GWAS Summary Data Research Articles

Association of lipid-lowering drugs with osteoarthritis outcomes from a drug-target Mendelian randomization study.

Osteoarthritis (OA), a prevalent musculoskeletal disorder, has been suggested to have a potential association with metabolic syndrome, particularly lipid metabolism. Studies exploring the effects of lipid-lowering drugs on OA have yielded conflicting results. This study employed a drug-targeted Mendelian randomization approach to investigate the association between genetically predicted lipid-modulating effects of commonly targeted lipid-lowering agents and the risk of OA, with the aim of providing a theoretical foundation for the use of lipid-lowering drugs in OA treatment. Employing Mendelian randomization (MR) analysis, we examined the potential causal relationship between lipid-lowering drugs and OA. Genetic variants associated with LDL cholesterol levels were selected from the GWAS summary data, and a series of statistical analyses, including inverse-variance weighted (IVW), weighted median (WM), and MR-Egger, were performed to estimate causal effects. We observed significant associations between genetically proxied lipid-lowering drug targets and OA risk. Notably, HMGCR-mediated LDL cholesterol showed an association with overall OA of the hip or knee (OR = 0.865, 95%CI: 0.762 to 0.983, p = 0.026, q = 0.07) and knee osteoarthritis specifically (OR = 0.746, 95%CI: 0.639 to 0.871, p = 2.180×10-4, q = 0.004). PCSK9-mediated LDL cholesterol also demonstrated an association with OA of the hip or knee (OR = 0.915, 95%CI: 0.847 to 0.988, p = 0.023, q = 0.07) and knee osteoarthritis (OR = 0.901, 95%CI: 0.821 to 0.990, p = 0.03, q = 0.07). NPC1L1-mediated LDL cholesterol showed a positive association with OA of the hip or knee (OR = 1.460, 95%CI: 1.127 to 1.890, p = 0.004, q = 0.033). Furthermore, LDLR-mediated LDL cholesterol demonstrated an association with OA of the hip or knee (OR = 0.882, 95%CI: 0.788 to 0.988, p = 0.03, q = 0.07) and hip osteoarthritis (OR = 0.867, 95%CI: 0.769 to 0.978, p = 0.02, q = 0.07). These findings provide preliminary evidence for the potential therapeutic use of lipid-lowering drugs in OA treatment. Further investigation is needed to validate these findings and explore the precise mechanisms underlying the observed associations.

Causal effect of serum 25 hydroxyvitamin D concentration on cardioembolic stroke: Evidence from two-sample Mendelian randomization

Background and aimsThe putative association between serum 25-hydroxyvitamin D concentration [25(OH)D] and the risk of cardioembolic stroke (CES) has been examined in observational studies, which indicate controversial findings. We performed Mendelian randomization (MR) analysis to determine the causal relationship of serum 25(OH)D with the risk of CES. Methods and resultsThe summary statistics dataset on the genetic variants related to 25(OH)D was used from the published GWAS of European descent participants in the UK Biobank, including 417,580 subjects, yielding 143 independent loci in 112 1-Mb regions. GWAS summary data of CES was obtained from GIGASTROKE Consortium, which included European individuals (10,804 cases, 1,234,808 controls). Our results unveiled a causal relationship between 25(OH)D and CES using IVW [OR = 0.82, 95% CI: 0.67–0.98, p = 0.037]. Horizontal pleiotropy was not seen [MR-Egger intercept = 0.001; p = 0.792], suggesting an absence of horizontal pleiotropy. Cochrane's Q [Q = 78.71, p-value = 0.924], Rucker's Q [Q = 78.64, p-value = 0.913], and I2 = 0.0% (95% CI: 0.0%, 24.6%) statistic suggested no heterogeneity. This result remained consistent using different MR methods and sensitivity analyses, including Maximum likelihood [OR = 0.82, 95%CI: 0.67–0.98, p-value = 0.036], Constrained maximum likelihood [OR = 0.76, 95%CI: 0.64–0.90, p-value = 0.002], Debiased inverse-variance weighted [OR = 0.82, 95%CI: 0.68–0.99, p-value = 0.002], MR-PRESSO [OR = 0.82, 95%CI 0.77–0.87, p-value = 0.022], RAPS [OR = 0.82, 95%CI 0.67–0.98, p-value = 0.038], MR-Lasso [OR = 0.82, 95%CI 0.68–0.99, p-value = 0.037]. ConclusionOur MR analysis provides suggestive evidence that increased 25(OH)D levels may play a protective role in the development of cardioembolic stroke. Determining the role of 25(OH)D in stroke subtypes has important clinical and public health implications.

