Exploring the role of rare and low-frequency coding variants in adult height using an ExomeChip. M. Graff1, K. Sin lo2, K. Stirrups3, C. Medina-Gomez4, T. Esko5,6, N. L. Heard-Costa7, A. E. Justice1, T. W. Winkler8, L. Southam3,9, C. Shurmann10, J. Czajkowski11, Y. Lu10, K. L. Young1, T. L. Edwards12, A. Giri12, C. Lindgren13, 9, I. B. Borecki11, K. E. North1, 14, M. McCarthy15,9, J. N. Hirschhorn4, 13,16, P. Deloukas3,17, F. Rivadeneira4, T. M. Frayling18, R. J. F. Loos10, G. Lettre19,20 For the BBMRI, the GOT2D, the CHARGE, and the GIANT Consortia. 1) Epidemiology, University of North Carolina at Chapel Hill, Chapel Hill, NC; 2) Montreal Heart Institute, Montreal, Quebec, Canada; 3) The Wellcome Trust Sanger Institute, Wellcome Trust Genome Campus, Hinxton, Cambridge, UK; 4) Departments of Epidemiology and Internal Medicine, Erasmus Medical Center, Rotterdam, The Netherlands; 5) Divisions of Endocrinology and Genetics and Center for Basic and Translational Obesity Research, Boston Children's Hospital, Boston, MA, USA; 6) Estonian Genome Center, University of Tartu, Tartu, Estonia; 7) National Heart, Lung, and Blood Institute, the Framingham Heart Study, Framingham MA, USA; 8) Department of Genetic Epidemiology, Institute of Epidemiology and Preventive Medicine, University of Regensburg, Regensburg, Germany; 9) Wellcome Trust Centre for Human Genetics, University of Oxford, Oxford, UK; 10) The Genetics of Obesity and Related Metabolic Traits Program, The Charles Bronfman Institute for Personalized Medicine, Icahn School of Medicine at Mount Sinai, New York, NY, USA;; 11) Department of Genetics, Washington University School of Medicine, St. Louis, MO, USA; 12) Center for Human Genetics Research, Division of Epidemiology, Department of Medicine, Vanderbilt University, Nashville TN, USA; 13) Broad Institute of the Massachusetts Institute of Technology and Harvard University, Cambridge, MA, USA; 14) Carolina Center for Genome Sciences, University of North Carolina at Chapel Hill, Chapel Hill, NC, USA; 15) Oxford Centre for Diabetes, Endocrinology & Metabolism,University of Oxford, Oxford, UK; 16) Department of Genetics, Harvard Medical School, Boston, MA, USA; 17) William Harvey Research Institute, Barts and The London School of Medicine and Dentistry Queen Mary University of London Charterhouse Square, London, UK; 18) Genetics of Complex Traits, University of Exeter Medical School, University of Exeter, Exeter, UK; 19) Montreal Heart Institute,Université de Montréal, Montréal, Québec , Canada; 20) Faculty of Medicine, Université de Montréal, Montreal, Québec, Canada.
Adult height is a model complex phenotype being highly heritable, polygenic and accurately measured in large numbers of individuals. Genome wide association studies (GWAS) of 250,000 individuals have identified 697 common genetic variants in 424 loci at p<5x10-8 that explain ~20% of heritability. All common variants captured by GWAS are likely to explain ~60% of heritability, leaving 40% unexplained. To investigate the contribution of rare (<1% MAF) and low frequency (>1<5% MAF) coding variants to adult height we examined exome array genotype data containing approximately 215,000 protein coding variants in >230,000 individuals from 86 studies. We identified 55 variants at MAF <5% (8<1%, 47 >1<5%) associated with height at p<5x10-8. Forty-three of these variants occurred in 26 known height loci, including rare missense variants in IHH (MAF 0.3% p = 4.7e-08) and FBN2 (MAF 0.7% p = 2.0e-21), two genes in which mutations cause monogenic growth defects. Twelve variants (2 rare, 10 low frequency) in 9 genes did not occur in known GWAS loci (>500kb) and strongly implicate the genes PDE5A, DLG5, AMOTL1, SERPINA1, ZNF646, ZBTB7B, LAMB2, DUSP1 and IL11 in human growth for the first time. The largest effect size was with a missense variant in PDE5A where heterozygous carriers were on ~2 cm taller/shorter (p=9.5e-12). In addition, we identified a previously undetected common insertion deletion (MAF 10%, p=1x10-20, ~0.4 cm change in height) in a known locus that results in a frameshift in exon 1 in CPNE1. We conclude that exome chip sample sizes of similar magnitude to GWAS will likely identify low frequency variants of larger effect. However, given that the exome array only covers ~1% of the genome we will expect fewer loci. A large fraction of low frequency coding variant associations overlapped with known loci, and further studies will be needed to dissect whether these are due to linkage disequilibrium with known signals or represent independent functional variants.
You may contact the first author (during and after the meeting) at