Multistage genome-wide association meta-analyses identified two new loci for bone mineral density

Human Molecular Genetics, 2014, Vol. 23, No. 7 doi:10.1093/hmg/ddt575 Advance Access published on November 17, 2013 1923–1933 Multistage genome-wide association meta-analyses identified two new loci for bone mineral density 1 Center of System Biomedical Sciences, University of Shanghai for Science and Technology, Shanghai, China, Department of Biostatistics and Bioinformatics and 3Department of Epidemiology, Tulane University, New Orleans, LA, USA, 4Department of Internal Medicine, Seoul National University College of Medicine, Seoul, Korea, 5Department of Internal Medicine and 6Department of Epidemiology, Erasmus Medical Center, Rotterdam, The Netherlands, 7 Netherlands Genomics Initiative (NGI)– Sponsored Netherlands Consortium for Healthy Aging (NCHA), Leiden, The Netherlands, 8Human Genetics Group, University of Queensland Diamantina Institute, Brisbane, Queensland, Australia, 9 Department of Endocrinology, Royal Brisbane and Women’s Hospital, Brisbane, Queensland, Australia, 10Center for Genome Science, National Institute of Health, Osong Health Technology Administration Complex, Chungcheongbuk-do, Korea, 11Department of Preventive Medicine, Ajou University School of Medicine, Youngtong-Gu, Suwon, Korea, 12Rural Clinical School, The University of Queensland, Toowoomba, Australia, 13Academic Unit of Bone Metabolism, Metabolic Bone Centre, University of Sheffield, Sheffield, UK, 14NIHR Musculoskeletal Biomedical Research Unit, Sheffield Teaching Hospitals Trust, Sheffield, UK, 15School of Medicine and Pharmacology, University of Western Australia, Perth, Australia, 16Department of Endocrinology and Diabetes, Sir Charles Gairdner Hospital, Perth, Australia, 17Garvan Institute of Medical Research, University of New South Wales, Sydney, Australia, 18Menzies Research Institute, University of Tasmania, Hobart, Australia, 19Department of Medicine, University of Auckland, Auckland, New Zealand, 20 Kolling Institute, Royal North Shore Hospital, University of Sydney, Sydney, Australia, 21Medical Research Council Lifecourse Epidemiology Unit, University of Southampton, Southampton, UK, 22Program in Personalized and Genomic Medicine, and Department of Medicine, Division of Endocrinology, Diabetes and Nutrition, University of Maryland School of Medicine, Baltimore, MD, USA, 23Geriatric Research and Education Clinical Center (GRECC), Veterans Administration Medical Center, Baltimore, MD, USA, 24Key Laboratory of Protein Chemistry and Developmental Biology of State Education Ministry of China, Hunan Normal University, Changsha, China and 25Department of Basic Medical Science, University of Missouri-Kansas City, Kansas City, MO, USA 2 Received October 23, 2013; Revised October 23, 2013; Accepted November 11, 2013 ∗ To whom correspondence should be addressed at: Center for Bioinformatics and Genomics, School of Public Health and Tropical Medicine, Tulane University, 1440 Canal Street, Suite 2001, New Orleans, LA 70112, USA. Tel: +1 5049881310. Email: These authors contributed equally to this work. Their orders of appearances are arranged in alphabetical order of their last names. † # The Author 2013. Published by Oxford University Press. All rights reserved. For Permissions, please email: Lei Zhang1,2, Hyung Jin Choi4,{, Karol Estrada5,6,7,{, Paul J. Leo8,{, Jian Li2, Yu-Fang Pei1,2, Yinping Zhang2, Yong Lin1, Hui Shen2, Yao-Zhong Liu2, Yongjun Liu2, Yingchun Zhao2, Ji-Gang Zhang2, Qing Tian2, Yu-ping Wang2, Yingying Han1, Shu Ran1, Rong Hai1, Xue-Zhen Zhu1, Shuyan Wu1, Han Yan2, Xiaogang Liu2, Tie-Lin Yang2, Yan Guo2, Feng Zhang2, Yan-fang Guo2, Yuan Chen2, Xiangding Chen2, Lijun Tan2, Lishu Zhang2, Fei-Yan Deng2, Hongyi Deng2, Fernando Rivadeneira5,6,7, Emma L Duncan8,9, Jong Young Lee10, Bok Ghee Han10, Nam H. Cho11, Geoffrey C. Nicholson12, Eugene McCloskey13,14, Richard Eastell13, Richard L. Prince15,16, John A. Eisman17, Graeme Jones18, Ian R. Reid19, Philip N. Sambrook20, Elaine M. Dennison21, Patrick Danoy8, Laura M. Yerges-Armstrong22, Elizabeth A. Streeten22,23, Tian Hu3, Shuanglin Xiang24, Christopher J. Papasian25, Matthew A. Brown8,{, Chan Soo Shin4,{, André G. Uitterlinden5,6,7,{ and Hong-Wen Deng1,2,∗ 1924 Human Molecular Genetics, 2014, Vol. 23, No. 7 INTRODUCTION Osteoporosis is the most common metabolic skeletal disorder in humans. It predisposes people to fragility fracture particularly at the hip and confers substantial morbidity and mortality (1), affecting over 200 million people worldwide (2). Osteoporosis is mainly characterized by low bone mineral density (BMD), which is highly heritable with heritability ranging from 0.5 to 0.8 (3). To date, genome-wide association studies (GWASs) and their meta-analyses have identified over 50 loci associated with variations in BMD (4 – 12). Cumulatively, however, genetic loci identified through GWAS account for no more than 6% of total BMD phenotypic variation (6). Therefore, there is little doubt that additional novel loci await to be uncovered. We here report a new multistage genome-wide association meta-analysis of samples of diverse ancestries and of imputed sequence variants with the 1000 genomes project (1KG) reference panels (13). RESULTS This study of meta-analysis comprises three stages (Fig. 1). Stage 1 incorporated seven GWAS samples encompassing 11 140 individuals of diverse ancestries (Supplementary Material, Table S1). The majority (7738; 69.5%) of the individuals were women. Principal component analysis (PCA) was applied to each individual sample (14), and no population outliers were observed. Imputation with the 1KG reference panels generated 5 842 825 SNPs that were qualified for analysis (Supplementary Material, Table S2). After adjusting phenotypes by PCA in each individual study (14), overall genomic control inflation factors were small or modest in both individual GWAS and meta-analysis for each of the spine (SPN-), total hip (HIP-) and femoral neck (FNK-) BMD traits (l ¼ 0.99–1.06, Supplementary Material, Table S2), implying the limited effects of potential population stratification. A logarithmic quantile–quantile plot of the metaanalysis test statistics showed a marked deviation in the tail of the distribution, both in the gender combined and female-specific samples, implying the possible existence of true associations in these samples (Fig. 2). In the combined sample, a total of 281 SNPs from 10 genomic loci were associated with BMD at the genome-wide significance (GWS; 5.0 × 1028) level (Supplementary Material, Table S3). Another 102 SNPs from 18 additional loci yielded P-values between 1.0 × 1026 and 5.0 × 1028, which was defined as a borderline association (Supplementary Material, Table S4). In the female-specific sample, 45 SNPs from four loci were associated with BMD at the GWS level (Supplementary Material, Table S3); all of these loci overlapped with those identified with the combined sample. Another seven SNPs from an additional four loci were associated at the borde (...truncated)