424 PXG cases and 326,434 controls. Two-sample Mendelian randomization (MR) analyses were performed to assess the association between the use of sun/UV protection and ease of skin tanning and risk of PXG. Inverse variance weighted regression of genetic susceptibility predicted that both use of sun/UV protection and ease of skin tanning were potentially positively associated with the decreased risk of PXG in the European ancestry (use of sun/UV protection: odds ratio [OR] = 0.47; 95% confidence interval [CI], 0.24-0.92; P = 0.028; ease of skin tanning: OR = 0.81; 95% CI, 0.67-0.97; P = 0.025). We found genetic evidence supporting a potential causal association between UV protection and a decreased risk of PXG in European population. Further research will help elucidate the underlying mechanisms and promote UV protection for eyes, especially in people with a high risk of PXG.

THE ASSOCIATION BETWEEN BLOOD PRESSURE VARIABILITY AND DEMENTIA: A TWO-SAMPLE MENDELIAN RANDOMIZATION STUDY

Objective: This study aims to investigate the causal inference by applying a mendelian randomization (MR) study using single-nucleotide polymorphisms (SNPs) from the genome-wide association study (GWAS) based on the UK biobank. Design and method: The genetic instruments were chosen from BPV GWAS based on UK Biobank data. Four GWA studies were used for the genetic evaluation of dementia, including (1) The International Genomic of Alzheimer's Project (IGAP) GWAS for dementia; (2) and (3) UK Biobank GWAS for proxy phenotypes of Alzheimer's dementia: the maternal and paternal family history (MFH and PFH-UKBB); (4) A combined GWAS meta-analysis from the abovementioned three GWAS results. The following exclusion criteria were applied for SNPs: (1) confounding SNPs by searching hypertension and artery stiffness GWAS; (2) significant SNPs that are associated with dementia in the datasets; (3) palindromic SNPs with an effect allele frequency between 0.4 and 0.7. Proxy SNPs were manually selected for some SNPs. (4) only SNPs with linkage disequilibrium r2 &lt; 0.01 were selected as genetic instruments, also known as independent SNPs. The effect/reference alleles were checked to ensure the consistency of the base pair between BPV and dementia GWAS, which is also known as data harmonization. After data harmonization, four MR methods were employed to measure the causal effect: IVW-MR with random effect, IVW-MR with fixed effect, MR-Egger, and weighted median MR. Results: Finally, six independent SNPs were chosen as instrument variants for SBPV, and five for DBPV. Significant causal effects of SBPV on dementia were found in the PFH-UKBB dataset, and the odds ratio of dementia per 10-unit increase in SBPV was 1.028, 1.015, and 1.015 for MR-Egger, weighted median, and IVW-MR, respectively (Table 1). In contrast, only one significant result was found for DBPV in MFH-UKBB. No significant results were found on other datasets. Conclusions: Along with the evidence from different observational studies, this GWAS study demonstrates that systolic blood pressure variability is a potential causal risk factor for Alzheimer's dementia, while the evidence from diastolic blood pressure variability is still uncertain.

Genetic basis of STEM occupational choice and regional economic performance: a UK biobank genome-wide association study

BackgroundScience, technology, engineering, and mathematics (STEM) professionals are regarded as the highly skilled labor force that fosters economic productivity, enterprise innovation, and international competitiveness of a country. This study aims to understand the genetic predisposition to STEM occupations and investigate its associations with regional economic performance. We conducted a genome-wide association study on the occupational choice of STEM jobs based on a sample of 178,976 participants from the UK Biobank database.ResultsWe identified two genetic loci significantly associated with participants’ STEM job choices: rs10048736 on chromosome 2 and rs12903858 on chromosome 15. The SNP heritability of STEM occupations was estimated to be 4.2%. We also found phenotypic and genetic evidence of assortative mating in STEM occupations. At the local authority level, we found that the average polygenic score of STEM is significantly and robustly associated with several metrics of regional economic performance.ConclusionsThe current study expands our knowledge of the genetic basis of occupational choice and potential regional disparities in socioeconomic developments.

Assessing the association of leukocyte telomere length with ankylosing spondylitis and rheumatoid arthritis: A bidirectional Mendelian randomization study.