Inflammatory cytokines and their potential role in Sjogren’s syndrome risk: insights from a mendelian randomization study

AimThis study aimed to investigate the causal impact of inflammatory cytokines on Sjogren’s Syndrome (SS) and to identify potential biomarkers for SS clinical management using Mendelian Randomization (MR).Materials and methodsLeveraging GWAS summary data of inflammatory cytokines and SS, we executed the first two-sample MR analysis. Genetic variants from prior GWASs associated with circulating inflammatory cytokines served as instrumental variables (IVs). Data regarding cytokines were analyzed using the Olink Target-96 Inflammation panel, synthesizing data from 14,824 participants. GWAS summary statistics for SS were procured from the UK Biobank, focusing on samples of European ancestry. To discern the causal relationship between inflammatory cytokines and SS, several MR methodologies, including inverse variance weighted (IVW) and MR-Egger regression, were applied.ResultsAfter rigorous IV quality control, 91 cytokines were incorporated into the MR analysis. The IVW analysis identified 8 cytokines with a positive association to SS: Axin-1 (OR 2.56, 95% CI 1.07–6.10), T-cell surface glycoprotein CD5 (OR 1.81, 95% CI 1.08–3.02), CUDP1 (OR 1.61, 95% CI 1.00-2.58), CXCL10 (OR 1.92, 95% CI 1.25–2.95), IL-4 (OR 2.18, 95% CI 1.22–3.91), IL-7 (OR 2.35, 95% CI 1.27–4.33), MCP-2 (OR 1.27, 95% CI 1.05–1.54), and TNFRSF9 (OR 1.83, 95% CI 1.03–3.24), suggesting their potential in increasing SS risk.ConclusionOur study conducted through MR, identified various inflammatory cytokines associated with SS risk, validating some previous research results and offering some new potential biomarkers for SS. However, these findings necessitate further research for validation and exploration of their precise role in the onset and progression of SS.

Association between Gut Microbiota and Biological Aging: A Two-Sample Mendelian Randomization Study.

Recent observational studies revealed an association between gut microbiota and aging, but whether gut microbiota are causally associated with the aging process remains unknown. We used a two-sample Mendelian randomization approach to investigate the causal association between gut microbiota and biological age acceleration using the largest available gut microbiota GWAS summary data from the MiBioGen consortium and GWAS data on biological age acceleration. We further conducted sensitivity analysis using MR-PRESSO, MR-Egger regression, Cochran Q test, and reverse MR analysis. Streptococcus (IVW, β = 0.16, p = 0.0001) was causally associated with Bioage acceleration. Eubacterium (rectale group) (IVW, β = 0.20, p = 0.0190), Sellimonas (IVW, β = 0.06, p = 0.019), and Lachnospira (IVW, β = -0.18, p = 0.01) were suggestive of causal associations with Bioage acceleration, with the latter being protective. Actinomyces (IVW, β = 0.26, p = 0.0083), Butyricimonas (IVW, β = 0.21, p = 0.0184), and Lachnospiraceae (FCS020 group) (IVW, β = 0.24, p = 0.0194) were suggestive of causal associations with Phenoage acceleration. This Mendelian randomization study found that Streptococcus was causally associated with Bioage acceleration. Further randomized controlled trials are needed to investigate its role in the aging process.

Separating the effects of life course adiposity on diabetic nephropathy: a comprehensive multivariable Mendelian randomization study.

Previous Mendelian randomization (MR) of obesity and diabetic nephropathy (DN) risk used small sample sizes or focused on a single adiposity metric. We explored the independent causal connection between obesity-related factors and DN risk using the most extensive GWAS summary data available, considering the distribution of adiposity across childhood and adulthood. To evaluate the overall effect of each obesity-related exposure on DN (Ncase = 3,676, Ncontrol = 283,456), a two-sample univariate MR (UVMR) analysis was performed. The independent causal influence of each obesity-related feature on DN was estimated using multivariable MR (MVMR) when accounting for confounding variables. It was also used to examine the independent effects of adult and pediatric obesity, adjusting for their interrelationships. We used data from genome-wide association studies, including overall general (body mass index, BMI) and abdominal obesity (waist-to-hip ratio with and without adjustment for BMI, i.e., WHR and WHRadjBMI), along with childhood obesity (childhood BMI). UVMR revealed a significant association between adult BMI (OR=1.24, 95%CI=1.03-1.49, P=2.06×10-2) and pediatric BMI (OR=1.97, 95%CI=1.59-2.45, P=8.55×10-10) with DN risk. At the same time, adult WHR showed a marginally significant increase in DN (OR =1.27, 95%CI = 1.01-1.60, P=3.80×10-2). However, the outcomes were adverse when the influence of BMI was taken out of the WHR (WHRadjBMI). After adjusting for childhood BMI, the causal effects of adult BMI and adult abdominal obesity (WHR) on DN were significantly attenuated and became nonsignificant in MVMR models. In contrast, childhood BMI had a constant and robust independent effect on DN risk(adjusted for adult BMI: IVW, OR=1.90, 95% CI=1.60-2.25, P=2.03×10-13; LASSO, OR=1.91, 95% CI=1.65-2.21, P=3.80×10-18; adjusted for adult WHR: IVW, OR=1.80, 95% CI=1.40-2.31, P=4.20×10-6; LASSO, OR=1.90, 95% CI=1.56-2.32, P=2.76×10-10). Our comprehensive analysis illustrated the hazard effect of obesity-related exposures for DN. In addition, we showed that childhood obesity plays a separate function in influencing the risk of DN and that the adverse effects of adult obesity (adult BMI and adult WHR) can be substantially attributed to it. Thus, several obesity-related traits deserve more attention and may become a new target for the prevention and treatment of DN and warrant further clinical investigation, especially in childhood obesity.

