Utilizing multimodal AI to improve genetic analyses of cardiovascular traits.

Electronic health records, biobanks, and wearable biosensors contain multiple high-dimensional clinical data (HDCD) modalities (e.g., ECG, Photoplethysmography (PPG), and MRI) for each individual. Access to multimodal HDCD provides a unique opportunity for genetic studies of complex traits because different modalities relevant to a single physiological system (e.g., circulatory system) encode complementary and overlapping information. We propose a novel multimodal deep learning method, M-REGLE, for discovering genetic associations from a joint representation of multiple complementary HDCD modalities. We showcase the effectiveness of this model by applying it to several cardiovascular modalities. M-REGLE jointly learns a lower representation (i.e., latent factors) of multimodal HDCD using a convolutional variational autoencoder, performs genome wide association studies (GWAS) on each latent factor, then combines the results to study the genetics of the underlying system. To validate the advantages of M-REGLE and multimodal learning, we apply it to common cardiovascular modalities (PPG and ECG), and compare its results to unimodal learning methods in which representations are learned from each data modality separately, but the downstream genetic analyses are performed on the combined unimodal representations. M-REGLE identifies 19.3% more loci on the 12-lead ECG dataset, 13.0% more loci on the ECG lead I + PPG dataset, and its genetic risk score significantly outperforms the unimodal risk score at predicting cardiac phenotypes, such as atrial fibrillation (Afib), in multiple biobanks.

NMR metabolomic modeling of age and lifespan: A multicohort analysis.

Metabolomic age models have been proposed for the study of biological aging, however, they have not been widely validated. We aimed to assess the performance of newly developed and existing nuclear magnetic resonance spectroscopy (NMR) metabolomic age models for prediction of chronological age (CA), mortality, and age-related disease. Ninety-eight metabolic variables were measured in blood from nine UK and Finnish cohort studies (N ≈31,000 individuals, age range 24-86 years). We used nonlinear and penalized regression to model CA and time to all-cause mortality. We examined associations of four new and two previously published metabolomic age models, with aging risk factors and phenotypes. Within the UK Biobank (N ≈102,000), we tested prediction of CA, incident disease (cardiovascular disease (CVD), type-2 diabetes mellitus, cancer, dementia, and chronic obstructive pulmonary disease), and all-cause mortality. Seven-fold cross-validated Pearson's r between metabolomic age models and CA ranged between 0.47 and 0.65 in the training cohort set (mean absolute error: 8-9 years). Metabolomic age models, adjusted for CA, were associated with C-reactive protein, and inversely associated with glomerular filtration rate. Positively associated risk factors included obesity, diabetes, smoking, and physical inactivity. In UK Biobank, correlations of metabolomic age with CA were modest (r = 0.29-0.33), yet all metabolomic model scores predicted mortality (hazard ratios of 1.01 to 1.06/metabolomic age year) and CVD, after adjustment for CA. While metabolomic age models were only moderately associated with CA in an independent population, they provided additional prediction of morbidity and mortality over CA itself, suggesting their wider applicability.

NMR metabolomic modelling of age and lifespan: a multi-cohort analysis.

Metabolomic age models have been proposed for the study of biological aging, however they have not been widely validated. We aimed to assess the performance of newly developed and existing nuclear magnetic resonance spectroscopy (NMR) metabolomic age models for prediction of chronological age (CA), mortality, and age-related disease. 98 metabolic variables were measured in blood from nine UK and Finnish cohort studies (N ≈ 31,000 individuals, age range 24-86 years). We used non-linear and penalised regression to model CA and time to all-cause mortality. We examined associations of four new and two previously published metabolomic age models, with ageing risk factors and phenotypes. Within the UK Biobank (N≈ 102,000), we tested prediction of CA, incident disease (cardiovascular disease (CVD), type-2 diabetes mellitus, cancer, dementia, chronic obstructive pulmonary disease) and all-cause mortality. Cross-validated Pearson's r between metabolomic age models and CA ranged between 0.47-0.65 in the training set (mean absolute error: 8-9 years). Metabolomic age models, adjusted for CA, were associated with C-reactive protein, and inversely associated with glomerular filtration rate. Positively associated risk factors included obesity, diabetes, smoking, and physical inactivity. In UK Biobank, correlations of metabolomic age with chronological age were modest (r = 0.29-0.33), yet all metabolomic model scores predicted mortality (hazard ratios of 1.01 to 1.06 / metabolomic age year) and CVD, after adjustment for CA. While metabolomic age models were only moderately associated with CA in an independent population, they provided additional prediction of morbidity and mortality over CA itself, suggesting their wider applicability.