Telomere length shortening can cause senescence and apoptosis in various immune cells, resulting in immune destabilization and ageing of the organism. In this study, we aimed to systematically assess the causal relationship of leukocyte telomere length (LTL) with ankylosing spondylitis (AS) and rheumatoid arthritis (RA) using a Mendelian randomization study. LTL (n=472174) was obtained from the UK Biobank genome-wide association study pooled data. AS (n=229640), RA (n=212472) were obtained from FinnGen database. MR-Egger, inverse variance weighting, and weighted median methods were used to estimate the effects of causes. Cochran's Q test, MR Egger intercept test, MR-PRESSO, leave-one-out analysis, and funnel plots were used to look at sensitivity, heterogeneity, and multiple effects. Forward MR analysis considered LTL as the exposure and AS, RA as the outcome. Reverse MR analysis considered AS, RA as the exposure and LTL as the outcome. In the forward MR analysis, inverse variance-weighted and weighted median analysis results indicated that longer LTL might be associated with increased risk of AS (IVW: OR = 1.55, 95% CI: 1.14-2.11, p = 0.006). MR Egger regression analysis showed no pleiotropy between instrumental variables (IVs) (Egger intercept= 0.008, p = 0.294). The leave-one-out analysis showed that each single nucleotide polymorphism (SNP) of AS was robust to each outcome. No significant causal effects were found between AS, RA and LTL in the reverse MR analysis. Longer LTL may be related with an increased risk of developing AS, and these findings provide a foundation for future clinical research on the causal association between LTL and AS.

A clustering linear combination method for multiple phenotype association studies based on GWAS summary statistics

There is strong evidence showing that joint analysis of multiple phenotypes in genome-wide association studies (GWAS) can increase statistical power when detecting the association between genetic variants and human complex diseases. We previously developed the Clustering Linear Combination (CLC) method and a computationally efficient CLC (ceCLC) method to test the association between multiple phenotypes and a genetic variant, which perform very well. However, both of these methods require individual-level genotypes and phenotypes that are often not easily accessible. In this research, we develop a novel method called sCLC for association studies of multiple phenotypes and a genetic variant based on GWAS summary statistics. We use the LD score regression to estimate the correlation matrix among phenotypes. The test statistic of sCLC is constructed by GWAS summary statistics and has an approximate Cauchy distribution. We perform a variety of simulation studies and compare sCLC with other commonly used methods for multiple phenotype association studies using GWAS summary statistics. Simulation results show that sCLC can control Type I error rates well and has the highest power in most scenarios. Moreover, we apply the newly developed method to the UK Biobank GWAS summary statistics from the XIII category with 70 related musculoskeletal system and connective tissue phenotypes. The results demonstrate that sCLC detects the most number of significant SNPs, and most of these identified SNPs can be matched to genes that have been reported in the GWAS catalog to be associated with those phenotypes. Furthermore, sCLC also identifies some novel signals that were missed by standard GWAS, which provide new insight into the potential genetic factors of the musculoskeletal system and connective tissue phenotypes.

A shared genetic architecture between adhesive capsulitis and Dupuytren disease.

The etiology of adhesive capsulitis involves inflammation, thickening, and fibrosis of the shoulder capsule. The underlying genetic factors are poorly understood. The purpose of this study was to identify genetic variants associated with adhesive capsulitis using the UK Biobank (UKB) cohort and compare them with variants associated with Dupuytren disease investigating a common etiology between the 2 fibrotic disorders. A genome-wide association study (GWAS) was performed using data from UKB with 10,773 cases of adhesive capsulitis, and a second GWAS was performed with 8891 cases of Dupuytren disease. Next, a comparison of association statistics was performed between adhesive capsulitis and Dupuytren disease using the data from both GWAS. Finally, single-nucleotide polymorphisms (SNPs) previously reported from candidate gene studies for adhesive capsulitis and Dupuytren disease were tested for association with adhesive capsulitis and Dupuytren disease using the summary statistics from their respective GWAS. The UKB GWAS for adhesive capsulitis identified 6 loci that reached genome-wide statistical significance: a cluster of 11 closely linked SNPs on chromosome 1; a single SNP on chromosome 2; a single SNP on chromosome 14; 2 closely linked SNPs on chromosome 21; 33 closely linked SNPs on chromosome 22; and 3 closely linked SNPs on the X chromosome. These SNPs were associated with 8 different genes including TSPAN2/NGF, SATB2, MRPL52/MMP14, ERG, WNT7B, and FGF13. A GWAS for Dupuytren disease was performed and a comparison to the adhesive capsulitis GWAS showed 13 loci significantly associated with both phenotypes. A validation attempt of 6 previously reported SNPs associated with adhesive capsulitis using UKB summary statistics was unable to confirm any of the previously reported SNPs (all P>.19). All 23 previously reported SNPs associated with Dupuytren disease were confirmed using the UKB summary statistics (P<2.1×10-3) CONCLUSION: This GWAS investigating adhesive capsulitis has identified 6 novel loci involving 8 different genes to be associated with adhesive capsulitis. A GWAS investigating Dupuytren disease was performed and compared to the adhesive capsulitis GWAS, and 13 common loci were identified between the 2 disorders with genes involved in pathologic fibrosis. We were unable to validate the SNPs in candidate genes previously reported to be associated with adhesive capsulitis although we were able to confirm all previously reported SNPs associated with Dupuytren disease. The strong genetic overlap between the adhesive capsulitis and Dupuytren disease loci suggests a similar etiology between the 2 diseases.

