Multi-trait meta-analysis of genome-wide association studies (GWAS) of lipid levels and BMI reveals pleiotropy. V. Lagou1,2, R. Mägi3, I. Surakka4,5, A.-P. Sarin4,5, M. Horikoshi1,2, L. Marullo6, T. Ferreira1, G. Thorleifsson7, S. Hägg8,9, M. Beekman10,11, C. Ladenvall12, A. Mahajan1, J.-J. Hottenga13, J. S. Ried14, T. W. Winkler15, C. Willenborg16, M. I. McCarthy1,2, A. P. Morris1, S. Ripatti4,5,17, I. Prokopenko1,2,18, ENGAGE (European Network for Genetic and Genomic Epidemiology) consortium 1) Wellcome Trust Centre for Human Genetics, University of Oxford, Oxford, United Kingdom; 2) Oxford Centre for Diabetes, Endocrinology and Metabolism, Radcliffe Department of Medicine, University of Oxford, Oxford, United Kingdom; 3) Estonian Genome Center, University of Tartu, Tartu, Estonia; 4) Institute for Molecular Medicine Finland FIMM, University of Helsinki, Helsinki, Finland; 5) National Institute for Health and Welfare, Helsinki, Finland; 6) Department of Life Sciences and Biotechnology, University of Ferrara, Ferrara, Italy; 7) deCODE Genetics, 101 Reykjavik, Iceland; 8) Department of Medical Epidemiology and Biostatistics, Karolinska Institutet, Stockholm, Sweden; 9) Department of Medical Sciences, Molecular Epidemiology, Uppsala University Hospital, Uppsala, Sweden; 10) Department of Molecular Epidemiology, Leiden University Medical Center, Leiden, The Netherlands; 11) Netherlands Consortium for Healthy Ageing, Leiden, The Netherlands; 12) Department of Clinical Sciences, Diabetes and Endocrinology, Lund University and Lund University Diabetes Centre, CRC at Skåne University Hospital, Malmö, Sweden; 13) Netherlands Twin Register, Dept Biological Psychology, VU Univ Amsterdam, The Netherlands; 14) Institute of Genetic Epidemiology, Helmholtz Zentrum München - German Research Center for Environmental Health, Neuherberg, Germany; 15) Department of Genetic Epidemiology, Institute of Epidemiology and Preventive Medicine, University of Regensburg, Regensburg, Germany; 16) AG Kardiovaskuläre Genomik, Medizinische Klinik II, Universität zu Lübeck, Lübeck, Germany; 17) Wellcome Trust Sanger Institute, United Kingdom; 18) Department of Genomics of Common Disease, School of Public Health, Imperial College London, Hammersmith Hospital, London, United Kingdom.
Serum lipid levels, fat storage and obesity share biochemical pathways and can be influenced by common genetic factors. Analysis of the genetic effects on multiple traits simultaneously allows dissection of variable patterns of multi-trait associations. Within the ENGAGE consortium, we assessed multi-trait genetic effects on four blood lipids (high-/low-density lipoprotein and total cholesterol, triglycerides [HDL/LDL/TC/TG]) and body-mass index (BMI). The 1000 Genomes reference panel (06/2011) was used for imputation in up to 50,539 individuals from 19 European GWAS. Each study carried out multi-trait analysis by fitting a multiple logistic regression on SNP genotypes allowing for joint effects of four lipid traits and BMI and combining evidence across study-specific likelihoods. Individual trait effects from each study multi-trait model were estimated by fixed effects inverse-variance meta-analyses. Single-trait meta-analyses, conditional on remaining traits, were used to verify the independence of trait-specific genetic effects. Joint analysis enabled identification of 26 signals with genome-wide significant (PLRT<5.0x10-8) multi-trait effects, including 11 loci with associations driven by the individual trait effects: a) TRIB1 on BMI, b) GCKR, FADS1, PLTP on TG, c) CELSR2 on HDL, d) CEPT on BMI/HDL, e) MLXIPL, LPL, APOA1 on BMI/TG, f) LIPC on HDL/TG and g) APOE on HDL/LDL/TG. At three loci previously associated with adiponectin (TRIB1), liver enzymes (TRIB1, MLXIPL) and lipodystrophy (LPL), genome-wide significant effects on obesity were observed for the first time, with higher BMI being related to lower TG indicating complex relationships between obesity and regulation of lipid levels. At the remaining 15 pleiotropic loci, multiple traits contributed to the signal. Pleiotropy was also observed at two loci [IRS1 (PLRT=5.0x10-4) and SH2B1 (PLRT=5.0x10-3)] previously associated with adiposity in GWAS with higher BMI being related to higher HDL at IRS1 and higher BMI with lower TG at SH2B1. We detected a substantial proportion of metabolic trait loci with complex patterns of genetic effects, some of which may not follow epidemiological correlations. This study highlights the emerging need of systematic investigation of multi-trait effects across the genome.
You may contact the first author (during and after the meeting) at