IFN-γ, SCF, MIP1b and IL-16 Were Associated with Risk of Diabetic Nephropathy: A Mendelian Randomization Study.

The impact of inflammatory factors on the risk of diabetic nephropathy (DN) is inconsistent. Two-sample Mendelian randomization (MR) analyses were used to detect the causal role of inflammatory factors in DN risk. Inflammatory factor GWAS summary data were collected from a meta-analysis including 8,293 Finnish participants, and DN information was extracted from a GWAS of 213,746 individuals from FinnGen. The MR Pleiotropy Residual Sum and Outlier (MR-PRESSO) outlier test was used for the removal of horizontal pleiotropic outliers. Multivariable MR analysis was also used to adjust for pleiotropy. IFN-γ [ORIVW: 1.33; 95% CI: 1.09-1.63; p=0.005] and SCF [ORIVW: 1.25, 1.02-1.52; p = 0.027] were associated with an increased risk of DN. MIP1b [ORIVW: 0.92; 95% CI: 0.85-0.98; p = 0.022] and IL-16 [ORIVW: 0.89, 0.81-0.99; p = 0.043] showed negative associations with the risk of DN. We validated our MR results with MR-PRESSO analyses. Significant horizontal pleiotropy was not found. Moreover, in the multivariable MR analysis, the associations between cytokines and DN risk remained. Our MR results based on genetic data contribute to a better understanding of the pathogenesis of DN and provide evidence for a causal effect of inflammatory factors on DN. These findings support targeting specific inflammatory factors to alleviate DN risk.

Networks of gut bacteria relate to cardiovascular disease in a multi-ethnic population: the HELIUS study.

Gut microbiota have been linked to blood lipid levels and cardiovascular diseases (CVDs). The composition and abundance of gut microbiota trophic networks differ between ethnicities. We aim to evaluate the relationship between gut microbiotal trophic networks and CVD phenotypes. We included cross-sectional data from 3860 individuals without CVD history from 6 ethnicities living in the Amsterdam region participating in the prospective Healthy Life in Urban Setting (HELIUS) study. Genetic variants were genotyped, faecal gut microbiota were profiled, and blood and anthropometric parameters were measured. A machine learning approach was used to assess the relationship between CVD risk (Framingham score) and gut microbiota stratified by ethnicity. Potential causal relationships between gut microbiota composition and CVD were inferred by performing two-sample Mendelian randomization with hard CVD events from the Pan-UK Biobank and microbiome genome-wide association studies summary data from a subset of the HELIUS cohort (n = 4117). Microbial taxa identified to be associated with CVD by machine learning and Mendelian randomization were often ethnic-specific, but some concordance across ethnicities was found. The microbes Akkermansia muciniphila and Ruminococcaceae UCG-002 were protective against ischaemic heart disease in African-Surinamese and Moroccans, respectively. We identified a strong inverse association between blood lipids, CVD risk, and the combined abundance of the correlated microbes Christensenellaceae-Methanobrevibacter-Ruminococcaceae (CMR). The CMR cluster was also identified in two independent cohorts and the association with triglycerides was replicated. Certain gut microbes can have a potentially causal relationship with CVD events, with possible ethnic-specific effects. We identified a trophic network centred around Christensenellaceae, Methanobrevibacter, and various Ruminococcaceae, frequently lacking in South-Asian Surinamese, to be protective against CVD risk and associated with low triglyceride levels.

From teeth to brain: dental caries causally affects the cortical thickness of the banks of the superior temporal sulcus