A fast and powerful linear mixed model approach for genotype-environment interaction tests in large-scale GWAS.

Genotype-by-environment interaction (GEI or GxE) plays an important role in understanding complex human traits. However, it is usually challenging to detect GEI signals efficiently and accurately while adjusting for population stratification and sample relatedness in large-scale genome-wide association studies (GWAS). Here we propose a fast and powerful linear mixed model-based approach, fastGWA-GE, to test for GEI effect and G + GxE joint effect. Our extensive simulations show that fastGWA-GE outperforms other existing GEI test methods by controlling genomic inflation better, providing larger power and running hundreds to thousands of times faster. We performed a fastGWA-GE analysis of ~7.27 million variants on 452 249 individuals of European ancestry for 13 quantitative traits and five environment variables in the UK Biobank GWAS data and identified 96 significant signals (72 variants across 57 loci) with GEI test P-values < 1 × 10-9, including 27 novel GEI associations, which highlights the effectiveness of fastGWA-GE in GEI signal discovery in large-scale GWAS.

Association of genetically predicted lipid traits and lipid-modifying targets with heart failure.

To assess the association of genetically predicted lipid traits and lipid-modification via licensed or investigational targets with heart failure (HF). Two-sample Mendelian randomization (MR) study was conducted using summary-level genome-wide association studies (GWASs) from UK Biobank and HERMES Consortium. Genetic variants obtained from UK Biobank GWAS data were selected as instrumental variables to predict the level of lipid traits [LDL cholesterol (LDL-C), HDL cholesterol (HDL-C), triglyceride (TG), apolipoprotein B (ApoB), and apolipoprotein AI (ApoAI)] and lipid-modifying effect of eight drug targets [HMGCR, PCSK9, NPC1L1, PPARA, lipoprotein lipase (LPL), ANGPTL3, APOC3, and cholesteryl ester transfer protein (CETP)]. In this study, we observed that genetically predicted LDL-C, TG, HDL-C or ApoB were significantly related to HF, which were mainly mediated by coronary heart disease (CHD). Drug target MR analyses identified PCSK9, CETP, and LPL as potential targets to prevent HF. The genetic proxy of LDL-C and ApoB increase modified by PCSK9 showed similar evidence in increasing risk of HF (PLDL-C = 1.27*10-4; PApoB = 1.94*10-4); CETP played a role in HF risk via modifying all investigational lipid traits with the strongest evidence though ApoB (P = 5.87*10-6); LPL exerted effects on HF via modifying most lipid traits with the strongest evidence observed via modifying TG (P = 3.73*10-12). This two-sample MR study provided genetic evidence of the associations between lipid traits and HF risk, which were mostly mediated by CHD. Besides, drug target MR studies indicated that PCSK9 inhibition, CETP inhibition, and LPL activation were effective in HF reduction.

Association of gut microbiome and primary liver cancer: A two-sample Mendelian randomization and case-control study.