Triangulating evidence from observational and Mendelian randomization studies of ketone bodies for cognitive performance.

BACKGROUND: Ketone bodies (KBs) are an alternative energy supply for brain functions when glucose is limited. The most abundant ketone metabolite, 3-ß-hydroxybutyrate (BOHBUT), has been suggested to prevent or delay cognitive impairment, but the evidence remains unclear. We triangulated observational and Mendelian randomization (MR) studies to investigate the association and causation between KBs and cognitive function. METHODS: In observational analyses of 5506 participants aged ≥ 45 years from the Whitehall II study, we used multiple linear regression to investigate the associations between categorized KBs and cognitive function scores. Two-sample MR was carried out using summary statistics from an in-house KBs meta-analysis between the University College London-London School of Hygiene and Tropical Medicine-Edinburgh-Bristol (UCLEB) Consortium and Kettunen et al. (N = 45,031), and publicly available summary statistics of cognitive performance and Alzheimer's disease (AD) from the Social Science Genetic Association Consortium (N = 257,841), and the International Genomics of Alzheimer's Project (N = 54,162), respectively. Both strong (P < 5 × 10-8) and suggestive (P < 1 × 10-5) sets of instrumental variables for BOHBUT were applied. Finally, we performed cis-MR on OXCT1, a well-known gene for KB catabolism. RESULTS: BOHBUT was positively associated with general cognitive function (ß = 0.26, P = 9.74 × 10-3). In MR analyses, we observed a protective effect of BOHBUT on cognitive performance (inverse variance weighted: ßIVW = 7.89 × 10-2, PIVW = 1.03 × 10-2; weighted median: ßW-Median = 8.65 × 10-2, PW-Median = 9.60 × 10-3) and a protective effect on AD (ßIVW = - 0.31, odds ratio: OR = 0.74, PIVW = 3.06 × 10-2). Cis-MR showed little evidence of therapeutic modulation of OXCT1 on cognitive impairment. CONCLUSIONS: Triangulation of evidence suggests that BOHBUT has a beneficial effect on cognitive performance. Our findings raise the hypothesis that increased BOHBUT may improve general cognitive functions, delaying cognitive impairment and reducing the risk of AD.

Validation of lipid-related therapeutic targets for coronary heart disease prevention using human genetics.

Drug target Mendelian randomization (MR) studies use DNA sequence variants in or near a gene encoding a drug target, that alter the target's expression or function, as a tool to anticipate the effect of drug action on the same target. Here we apply MR to prioritize drug targets for their causal relevance for coronary heart disease (CHD). The targets are further prioritized using independent replication, co-localization, protein expression profiles and data from the British National Formulary and clinicaltrials.gov. Out of the 341 drug targets identified through their association with blood lipids (HDL-C, LDL-C and triglycerides), we robustly prioritize 30 targets that might elicit beneficial effects in the prevention or treatment of CHD, including NPC1L1 and PCSK9, the targets of drugs used in CHD prevention. We discuss how this approach can be generalized to other targets, disease biomarkers and endpoints to help prioritize and validate targets during the drug development process.

Genome-wide association study of circulating interleukin 6 levels identifies novel loci.

Interleukin 6 (IL-6) is a multifunctional cytokine with both pro- and anti-inflammatory properties with a heritability estimate of up to 61%. The circulating levels of IL-6 in blood have been associated with an increased risk of complex disease pathogenesis. We conducted a two-staged, discovery and replication meta genome-wide association study (GWAS) of circulating serum IL-6 levels comprising up to 67 428 (ndiscovery = 52 654 and nreplication = 14 774) individuals of European ancestry. The inverse variance fixed effects based discovery meta-analysis, followed by replication led to the identification of two independent loci, IL1F10/IL1RN rs6734238 on chromosome (Chr) 2q14, (Pcombined = 1.8 × 10-11), HLA-DRB1/DRB5 rs660895 on Chr6p21 (Pcombined = 1.5 × 10-10) in the combined meta-analyses of all samples. We also replicated the IL6R rs4537545 locus on Chr1q21 (Pcombined = 1.2 × 10-122). Our study identifies novel loci for circulating IL-6 levels uncovering new immunological and inflammatory pathways that may influence IL-6 pathobiology.

