Search

Publisher's version (útgefin grein), Blood pressure (BP) was inconsistently associated with migraine and the mechanisms of BP-lowering medications in migraine prophylaxis are unknown. Leveraging large-scale summary statistics for migraine (Ncases/Ncontrols = 59,674/316,078) and BP (N = 757,601), we find positive genetic correlations of migraine with diastolic BP (DBP, rg = 0.11, P = 3.56 × 10−06) and systolic BP (SBP, rg = 0.06, P = 0.01), but not pulse pressure (PP, rg = −0.01, P = 0.75). Cross-trait meta-analysis reveals 14 shared loci (P ≤ 5 × 10−08), nine of which replicate (P, This research has been conducted using the UK Biobank Resource under Application Number 29273. We would like to thank the participants and researchers from the UK Biobank, 23andMe, Inc., International Headache Genetics Consortium (IHGC), MEGASTROKE, CARDIoGRAM, and International Consortium of Blood Pressure-Genome Wide Association Studies (ICBP) who contributed or collected data. Daniel I. Chasman is funded by US National Institutes of Health and US National Institute of Neurological Disorders and Stroke (R21NS09296 and R21NS104398). Pamela M. Rist is funded by K01 HL128791. The MEGASTROKE project received funding from sources specified at http://www.megastroke.org/acknowledgments.html.

In the past seven years, Genome-Wide Association Studies (GWAS) have identified hundreds of variants associated with cardiometabolic quantitative traits and diseases. Many genetic loci appear to harbour variants associated with multiple phenotypes (cross-phenotype associations, CP). CP associations highlight that phenotypes may share common underlying genetic mechanisms that might, or might not, be consistent with epidemiological expectations and, therefore, add complexity to the relationships between human phenotypes. Pleiotropy occurs when the same genetic causal element affects more than one phenotype “in parallel” and can explain the presence of CP associations. It can appear at a single variant level, where a single causal variant is related to multiple phenotypes, or at a locus level, that is when multiple variants in the same gene or locus are associated with different phenotypes by affecting the same functional element. However, other potential genetic mechanisms, that can explain CP associations, exist. Among them, mediation occurs when a genetic variant is directly associated with a phenotype and that phenotype is itself causal for a second phenotype or more phenotypes; multi-phenotype allelic heterogeneity is a phenomenon which involves independent uncorrelated variants within the same locus which cause changes in multiple phenotypes, by affecting them through independent pathways related to distinct functional elements. The identification and characterisation of CP associations across the genome may help uncovering the mechanistic basis of physiological processes that underlie variability of cardiometabolic quantitative traits, and of pathogenetic processes leading to metabolic disorders. The definition of specific patterns of effect combinations on cardiometabolic phenotypes will highlight novel biological pathways, targets for translational research, for therapeutic intervention, and for the understanding of the pathophysiology of human metabolism. Based on this hypothesis, and in collaboration with the Cross-Consortia pleiotropy group and with the European Network for Genetic and Genomic Epidemiology (ENGAGE) consortium, my PhD project focused on dissection of CP effects, pleiotropy in particular, at common variants across the genome in association with cardiometabolic phenotypes. The objective was to improve our understanding of the extent of shared genetics between cardiometabolic phenotypes and of the influences of DNA sequence variation on risk of metabolic diseases, considering phenotypes as a range of inter-related manifestations of biological mechanisms rather than as isolated events. My research has been divided into three sub-projects: Project 1: Clustering and pathway analysis of univariate GWAS results for the detection of pleiotropic effects. We explored multi-phenotype effects at hundreds of established cardiometabolic genetic variants from published univariate GWAS meta-analyses on more than 20 respective phenotypes, by defining clusters of loci with similar multiple effects, comparing them to known epidemiological expectations, and identifying enriched biological networks within the most interesting groups of loci. Our results highlighted that many variants at cardiometabolic loci have multiple associations that characterise different aspects of metabolism. Cardiometabolic loci can be grouped according to their shared multi-phenotype effects and metabolic syndrome represents just one possible combination; in fact, several other unexpected combinations might be observed, for example healthy obesity/unhealthy leanness. We also highlighted that genetic loci with similar cardiometabolic effects are involved in shared biological pathways. Some of these may be expected, for instance, regulation of lipids metabolism or cholesterol transport for groups of loci with strong effects on lipids, and circulatory system processes for genes near blood pressure-association signals. Sometimes groups of loci affected fundamental cell functions, such as regulation of cellular processes, for the loci with effects on obesity and anthropometric traits. The enriched connectivity within pathway networks revealed new potential candidate genes and tissues of action that are more likely to have causal effect on phenotypes. Project 2: Validating pleiotropy and analysis of locus architecture in potential pleiotropic regions. We aimed to dissect the architecture of established cardiometabolic loci showing multiple associations for a better definition of the underlying mechanisms of multi-phenotype effects and for the discernment of potential pleiotropy from allelic heterogeneity. To this aim, we applied an approximate conditional analysis, based on observed linkage disequilibrium patterns, which led us to the discovery of multiple associations at adjacent variants that underlie the same genetic cause for variability of different phenotypes. Our results also highlighted that a substantial proportion of metabolic loci incorporate complex patterns of multi-phenotype allelic heterogeneity, thus suggesting an important contribution of this mechanism into cross-phenotype effects. Project 3: Application of a multivariate statistical approach for the study of pleiotropy within cardiometabolic phenotypes. We developed and applied a statistical strategy for joint multivariate analysis of multiple correlated phenotypes using individual genetic data from the ENGAGE consortium to discover new uncovered multiple associations and to follow-up GWAS meta-analysis at two loci, FTO and FADS1. Using this approach we were able to take into account correlation between phenotypes, and we achieved a boost in power; moreover, we improved precision of parameter estimates and of the identification of novel candidate genes. Our results allowed us to identify several variants jointly associated with multiple lipid traits and body mass index. Our approach was useful for the identification of mediation: we, in fact, confirmed mediation underlying causal relationship between adiposity and other cardiometabolic phenotypes at the FTO locus. Additionally, we demonstrated that multiple effects on cardiometabolic phenotypes attributable to the FADS1 locus are mediated by its independent, thus pleiotropic, effect on total cholesterol and triglycerides. In conclusion, we applied several statistical approaches which allowed dissecting suggestive CP effects and their mechanisms, including pleiotropy, mediation and allelic heterogeneity. Our analyses have demonstrated the complexity of the relationships between cardiometabolic phenotypes related to the variability of both, underlying genetic mechanisms and genetic loci architecture.