ObjectivesDental caries is one of the most prevalent oral diseases and causes of tooth loss. Cross-sectional studies observed epidemiological associations between dental caries and brain degeneration disorders, while it is unknown whether dental caries causally affect the cerebral structures. This study tested whether genetically proxied DMFS (the sum of Decayed, Missing, and Filled tooth Surfaces) causally impacts the brain cortical structure using Mendelian randomization (MR).MethodsThe summary-level GWAS meta-analysis data from the GLIDE consortium were used for DMFS, including 26,792 participants. ENIGMA (Enhancing NeuroImaging Genetics through Meta Analysis) consortium GWAS summary data of 51,665 patients were used for brain structure. This study estimated the causal effects of DMFS on the surface area (SA) and thickness (TH) of the global cortex and functional cortical regions accessed by magnetic resonance imaging (MRI). Inverse-variance weighted (IVW) was used as the primary estimate, the MR pleiotropy residual sum and outlier (MR-PRESSO), the MR-Egger intercept test, and leave-one-out analyses were used to examine the potential horizontal pleiotropy.ResultsGenetically proxied DMFS decreases the TH of the banks of the superior temporal sulcus (BANSSTS) with or without global weighted (weighted, β = − 0.0277 mm, 95% CI: − 0.0470 mm to − 0.0085 mm, P = 0.0047; unweighted, β = − 0.0311 mm, 95% CI: − 0.0609 mm to − 0.0012 mm, P = 0.0412). The causal associations were robust in various sensitivity analyses.ConclusionsDental caries causally decrease the cerebral cortical thickness of the BANKSSTS, a cerebral cortical region crucial for language-related functions, and is the most affected brain region in Alzheimer’s disease. This investigation provides the first evidence that dental caries causally affects brain structure, proving the existence of teeth-brain axes. This study also suggested that clinicians should highlight the causal effects of dental caries on brain disorders during the diagnosis and treatments, the cortical thickness of BANKSSTS is a promising diagnostic measurement for dental caries-related brain degeneration.

GWAS meta-analysis of kidney function traits in Japanese populations.

Genetic epidemiological evidence for the kidney function traits in East Asian population including Japanese remain still relatively unclarified. Especially, the number of GWASs for kidney traits reported still remains limited, and the sample size of each independent study is relatively small. Given the genetic variability between ancestries/ethnicities, implementation of GWAS with sufficiently large sample sizes in specific population of Japanese is considered meaningful. We conducted the GWAS meta-analyses of kidney traits by leveraging the GWAS summary data of the representative large genome cohort studies with about 200,000 Japanese participants (n = 202,406 for estimated glomerular filtration rate [eGFR] and n = 200,845 for serum creatinine [SCr]). In the present GWAS meta-analysis, we identified 110 loci with 169 variants significantly associated with eGFR (on chromosomes 1-13 and 15-22; p < 5×10-8), whereas we also identified 112 loci with 176 variants significantly associated with SCr (on chromosomes 1-22; p < 5×10-8), of which one locus (more than 1Mb distant from known loci) with one variant (CD36 rs146148222 on chromosome 7) for SCr was considered as the truly novel finding. The present GWAS meta-analysis of largest genome cohort studies in Japanese provided some original genomic loci associated with kidney function in Japanese, which may contribute to the possible development of personalized prevention of kidney diseases based on genomic information in the near future.

Investigating potential novel therapeutic targets and biomarkers for ankylosing spondylitis using plasma protein screening.

Ankylosing spondylitis (AS) is a chronic inflammatory disease affecting the spine and sacroiliac joints. Recent genetic studies suggest certain plasma proteins may play a causal role in AS development. This study aims to identify and characterize these proteins using Mendelian randomization (MR) and colocalization analyses. Plasma protein data were obtained from recent publications in Nature Genetics, integrating data from five previous GWAS datasets, including 738 cis-pQTLs for 734 plasma proteins. GWAS summary data for AS were sourced from IGAS and other European cohorts. MR analyses were conducted using "TwoSampleMR" to assess causal links between plasma protein levels and AS. Colocalization analysis was performed with the coloc R package to identify shared genetic variants. Sensitivity analyses and protein-protein interaction (PPI) network analyses were conducted to validate findings and explore therapeutic targets. We performed Phenome-wide association study (PheWAS) to examine the potential side effects of drug protein on AS treatment. After FDR correction, eight significant proteins were identified: IL7R, TYMP, IL12B, CCL8, TNFAIP6, IL18R1, IL23R, and ERAP1. Elevated levels of IL7R, IL12B, CCL8, IL18R1, IL23R, and ERAP1 increased AS risk, whereas elevated TYMP and TNFAIP6 levels decreased AS risk. Colocalization analysis indicated that IL23R, IL7R, and TYMP likely share causal variants with AS. PPI network analysis identified IL23R and IL7R as potential new therapeutic targets. This study identified eight plasma proteins with significant associations with AS risk, suggesting IL23R, IL7R, and TYMP as promising therapeutic targets. Further research is needed to explore underlying mechanisms and potential for drug repurposing.

The cause-and-effect relationship between gut microbiota abundance and carcinoid syndrome: a bidirectional Mendelian randomization study.