Establishing reference intervals for triglyceride-containing lipoprotein subfraction metabolites measured using nuclear magnetic resonance spectroscopy in a UK population.

BACKGROUND: Nuclear magnetic resonance (NMR) spectroscopy allows triglycerides to be subclassified into 14 different classes based on particle size and lipid content. We recently showed that these subfractions have differential associations with cardiovascular disease events. Here we report the distributions and define reference interval ranges for 14 triglyceride-containing lipoprotein subfraction metabolites. METHODS: Lipoprotein subfractions using the Nightingale NMR platform were measured in 9073 participants from four cohort studies contributing to the UCL-Edinburgh-Bristol consortium. The distribution of each metabolite was assessed, and reference interval ranges were calculated for a disease-free population, by sex and age group (<55, 55-65, >65 years), and in a subgroup population of participants with cardiovascular disease or type 2 diabetes. We also determined the distribution across body mass index and smoking status. RESULTS: The largest reference interval range was observed in the medium very-low density lipoprotein subclass (2.5th 97.5th percentile; 0.08 to 0.68 mmol/L). The reference intervals were comparable among male and female participants, with the exception of triglyceride in high-density lipoprotein. Triglyceride subfraction concentrations in very-low density lipoprotein, intermediate-density lipoprotein, low-density lipoprotein and high-density lipoprotein subclasses increased with increasing age and increasing body mass index. Triglyceride subfraction concentrations were significantly higher in ever smokers compared to never smokers, among those with clinical chemistry measured total triglyceride greater than 1.7 mmol/L, and in those with cardiovascular disease, and type 2 diabetes as compared to disease-free subjects. CONCLUSION: This is the first study to establish reference interval ranges for 14 triglyceride-containing lipoprotein subfractions in samples from the general population measured using the nuclear magnetic resonance platform. The utility of nuclear magnetic resonance lipid measures may lead to greater insights for the role of triglyceride in cardiovascular disease, emphasizing the importance of appropriate reference interval ranges for future clinical decision making.

Metabolic profiles of socio-economic position: a multi-cohort analysis.

BACKGROUND: Low socio-economic position (SEP) is a risk factor for multiple health outcomes, but its molecular imprints in the body remain unclear. METHODS: We examined SEP as a determinant of serum nuclear magnetic resonance metabolic profiles in ∼30 000 adults and 4000 children across 10 UK and Finnish cohort studies. RESULTS: In risk-factor-adjusted analysis of 233 metabolic measures, low educational attainment was associated with 37 measures including higher levels of triglycerides in small high-density lipoproteins (HDL) and lower levels of docosahexaenoic acid (DHA), omega-3 fatty acids, apolipoprotein A1, large and very large HDL particles (including levels of their respective lipid constituents) and cholesterol measures across different density lipoproteins. Among adults whose father worked in manual occupations, associations with apolipoprotein A1, large and very large HDL particles and HDL-2 cholesterol remained after adjustment for SEP in later life. Among manual workers, levels of glutamine were higher compared with non-manual workers. All three indicators of low SEP were associated with lower DHA, omega-3 fatty acids and HDL diameter. At all ages, children of manual workers had lower levels of DHA as a proportion of total fatty acids. CONCLUSIONS: Our work indicates that social and economic factors have a measurable impact on human physiology. Lower SEP was independently associated with a generally unfavourable metabolic profile, consistent across ages and cohorts. The metabolites we found to be associated with SEP, including DHA, are known to predict cardiovascular disease and cognitive decline in later life and may contribute to health inequalities.

Novel Insights Into the Effects of Interleukin 6 Antagonism in Non-ST-Segment-Elevation Myocardial Infarction Employing the SOMAscan Proteomics Platform.