Observational epidemiology studies suggested a relationship between the gut microbiome and primary liver cancer. However, the causal relationship remains unclear because of confounding factors and reverse causality. We aimed to explore the causal role of the gut microbiome in the development of primary liver cancer, including hepatocellular carcinoma (HCC) and intrahepatic cholangiocarcinoma (ICC). Mendelian randomization (MR) study was conducted using summary statistics from genome-wide association studies (GWAS) of the gut microbiome and liver cancer, and sequencing data from a case-control study validated the findings. A 5-cohort GWAS study in Germany (N=8956) served as exposure, whilst the UK biobank GWAS study (N=456 348) served as an outcome. The case-control study was conducted at the First Affiliated Hospital of Wenzhou Medical University from December 2018 to October 2020 and included 184 HCC patients, 63 ICC patients and 40 healthy controls. A total of 57 features were available for MR analysis, and protective causal associations were identified for Family_Ruminococcaceae (OR=0.46 [95% CI, 0.26-0.82]; p=.009) and Genus_Porphyromonadaceae (OR=0.59 [95% CI, 0.42-0.83]; p=.003) with HCC, and for Family_Porphyromonadaceae (OR=0.36 [95% CI, 0.14-0.94]; p=.036) and Genus_Bacteroidetes (OR=0.55 [95% CI, 0.34-0.90]; p=.017) with ICC respectively. The case-control study results showed that the healthy controls had a higher relative abundance of Family_Ruminococcaceae (p=.00033), Family_Porphyromonadaceae (p=.0055) and Genus_Bacteroidetes (p=.021) than the liver cancer patients. This study demonstrates that Ruminococcaceae, Porphyromonadaceae and Bacteroidetes are related to a reduced risk of liver cancer (HCC or ICC), suggesting potential significance for the prevention and control of liver cancer.

Association of autoimmune diseases with Alzheimer's disease: A mendelian randomization study

BackgroundAlzheimer's disease may have an autoimmune component, but the association is unclear. ObjectiveThe objective of this Mendelian randomization (MR) study was to evaluate the association of liability to autoimmune diseases with Alzheimer's disease. MethodsA systematic search was done using PubMed to identify autoimmune diseases that have been suggested as associated with Alzheimer's disease. Genetic predictors of these autoimmune diseases were obtained from the largest and most recent genome-wide association studies (GWAS). Genetic associations with clinically-diagnosed Alzheimer's disease were obtained from the International Genomics of Alzheimer's Project GWAS (21982 cases; 41944 controls); and with parental and sibling history of Alzheimer's disease from the UK Biobank GWAS (27696 maternal, 14338 paternal and 2171 sibling cases). MR estimates were obtained using inverse variance weighting, MR-Egger and weighted median. To address possible selection bias due to inevitably recruiting only survivors, the analysis was repeated in younger people, i.e., UK Biobank siblings and adjusting for competing risk of Alzheimer's disease. ResultsOf the 7 autoimmune diseases considered, liability to psoriasis and sarcoidosis were not associated with Alzheimer's disease. Some evidence was found for liability to multiple sclerosis being associated with higher risk and liability to Sjogren's syndrome with lower risk of Alzheimer's disease. Associations found for liability to giant cell arteritis, type 1 diabetes and rheumatoid arthritis were inconsistent in sensitivity analyses. ConclusionLiability to multiple sclerosis and Sjogren's syndrome could be associated with Alzheimer's disease. The underlying mechanisms, such as the role of myelin and neuroinflammation, should be further investigated.

Coronary artery disease risk factors affected by RNA modification-related genetic variants.

BackgroundSingle nucleotide polymorphisms that affect RNA modification (RNAm-SNPs) may have functional roles in coronary artery disease (CAD). The aim of this study was to identify RNAm-SNPs in CAD susceptibility loci and highlight potential risk factors.MethodsCAD-associated RNAm-SNPs were identified in the CARDIoGRAMplusC4D and UK Biobank genome-wide association studies. Gene expression and circulating protein levels affected by the RNAm-SNPs were identified by QTL analyses. Cell experiments and Mendelian randomization (MR) methods were applied to test whether the gene expression levels were associated with CAD.ResultsWe identified 81 RNAm-SNPs that were associated with CAD or acute myocardial infarction (AMI), including m6A-, m1A-, m5C-, A-to-I- and m7G-related SNPs. The m6A-SNPs rs3739998 in JCAD, rs148172130 in RPL14 and rs12190287 in TCF21 and the m7G-SNP rs186643756 in PVT1 were genome-wide significant. The RNAm-SNPs were associated with gene expression (e.g., MRAS, DHX36, TCF21, JCAD and SH2B3), and the expression levels were associated with CAD. Differential m6A methylation and differential expression in FTO-overexpressing human aorta smooth muscle cells and peripheral blood mononuclear cells of CAD patients and controls were detected. The RNAm-SNPs were associated with circulating levels of proteins with specific biological functions, such as blood coagulation, and the proteins (e.g., cardiotrophin-1) were confirmed to be associated with CAD and AMI in MR analyses.ConclusionThe present study identified RNAm-SNPs in CAD susceptibility genes, gene expression and circulating proteins as risk factors for CAD and suggested that RNA modification may play a role in the pathogenesis of CAD.