Carcinoid syndrome (CS) commonly results from neuroendocrine tumors. While active substances are recognized as the main causes of the typical symptoms such as diarrhea and skin flush, the cause-and-effect relationship between gut microbiota abundance and CS remains unclear. The Single Nucleotide Polymorphisms (SNPs) related to gut microbiota abundance and CS were obtained from the GWAS summary data. The inverse variance weighted (IVW) method was used to assess the causal relationship between gut microbiota abundance and CS. Additionally, the MR-Egger, Weighted Median model, and Weighted model were employed as supplementary approaches. The heterogeneity function of the TwoSampleMR package was utilized to assess whether SNPs exhibit heterogeneity. The Egger intercept and Presso test were used to assess whether SNPs exhibit pleiotropy. The Leave-One-Out test was employed to evaluate the sensitivity of SNPs. The Steiger test was utilized to examine whether SNPs have a reverse causal relationship. A bidirectional mendelian randomization (MR) study was conducted to elucidate the inferred cause-and-effect relationship between gut microbiota abundance and CS. The IVW results indicated a causal relationship between 6 gut microbiota taxa and CS. Among the 6 gut microbiota taxa, the genus Anaerofilum (IVW OR: 0.3606, 95%CI: 0.1554-0.8367, p-value: 0.0175) exhibited a protective effect against CS. On the other hand, the family Coriobacteriaceae (IVW OR: 3.4572, 95%CI: 1.0571-11.3066, p-value: 0.0402), the genus Enterorhabdus (IVW OR: 4.2496, 95%CI: 1.3314-13.5640, p-value: 0.0146), the genus Ruminiclostridium6 (IVW OR: 4.0116, 95%CI: 1.2711-12.6604, p-value: 0.0178), the genus Veillonella (IVW OR: 3.7023, 95%CI: 1.0155-13.4980, p-value: 0.0473) and genus Holdemanella (IVW OR: 2.2400, 95%CI: 1.0376-4.8358, p-value: 0.0400) demonstrated a detrimental effect on CS. The CS was not found to have a reverse causal relationship with the above 6 gut microbiota taxa. Six microbiota taxa were found to have a causal relationship with CS, and further randomized controlled trials are needed for verification.

Novel insights into the complex interplay of immune dysregulation and inflammatory biomarkers in preeclampsia and fetal growth restriction: A two-step Mendelian randomization analysis

BackgroundThe relationship between genetic immune dysregulation and the occurrence of preeclampsia (PE) or PE with fetal growth restriction (PE with FGR) has yielded inconsistent findings, and the underlying mediators of this association remain elusive. We aimed to explore the causal impact of genetic immune dysregulation on the risk of PE or PE with FGR and to elucidate the role of specific transcriptomes in mediating this relationship. MethodsA two-step Mendelian randomization (MR) analysis was performed to explore the link between immune dysregulation and PE or PE with FGR, as well as to identify potential inflammatory biomarkers that act as mediators. GWAS summary data for outcomes were obtained from the FinnGen dataset. The analyses encompassed five systemic immune-associated diseases, four chronic genital inflammatory diseases, and thirty-one inflammatory biomarkers. Summary-data-based MR (SMR) and HEIDI analysis were conducted to test whether the effect size of single nucleotide polymorphisms (SNPs) on outcomes was mediated by the expression of immune-associated genes. ResultsThe primary univariable analysis revealed a significant positive correlation between systemic lupus erythematosus (SLE), type 1 diabetes (T1D), type 2 diabetes (T2D), and rheumatoid arthritis (RA) with the risk of PE or PE with FGR. Surprisingly, a counterintuitive finding showed a significant negative association between endometriosis of pelvic peritoneum (EMoP) and the risk of PE with FGR. None of the inflammatory factors had a causal relationship with PE or PE with FGR. However, there was a significant association between lymphocyte count and the risk of PE with FGR. Within the lymphocyte subset, both the proportion of Natural Killer (NK) cells and absolute counts of naïve CD4+ T cells demonstrated significant effects on the risk of PE with FGR. Two-step MR analysis underscored the genetically predicted lymphocyte count as a significant mediator between T1D and PE with FGR. Additionally, SMR analysis indicated the potential involvement of SH2B3 in the occurrence of PE with FGR. ConclusionsOur findings provided substantial evidence of the underlying causal relationship between immune dysregulation and PE or PE with FGR and some of these diseases proved to accelerate immune cells disorders and then contribute to the risk of incident PE or PE with FGR.

Specific gut microbiota may increase the risk of erectile dysfunction: a two-sample Mendelian randomization study.