Background Interleukin 6 concentration is associated with myocardial injury, heart failure, and mortality after myocardial infarction. In the Norwegian tocilizumab non-ST-segment-elevation myocardial infarction trial, the first randomized trial of interleukin 6 blockade in myocardial infarction, concentration of both C-reactive protein and troponin T were reduced in the active treatment arm. In this follow-up study, an aptamer-based proteomic approach was employed to discover additional plasma proteins modulated by tocilizumab treatment to gain novel insights into the effects of this therapeutic approach. Methods and Results Plasma from percutaneous coronary intervention-treated patients, 24 in the active intervention and 24 in the placebo-control arm, drawn 48 hours postrandomization were randomly selected for analysis with the SOMAscan assay. Employing slow off-rate aptamers, the relative abundance of 1074 circulating proteins was measured. Proteins identified as being significantly different between groups were subsequently measured by enzyme immunoassay in the whole trial cohort (117 patients) at all time points (days 1-3 [7 time points] and 3 and 6 months). Five proteins identified by the SOMAscan assay, and subsequently confirmed by enzyme immunoassay, were significantly altered by tocilizumab administration. The acute-phase proteins lipopolysaccharide-binding protein, hepcidin, and insulin-like growth factor-binding protein 4 were all reduced during the hospitalization phase, as was the monocyte chemoattractant C-C motif chemokine ligand 23. Proteinase 3, released primarily from neutrophils, was significantly elevated. Conclusions Employing the SOMAscan aptamer-based proteomics platform, 5 proteins were newly identified that are modulated by interleukin 6 antagonism and may mediate the therapeutic effects of tocilizumab in non-ST-segment-elevation myocardial infarction.

Triglyceride-containing lipoprotein sub-fractions and risk of coronary heart disease and stroke: A prospective analysis in 11,560 adults.

AIMS: Elevated low-density lipoprotein cholesterol (LDL-C) is a risk factor for cardiovascular disease; however, there is uncertainty about the role of total triglycerides and the individual triglyceride-containing lipoprotein sub-fractions. We measured 14 triglyceride-containing lipoprotein sub-fractions using nuclear magnetic resonance and examined associations with coronary heart disease and stroke. METHODS: Triglyceride-containing sub-fraction measures were available in 11,560 participants from the three UK cohorts free of coronary heart disease and stroke at baseline. Multivariable logistic regression was used to estimate the association of each sub-fraction with coronary heart disease and stroke expressed as the odds ratio per standard deviation increment in the corresponding measure. RESULTS: The 14 triglyceride-containing sub-fractions were positively correlated with one another and with total triglycerides, and inversely correlated with high-density lipoprotein cholesterol (HDL-C). Thirteen sub-fractions were positively associated with coronary heart disease (odds ratio in the range 1.12 to 1.22), with the effect estimates for coronary heart disease being comparable in subgroup analysis of participants with and without type 2 diabetes, and were attenuated after adjustment for HDL-C and LDL-C. There was no evidence for a clear association of any triglyceride lipoprotein sub-fraction with stroke. CONCLUSIONS: Triglyceride sub-fractions are associated with increased risk of coronary heart disease but not stroke, with attenuation of effects on adjustment for HDL-C and LDL-C.

Liver Function and Risk of Type 2 Diabetes: Bidirectional Mendelian Randomization Study.

Liver dysfunction and type 2 diabetes (T2D) are consistently associated. However, it is currently unknown whether liver dysfunction contributes to, results from, or is merely correlated with T2D due to confounding. We used Mendelian randomization to investigate the presence and direction of any causal relation between liver function and T2D risk including up to 64,094 T2D case and 607,012 control subjects. Several biomarkers were used as proxies of liver function (i.e., alanine aminotransferase [ALT], aspartate aminotransferase [AST], alkaline phosphatase [ALP], and γ-glutamyl transferase [GGT]). Genetic variants strongly associated with each liver function marker were used to investigate the effect of liver function on T2D risk. In addition, genetic variants strongly associated with T2D risk and with fasting insulin were used to investigate the effect of predisposition to T2D and insulin resistance, respectively, on liver function. Genetically predicted higher circulating ALT and AST were related to increased risk of T2D. There was a modest negative association of genetically predicted ALP with T2D risk and no evidence of association between GGT and T2D risk. Genetic predisposition to higher fasting insulin, but not to T2D, was related to increased circulating ALT. Since circulating ALT and AST are markers of nonalcoholic fatty liver disease (NAFLD), these findings provide some support for insulin resistance resulting in NAFLD, which in turn increases T2D risk.