Frailty and hearing loss: From association to causation.

BackgroundObservational studies suggest that frailty is associated with hearing loss (HL) but with inconsistent results. This study aims to examine such association and to assess its causality.Materials and methodsThe cross-sectional data from the National Health and Nutrition Examination Survey (NHANES). Multivariate logistic regression models were used to assess the association between HL and frailty index (FI). Genetic variants associated with the FI and HL were obtained from a large genome-wide association study (GWAS) meta-analysis and UK Biobank GWAS. The inverse variance weighting (IVW) method was used to estimate causal effects. Sensitivity analyses were performed to further validate the robustness of results.ResultsIn this cross-sectional analysis, results support the possibility that frailty may be associated with a higher risk of developing HL, with self-reported [odds ratio (OR) = 2.813; 95% CI, 2.386, 3.317; p < 0.001], speech frequency HL (OR = 1.975; 95% CI, 1.679–2.323; p < 0.001), and high frequency HL (OR = 1.748; 95% CI, 1.459–2.094; p < 0.001). In the adjusted model, frail participants remained at high risk of HL. Mendelian randomization (MR) studies showed a bidirectional causal association between genetically predicted FI and risk of HL (FI for exposure: OR = 1.051; 95% CI, 1.020–1.083; p = 0.001; HL for exposure: OR = 1.527; 95% CI, 1.227–1.901; p < 0.001).ConclusionOur observational study found that inter-individual differences in frailty were associated with the risk of developing HL. Genetic evidence suggests a potential bidirectional causal association between FI and HL. Furthermore, the potential mechanisms of this association require investigation.

Statistical power of transcriptome-wide association studies.

Transcriptome-Wide Association Studies (TWASs) have become increasingly popular in identifying genes (or other endophenotypes or exposures) associated with complex traits. In TWAS, one first builds a predictive model for gene expressions using an expression quantitative trait loci (eQTL) data set in stage 1, then tests the association between the predicted gene expression and a trait based on a large, independent genome-wide association study (GWAS) data set in stage 2. However, since the sample size of the eQTL data set is usually small and the coefficient of multiple determination (i.e., ) of the model for many genes is also small, a question of interest is to what extent these factors affect the statistical power of TWAS. In addition, in contrast to a standard (univariate) TWAS (UV-TWAS) considering only a single gene at a time, multivariate TWAS (MV-TWAS) methods have recently emerged to account for the effects of multiple genes, or a gene's nonlinear effects, simultaneously. With the absence of the power analysis for these MV-TWAS methods, it would be of interest to investigate whether one can gain or lose power by using the newly proposed MV-TWAS instead of UV-TWAS. In this paper, we first outline a general method for sample size/power calculations for two-sample TWAS, then use real data-the Alzheimer's Disease Neuroimaging Initiative (ADNI) expression quantitative trait loci (eQTL) data and the Genotype-Tissue Expression (GTEx) eQTL data for stage 1, the International Genomics of Alzheimer's Project Alzheimer's disease (AD) GWAS summary data and UK Biobank (UKB) individual-level data for stage 2-to empirically address these questions. Our most important conclusions are the following. First, a sample size of a few thousands (~8000) would suffice in stage 1, where the power of TWAS would be more determined by cis-heritability of gene expression. Second, as in the general case of simple regression versus multiple regression, the power of MV-TWAS may be higher or lower than that of UV-TWAS, depending on the specific relationships among the GWAS trait and multiple genes (or linear and nonlinear terms of the same gene's expression levels), such as their correlations and effect sizes. Interestingly, several top genes with large power gains in MV-TWAS (over that in UV-TWAS)were known to be (and in our data more significantly) associated with AD. We also reached similar conclusions in an application to the GTEx whole blood gene expression data and UKB GWAS data of high-density lipoprotein cholesterol. The proposed method and the conclusions are expected to be useful in planning and designing future TWAS and other related studies (e.g., Proteome- or Metabolome-Wide Association Studies) when determining the sample sizes for the two stages.