Studies have found that gut microbiota may be associated with the development of erectile dysfunction (ED); however, the exact link between the two remains unclear. This study aimed to elucidate the relationship between the gut microbiota and the risk of ED from a genetic perspective. We investigated the relationship between the gut microflora and ED using two-sample Mendelian randomization. GWAS-pooled data for ED were obtained from 223805 participants in Europe. GWAS summary data for ED were obtained from 223805 subjects in Europe and that for the gut microbiota were obtained from 18340 participants in 24 cohorts. We used the inverse-variance weighted (IVW) estimator as the primary method for the preliminary analysis, and the MR-Egger, weighted median (WM), simple model, and weighted model as secondary methods. We used Cochrane's Q-test, to detect heterogeneity, MREgger to detect pleiotropy, and the leave-one-out method to test the stability of the MR results. Ultimately, we genetically predicted a causal relationship between 211 gut microbiota and ED. A total of 2818 SNPs associated with gut microflora were screened in the ED correlation analysis based on the assumption of instrumental variables. The results of MR analysis showed a causal relationship between the six gut microbes and ED occurrence. The results of the fixed effects IVW method revealed five gut microflora, including Lachnospiraceae (OR, 1.265; P = 0.008), Lachnospiraceae NC2004 group (OR, 1.188; P = 0.019), Oscillibacter (OR, 1.200; P = 0.015), Senegalimassilia (OR, 1.355; P = 0.002), Tyzzerella3 (OR, 1.133; P = 0.022), to be negatively associated with ED. In addition, the IVW method revealed Ruminococcaceae UCG-013 (OR, 0.827; P = 0.049) to be positively associated with ED. Quality control results showed no heterogeneity or horizontal pleiotropy in the MR analysis (P > 0.05). Six gut microbes were genetically associated with ED; of which, Ruminococcaceae UCG-013 was causally associated with a reduced risk of ED development. Our findings provide a new direction for research on the prevention and treatment of ED; however, the mechanisms and details require further investigation.

Statistical inference with large-scale trait imputation.

Recently a nonparametric method called LS-imputation has been proposed for large-scale trait imputation based on a GWAS summary dataset and a large set of genotyped individuals. The imputed trait values, along with the genotypes, can be treated as an individual-level dataset for downstream genetic analyses, including those that cannot be done with GWAS summary data. However, since the covariance matrix of the imputed trait values is often too large to calculate, the current method imposes a working assumption that the imputed trait values are identically and independently distributed, which is incorrect in truth. Here we propose a "divide and conquer/combine" strategy to estimate and account for the covariance matrix of the imputed trait values via batches, thus relaxing the incorrect working assumption. Applications of the methods to the UK Biobank data for marginal association analysis showed some improvement by the new method in some cases, but overall the original method performed well, which was explained by nearly constant variances of and mostly weak correlations among imputed trait values.

Mendelian Randomization Identifies Genetically Supported Drug Targets for Amyotrophic Lateral Sclerosis and Frontotemporal Dementia.

Currently, amyotrophic lateral sclerosis (ALS) and frontotemporal dementia (FTD) have no effective treatments. Drug repurposing offers a rapid method to meet therapeutic need for ALS and FTD. To identify therapeutic targets associated with ALS and FTD, Mendelian randomization (MR) analysis and colocalization were performed. Genetic instruments were based on transcriptomic and proteomic data for 422 actionable proteins targeted by approved drugs or clinical drug candidates. The publicly available ALS GWAS summary data (including a total of 20,806 ALS cases and 59,804 controls) and FTD GWAS summary data (including a total of 2154 patients with FTD and 4308 controls) were used. Using cis-expression quantitative trait loci and cis-protein quantitative trait loci genetic instruments, we identified several drug targets for repurposing (ALS: MARK3, false-discovery rate (FDR) = 0.043; LTBR, FDR = 0.068) (FTD: HLA-DRB1, FDR = 0.083; ADH5, FDR = 0.056). Our MR study analyzed the actionable druggable proteins and provided potential therapeutic targets for ALS and FTD. Future studies should further elucidate the underlying mechanism of corresponding drug targets in the pathogenesis of ALS and FTD.

Predicting the impact of low-density lipoprotein cholesterol lowering through different drug targets on risk of congenital heart disease

Abstract Background Low-density lipoprotein cholesterol (LDL) lowering in pregnancy has been a controversial topic as observational studies have described an increased risk of congenital malformations. With the advent of new, long-lasting LDL lowering therapies such as silencing RNA therapies that exert clinical effects for up to six months after administration, there is substantial probability that women of childbearing age might conceive during the period of drug activity. Where clinical data is not available, the Mendelian randomization (MR) paradigm can be used to predict effects of drug target perturbation using genetic data. Purpose We aimed to leverage genetic data to explore the predicted impact of LDL lowering via the statin, ezetimibe and PCSK9 drug targets, on congenital heart disease. Methods Uncorrelated genome-wide significant (p&amp;lt;5x10-8) variants for LDL and their association estimates, overall and restricted to the HMGCR, NPC1L1, PCSK9 gene regions +/- 10kB, were extracted from GWAS summary data on the UK Biobank (n=469,897). Gene-outcome association data was extracted from GWAS results of FinnGen 8thRelease (n=342,499) for conotruncal defects, congenital malformations of the great arteries; septal defects; congenital disease of the aortic or mitral valves; and the vertebral, anorectal, cardiovascular, tracheo-esophageal, renal and limb anomaly (VACTERL) association. A drug-target MR analysis was carried out using inverse-variance weighted MR. Results Genetically-proxied LDL-lowering through PCSK9 associated with higher odds of VACTREL association (OR 1.70, 95%ci 1.25-2.32; p&amp;lt;0.001), conotruncal defects (OR 3.97, 95%CI 1.37-11.53, p=0.011) and overall congenital heart disease of any type (OR 1.49, 95%CII 1.08-2.05, p=0.015). Genetically-proxied LDL lowering via HMGCR was associated with higher odds of congenital malformations involving the great arteries (OR 5.85, 95%cCI 1.07-31.84, p=0.041). Genetically-proxied LDL lowering via NPC1L1 drug targets was not associated with differences in risk of congenital heart disease, though the power of these analyses was limited. Conclusions These data provide genetic evidence suggesting that use of LDL-lowering agents during pregnancy might influence risk of congenital heart disease. Importantly, due to underlying assumptions of the analysis, the results of this analysis are only applicable to drugs that are able to cross the placenta and exert direct effects on LDL metabolism in the fetus and they cannot be used to infer the indirect influence of lowering maternal LDL. The data support ongoing caution regarding statin use, and additionally suggest that long-lasting PCSK9-inhibiting agents that are able to cross the placenta, such as monoclonal antibodies and silencing RNA therapies, should be used with utmost caution in women planning to conceive.Figure 1