Genetic studies of accelerometer-based sleep measures yield new insights into human sleep behaviour.

Sleep is an essential human function but its regulation is poorly understood. Using accelerometer data from 85,670 UK Biobank participants, we perform a genome-wide association study of 8 derived sleep traits representing sleep quality, quantity and timing, and validate our findings in 5,819 individuals. We identify 47 genetic associations at P < 5 × 10-8, of which 20 reach a stricter threshold of P < 8 × 10-10. These include 26 novel associations with measures of sleep quality and 10 with nocturnal sleep duration. The majority of identified variants associate with a single sleep trait, except for variants previously associated with restless legs syndrome. For sleep duration we identify a missense variant (p.Tyr727Cys) in PDE11A as the likely causal variant. As a group, sleep quality loci are enriched for serotonin processing genes. Although accelerometer-derived measures of sleep are imperfect and may be affected by restless legs syndrome, these findings provide new biological insights into sleep compared to previous efforts based on self-report sleep measures.

Vitamin D and cognitive function: A Mendelian randomisation study.

The causal nature of the association between hypovitaminosis D and poor cognitive function in mid- to later-life is uncertain. Using a Mendelian randomisation(MR) approach, we examined the causal relationship between 25(OH)D and cognitive function. Data came from 172,349 participants from 17 cohorts. DHCR7(rs12785878), CYP2R1 rs12794714) and their combined synthesis score were chosen to proxy 25(OH)D. Cognitive tests were standardised into global and memory scores. Analyses were stratified by 25(OH)D tertiles, sex and age. Random effects meta-analyses assessed associations between 25(OH)D and cognitive function. Associations of serum 25(OH)D with global and memory-related cognitive function were non-linear (lower cognitive scores for both low and high 25(OH)D, p curvature ≤ 0.006), with much of the curvature attributed to a single study. DHCR7, CYP2R1, and the synthesis score were associated with small reductions in 25(OH)D per vitamin D-decreasing allele. However, coefficients for associations with global or memory-related cognitive function were non-significant and in opposing directions for DHCR7 and CYP2R1, with no overall association observed for the synthesis score. Coefficients for the synthesis score and global and memory cognition were similar when stratified by 25(OH)D tertiles, sex and age. We found no evidence for serum 25(OH)D concentration as a causal factor for cognitive performance in mid- to later life.

Causal Associations of Adiposity and Body Fat Distribution With Coronary Heart Disease, Stroke Subtypes, and Type 2 Diabetes Mellitus: A Mendelian Randomization Analysis.

BACKGROUND: The implications of different adiposity measures on cardiovascular disease etiology remain unclear. In this article, we quantify and contrast causal associations of central adiposity (waist-to-hip ratio adjusted for body mass index [WHRadjBMI]) and general adiposity (body mass index [BMI]) with cardiometabolic disease. METHODS: Ninety-seven independent single-nucleotide polymorphisms for BMI and 49 single-nucleotide polymorphisms for WHRadjBMI were used to conduct Mendelian randomization analyses in 14 prospective studies supplemented with coronary heart disease (CHD) data from CARDIoGRAMplusC4D (Coronary Artery Disease Genome-wide Replication and Meta-analysis [CARDIoGRAM] plus The Coronary Artery Disease [C4D] Genetics; combined total 66 842 cases), stroke from METASTROKE (12 389 ischemic stroke cases), type 2 diabetes mellitus from DIAGRAM (Diabetes Genetics Replication and Meta-analysis; 34 840 cases), and lipids from GLGC (Global Lipids Genetic Consortium; 213 500 participants) consortia. Primary outcomes were CHD, type 2 diabetes mellitus, and major stroke subtypes; secondary analyses included 18 cardiometabolic traits. RESULTS: Each one standard deviation (SD) higher WHRadjBMI (1 SD≈0.08 U) associated with a 48% excess risk of CHD (odds ratio [OR] for CHD, 1.48; 95% confidence interval [CI], 1.28-1.71), similar to findings for BMI (1 SD≈4.6 kg/m2; OR for CHD, 1.36; 95% CI, 1.22-1.52). Only WHRadjBMI increased risk of ischemic stroke (OR, 1.32; 95% CI, 1.03-1.70). For type 2 diabetes mellitus, both measures had large effects: OR, 1.82 (95% CI, 1.38-2.42) and OR, 1.98 (95% CI, 1.41-2.78) per 1 SD higher WHRadjBMI and BMI, respectively. Both WHRadjBMI and BMI were associated with higher left ventricular hypertrophy, glycemic traits, interleukin 6, and circulating lipids. WHRadjBMI was also associated with higher carotid intima-media thickness (39%; 95% CI, 9%-77% per 1 SD). CONCLUSIONS: Both general and central adiposity have causal effects on CHD and type 2 diabetes mellitus. Central adiposity may have a stronger effect on stroke risk. Future estimates of the burden of adiposity on health should include measures of central and general adiposity.

The druggable genome and support for target identification and validation in drug development.

Target identification (determining the correct drug targets for a disease) and target validation (demonstrating an effect of target perturbation on disease biomarkers and disease end points) are important steps in drug development. Clinically relevant associations of variants in genes encoding drug targets model the effect of modifying the same targets pharmacologically. To delineate drug development (including repurposing) opportunities arising from this paradigm, we connected complex disease- and biomarker-associated loci from genome-wide association studies to an updated set of genes encoding druggable human proteins, to agents with bioactivity against these targets, and, where there were licensed drugs, to clinical indications. We used this set of genes to inform the design of a new genotyping array, which will enable association studies of druggable genes for drug target selection and validation in human disease.

Marginal role for 53 common genetic variants in cardiovascular disease prediction.

OBJECTIVE: We investigated discrimination and calibration of cardiovascular disease (CVD) risk scores when genotypic was added to phenotypic information. The potential of genetic information for those at intermediate risk by a phenotype-based risk score was assessed. METHODS: Data were from seven prospective studies including 11 851 individuals initially free of CVD or diabetes, with 1444 incident CVD events over 10 years' follow-up. We calculated a score from 53 CVD-related single nucleotide polymorphisms and an established CVD risk equation 'QRISK-2' comprising phenotypic measures. The area under the receiver operating characteristic curve (AUROC), detection rate for given false-positive rate (FPR) and net reclassification improvement (NRI) index were estimated for gene scores alone and in addition to the QRISK-2 CVD risk score. We also evaluated use of genetic information only for those at intermediate risk according to QRISK-2. RESULTS: The AUROC was 0.635 for QRISK-2 alone and 0.623 with addition of the gene score. The detection rate for 5% FPR improved from 11.9% to 12.0% when the gene score was added. For a 10-year CVD risk cut-off point of 10%, the NRI was 0.25% when the gene score was added to QRISK-2. Applying the genetic risk score only to those with QRISK-2 risk of 10%-<20% and prescribing statins where risk exceeded 20% suggested that genetic information could prevent one additional event for every 462 people screened. CONCLUSION: The gene score produced minimal incremental population-wide utility over phenotypic risk prediction of CVD. Tailored prediction using genetic information for those at intermediate risk may have clinical utility.

Replication and Characterization of Association between ABO SNPs and Red Blood Cell Traits by Meta-Analysis in Europeans.