BMI and plasma lipid levels with risk of proliferative diabetic retinopathy: a univariable and multivariable Mendelian randomization study.

The study aimed to determine whether a causal effect exists between body mass index (BMI) or plasma lipid levels and proliferative diabetic retinopathy (PDR) risk in humans. We utilized univariable (UVMR) and multivariable two-sample Mendelian randomization (MVMR) analyses to confirm the effects of BMI and plasma lipid levels on the risk of PDR. Genetic variants associated with BMI and three plasma lipids were obtained from GWAS summary datasets generated by many different consortia and were deposited in the MR-Base database. The GWAS summary data for PDR from the FinnGen biobank included 2,12,889 participants of European ancestry (8,681 cases and 2,04,208 controls). Inverse variance weighted (IVW) was applied as the main MR analysis. Sensitivity analysis was used to evaluate the robustness of our findings. In the UVMR analysis, the causal associations of genetically predicted BMI with PDR presented a positive association (OR = 1.120, 95% CI = 1.076-1.167, P < 0.001), and the lower HDL-C level was associated with a higher risk of PDR (OR = 0.898, 95% CI = 0.811-0.995, P = 0.040). No evidence of an association between LDL-C or TG levels (P > 0.05) and PDR risk was found. In the MVMR analysis controlling for the HDL-C level, there was strong evidence for a direct causal effect of BMI on the risk of PDR (OR = 1.106, 95%CI = 1.049, 1.166, P < 0.001, IVW). After adjusting for BMI, there was no evidence for a direct causal effect of the HDL-C level on the risk of PDR (OR = 0.911, 95% CI = 0.823, 1.008, P = 0.072). Sensitivity analyses confirmed that the results were reliable and stable. Robust evidence was demonstrated for an independent, causal effect of BMI in increasing the risk of PDR. Further studies are required to understand the potential biological mechanisms underlying this causal relationship.

Transcriptome-wide association study identifies new susceptibility genes and pathways for spondyloarthritis

BackgroundSpondyloarthritis (SpA) is a group of multifactorial bone diseases influenced by genetic factors, the environment and lifestyle. However, current studies have found a limited number of SpA-related genes, and the genetic and pathogenic mechanisms of SpA are still unclear.MethodsA tissue-specific transcriptome-wide association study (TWAS) of SpA was performed using GWAS (including 3966 SpA patients and 448,298 controls) summary data and gene expression weights of whole blood and skeletal muscle. The SpA-associated genes identified by TWAS were further compared with the differentially expressed genes (DEGs) identified in the SpA gene expression profile acquired from the Gene Expression Omnibus database (GEO, GSE58667). Finally, functional enrichment and annotation analyses of the identified genes were performed.ResultsThe TWAS detected 499 suggestive genes associated with SpA in whole blood and skeletal muscle, such as CTNNAL1 (PSM = 3.04 × 10−2, PWB = 9.58 × 10−3). The gene expression profile of SpA identified 20 candidate genes that overlapped in the TWAS data, such as MCM4 (PTWAS = 1.32 × 10−2, PDEG = 2.75 × 10−2) and KIAA1109 (PTWAS = 3.71 × 10−2, PDEG = 4.67 × 10−2). Enrichment analysis of the genes identified by TWAS identified 93 significant GO terms and 33 KEGG pathways, such as mitochondrion organization (GO: 0007005) and axon guidance (hsa04360).ConclusionWe identified multiple candidate genes that were genetically related to SpA. Our study may provide novel clues regarding the genetic mechanism, diagnosis, and treatment of SpA.

The causal effect of bioavailable testosterone on primary biliary cholangitis in female patients: A Bidirectional Mendelian randomization analysis.

Primary biliary cholangitis (PBC) primarily affects female in their 4th to 6th decade of life at which stage female's sex hormones change dramatically. Sex hormones have a wide range of effects on the liver and immune system. However, it remains unclear whether sex hormonal changes mediate the onset of PBC. This study investigated the causal effect between total testosterone (TT), bioavailable testosterone (BAT), sex hormone binding globulin (SHBG) in female and PBC using public GWAS summary data. Data on TT, BAT, SHBG in female were obtained from a previous study based on the UK Biobank. PBC GWAS summary data was obtained from a genome-wide meta-analysis. We used several methods to make the conclusion robust. Various sensitivity analyses had been conducted to assess the consistency of our findings. Our Mendelian randomization analysis revealed that the genetically predicted BAT in female was positively associated with PBC and SHBG in female was negatively associated with PBC. The results obtained from different methods are similar, which proves the reliability of the results. Our study supported a causal relationship of BAT and SHBG in female on PBC. To confirm the association between testosterone and PBC, more research should be conducted.

P-477 Bidirectional association between ovarian reserve and spontaneous miscarriage: findings from the clinical IVF data to epidemiological evidence and genetic links

Abstract Study question To determine whether the association between ovarian reserve and spontaneous miscarriage risk is bidirectional and caused by shared genetic risk loci. Summary answer Bidirectional association between ovarian reserve and spontaneous miscarriage was observed. This may be attributed to five shared risk genes (SYCP2L, TP63, GSPT1, RIF1, and PRRC2A). What is known already Previous studies have shown that women with lower serum anti-Müllerian hormone (AMH) had an increased risk of embryo aneuploidy or spontaneous miscarriage risk, compared to women with higher AMH concentration, although not all studies agree. The most recent meta-analysis postulated that the association between diminished ovarian reserve (DOR) and miscarriage may be not causal, but that the two occurrences merely share a common cause, such as other systematic clinical conditions or past exposure. Inspired by this hypothesis, we conducted the present analyses to clarify this bidirectional association between DOR and miscarriage and further elucidate plausible mechanisms by using genetic data. Study design, size, duration There were four data sources. 1) The Clinical IVF data were retrospectively collected from an academically affiliated Reproductive Medicine Center (17,786 cycles included). 2) We further analyzed data from the UK Biobank (UKB) which is a large-scale, population-based, prospective cohort study (35,316 white women included). 3) The GWAS summary statistics for age at natural menopause (ANM) and spontaneous miscarriage were obtained from published GWAS studies. 4) The individual-level genotype data were also extracted from UKB. Participants/materials, setting, methods There were four analysis modules, 1) IVF data to test the association of ovarian reserve and miscarriage risk by using logistic regression models, 2) UKB data to test the association of spontaneous miscarriage history and ANM by using linear regression models, 3) GWAS summary data to test the overall genetic enrichment between miscarriage and ANM by plotting Q-Q plot, and 4) individual-level genotype data to identify specific shared genes by using ‘pleio’ R package. Main results and the role of chance In the analysis of clinical data, the risk of miscarriage was 1.5 (95% CI: 1.22-1.85, p &amp;lt; 0.001) times in the group with AMH&amp;lt;1.1 ng/ml, compared to the group with AMH &amp;gt; =1.1 ng/ml after adjustment. In the analysis of UK Biobank data, one more spontaneous miscarriage history would lead to 0.02 (95% CI: 0.00-0.03, p = 0.032) years earlier in the standardized ANM after adjustment. Participants with &amp;gt; =2 spontaneous miscarriage history would have 0.05 (95% CI:0.01-0.08, p = 0.011) years earlier in standardized ANM, compared with participants with &amp;lt; =1 spontaneous miscarriage. In the analysis of GWAS summary statistics, we found that there were 11,736,532 single nucleotide polymorphisms (SNPs) tested by both GWAS studies of ANM and miscarriage. In the stratified conditional Q-Q plot, successive leftward deflections did not show which suggested that an overall genetic overlap between miscarriage and ANM was not observed. In the analysis of the individual-level genotype data, we identified five shared genetic risk loci that affect both miscarriage and menopause. They were all novel risk genes for spontaneous miscarriage. Of these five genes, SYCP2L, TP63, GSPT1, and RIF1 were associated with congenital malformations and PRRC2A was associated with autoimmune disease, which are factors that possibly play a role in the occurrence of miscarriage. Limitations, reasons for caution The UKB epidemiology data and genetic data were restricted to participants of European ancestry. Further studies are needed in non-white populations. Second, the predictive value of GWAS summary data for miscarriage may be limited which results in a null finding in the conditional Q-Q plot. Wider implications of the findings Our findings could be of interest to IVF clinicians in patient counseling regarding the prognosis of IVF treatment and genetic counseling for miscarriage. Our findings encourage further research focus on the shared genetic architecture and common pathophysiological basis of diminished ovarian reserve and miscarriage which may bring new therapeutic opportunities. Trial registration number not applicable