Red blood cell (RBC) traits are routinely measured in clinical practice as important markers of health. Deviations from the physiological ranges are usually a sign of disease, although variation between healthy individuals also occurs, at least partly due to genetic factors. Recent large scale genetic studies identified loci associated with one or more of these traits; further characterization of known loci and identification of new loci is necessary to better understand their role in health and disease and to identify potential molecular mechanisms. We performed meta-analysis of Metabochip association results for six RBC traits-hemoglobin concentration (Hb), hematocrit (Hct), mean corpuscular hemoglobin (MCH), mean corpuscular hemoglobin concentration (MCHC), mean corpuscular volume (MCV) and red blood cell count (RCC)-in 11 093 Europeans from seven studies of the UCL-LSHTM-Edinburgh-Bristol (UCLEB) Consortium. We identified 394 non-overlapping SNPs in five loci at genome-wide significance: 6p22.1-6p21.33 (with HFE among others), 6q23.2 (with HBS1L among others), 6q23.3 (contains no genes), 9q34.3 (only ABO gene) and 22q13.1 (with TMPRSS6 among others), replicating previous findings of association with RBC traits at these loci and extending them by imputation to 1000 Genomes. We further characterized associations between ABO SNPs and three traits: hemoglobin, hematocrit and red blood cell count, replicating them in an independent cohort. Conditional analyses indicated the independent association of each of these traits with ABO SNPs and a role for blood group O in mediating the association. The 15 most significant RBC-associated ABO SNPs were also associated with five cardiometabolic traits, with discordance in the direction of effect between groups of traits, suggesting that ABO may act through more than one mechanism to influence cardiometabolic risk.

Mendelian Randomisation study of the influence of eGFR on coronary heart disease.

Impaired kidney function, as measured by reduced estimated glomerular filtration rate (eGFR), has been associated with increased risk of coronary heart disease (CHD) in observational studies, but it is unclear whether this association is causal or the result of confounding or reverse causation. In this study we applied Mendelian randomisation analysis using 17 genetic variants previously associated with eGFR to investigate the causal role of kidney function on CHD. We used 13,145 participants from the UCL-LSHTM-Edinburgh-Bristol (UCLEB) Consortium and 194,427 participants from the Coronary ARtery DIsease Genome-wide Replication and Meta-analysis plus Coronary Artery Disease (CARDIoGRAMplusC4D) consortium. We observed significant association of an unweighted gene score with CHD risk (odds ratio = 0.983 per additional eGFR-increasing allele, 95% CI = 0.970-0.996, p = 0.008). However, using weights calculated from UCLEB, the gene score was not associated with disease risk (p = 0.11). These conflicting results could be explained by a single SNP, rs653178, which was not associated with eGFR in the UCLEB sample, but has known pleiotropic effects that prevent us from drawing a causal conclusion. The observational association between low eGFR and increased CHD risk was not explained by potential confounders, and there was no evidence of reverse causation, therefore leaving the remaining unexplained association as an open question.

Exome-wide analysis of rare coding variation identifies novel associations with COPD and airflow limitation in MOCS3, IFIT3 and SERPINA12.

BACKGROUND: Several regions of the genome have shown to be associated with COPD in genome-wide association studies of common variants. OBJECTIVE: To determine rare and potentially functional single nucleotide polymorphisms (SNPs) associated with the risk of COPD and severity of airflow limitation. METHODS: 3226 current or former smokers of European ancestry with lung function measures indicative of Global Initiative for Chronic Obstructive Lung Disease (GOLD) 2 COPD or worse were genotyped using an exome array. An analysis of risk of COPD was carried out using ever smoking controls (n=4784). Associations with %predicted FEV1 were tested in cases. We followed-up signals of interest (p<10(-5)) in independent samples from a subset of the UK Biobank population and also undertook a more powerful discovery study by meta-analysing the exome array data and UK Biobank data for variants represented on both arrays. RESULTS: Among the associated variants were two in regions previously unreported for COPD; a low frequency non-synonymous SNP in MOCS3 (rs7269297, pdiscovery=3.08×10(-6), preplication=0.019) and a rare SNP in IFIT3, which emerged in the meta-analysis (rs140549288, pmeta=8.56×10(-6)). In the meta-analysis of % predicted FEV1 in cases, the strongest association was shown for a splice variant in a previously unreported region, SERPINA12 (rs140198372, pmeta=5.72×10(-6)). We also confirmed previously reported associations with COPD risk at MMP12, HHIP, GPR126 and CHRNA5. No associations in novel regions reached a stringent exome-wide significance threshold (p<3.7×10(-7)). CONCLUSIONS: This study identified several associations with the risk of COPD and severity of airflow limitation, including novel regions MOCS3, IFIT3 and SERPINA12, which warrant further study.