Testing the role of predicted gene knockouts in human anthropometric trait variation

Human Molecular Genetics, 2016, Vol. 25, No. 10 2082–2092 doi: 10.1093/hmg/ddw055 Advance Access Publication Date: 21 February 2016 Association Studies Article A S S O C I AT I O N S T U D I E S A R T I C L E Testing the role of predicted gene knockouts in human anthropometric trait variation 1 Montreal Heart Institute, Montréal, Québec H1T 1C8, Canada, 2Faculté de Médecine, Université de Montréal, Montréal, Québec H3T 1J4, Canada, 3Medical and Population Genetics Program, Broad Institute, Cambridge, MA 02142, USA, 4Center for Human Genetics Research, Massachusetts General Hospital, Boston, MA 02114, USA, 5 Department of Medicine and, 6Department of Genetics, Harvard Medical School, Boston, MA 02115, USA, 7School of Public Health, University of Wisconsin-Milwaukee, Milwaukee, WI 53201-0413, USA, 8Division of Epidemiology, Institute for Medicine and Public Health and, 9Department of Medicine, Vanderbilt Genetics Institute, Vanderbilt University Medical Center, Nashville, TN 37203, USA, 10University of North Carolina Gillings School of Global Public Health, Chapel Hill, NC 27599, USA, 11The Charles Bronfman Institute for Personalized Medicine and, 12The Mindich Child Health and Development Institute, the Genetics of Obesity and Related Metabolic Traits Program, The Icahn School of Medicine at Mount Sinai, New York, NY 10029, USA, 13Genetics of Complex Traits, University of Exeter Medical School, University of Exeter, Exeter EX2 5DW, UK, 14Medical Research Council Epidemiology Unit, University of Cambridge, Cambridge CB2 0QQ, UK, 15Estonian Genome Center, University of Tartu, Tartu, Estonia, 16Division of Endocrinology, Genetics and Basic and Translational Obesity Research, Children’s Hospital Boston, Boston, MA 02115, USA, 17Wellcome Trust Centre for Human Genetics, University of Oxford, Oxford OX3 7BN, UK, 18Division of Public Health Sciences, Fred Hutchinson Cancer Research Center, Seattle, WA 98109-1024, USA, 19Clinical Pharmacology, William Harvey Research Institute and, 20NIHR Barts Cardiovascular Biomedical Research Unit, Barts and The London School of Medicine and Dentistry, Queen Mary University of London, London EC1M 6BQ, UK, 21Division of Endocrinology, Diabetes and Metabolism, Ohio State University, Columbus, OH 43210, USA, 22Department of Biostatistics, University of Washington, Seattle, WA 98195, USA, 23Division of † A full list of members and affiliations appears in the Supplementary Text. Received: October 14, 2015. Revised: February 2, 2016. Accepted: February 15, 2016 © The Author 2016. Published by Oxford University Press. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted reuse, distribution, and reproduction in any medium, provided the original work is properly cited. 2082 Samuel Lessard1,2, Alisa K. Manning3,4,5, Cécile Low-Kam1,2, Paul L. Auer7, Ayush Giri8, Mariaelisa Graff10, Claudia Schurmann11,12, Hanieh Yaghootkar13, Jian’an Luan14, Tonu Esko3,15,16, Tugce Karaderi17, NHLBI GO Exome Sequence Project†, GOT2D†, T2D-GENES†, GIANT Consortium†, Erwin P. Bottinger11, Yingchang Lu11, Chris Carlson18, Mark Caulfield19,20, Marie-Pierre Dubé1,2, Rebecca D. Jackson21, Charles Kooperberg18, Barbara McKnight22, Ian Mongrain1, Ulrike Peters18, Alex P. Reiner18, David Rhainds1, Nona Sotoodehnia23, Joel N. Hirschhorn3,6,16, Robert A. Scott14, Patricia B. Munroe19,20, Timothy M. Frayling13, Ruth J.F. Loos11,12, Kari E. North10, Todd L. Edwards8,9, Jean-Claude Tardif1,2, Cecilia M. Lindgren3,17,24 and Guillaume Lettre1,2* Human Molecular Genetics, 2016, Vol. 25, No. 10 | 2083 Cardiology, Cardiovascular Health Research Unit, University of Washington, Seattle, WA 98195-6422, USA and 24The Big Data Institute, University of Oxford, Oxford, UK *To whom correspondence should be addressed at: Montreal Heart Institute, 5000 Bélanger Street, Montréal, Québec H1T 1C8, Canada. Email: guillaume. Abstract Introduction The identification of complete loss-of-function (LoF) alleles (i.e. genetic null or amorphic alleles) is a powerful strategy to characterize gene functions through random (e.g. chemical mutagenesis) or targeted [e.g. knockout (KO) methodology in the mouse, RNAi] genetic experiments. In contrast to model organisms, humans are not amenable to such genetic manipulations. Yet, there is tremendous biomedical interest in understanding how the complete disruption of both copies of a gene may impact human biology (1). Our complex physiology, interactions with our environment, and gene redundancy within our genome are only few of the reasons highlighting the importance of describing the phenotypic consequences of gene inactivation in humans. From a drug development perspective, the identification of humans with gene KOs also offers naturally occurring genetic experiments to assess the potential pleiotropic effects of candidate target genes (2). Mendelian diseases, such as sickle cell anemia [MIM 603903] and cystic fibrosis [MIM 219700], offer an entry point into the study of gene functions in humans. Indeed, the study of these conditions continues to yield important insights into human biology in health and disease (3). But only a limited number of genes have been implicated in Mendelian diseases: as of October 13, 2015, there were 4651 genes in the Online Mendelian Inheritance in Man (OMIM) database with phenotype-causing mutations. Furthermore, these mutations are often rare such that it is difficult to assemble sufficiently large cohorts of patients to study their pleiotropic effects. Gene KOs can have strong phenotypic effects on anthropometric traits in the context of Mendelian disorders or syndromes, as evident by mutations causing earlyonset morbid obesity (PCSK1, LEPR) or dwarfism (GH1 GHR, ATR) (4–6). These mutations are rare (often private) and unlikely to be involved in anthropometric trait variation in the general population. However, the possibility that gene KOs of more subtle effect might influence normal variation in anthropometric traits remains to be investigated. Large-scale whole-exome and -genome sequencing projects are beginning to systematically catalogue coding genetic variation in the human genome, including predicted LoF variants (7–11). On average, there are ∼100–200 LoF variants per individual, resulting in ∼20 genes that are inactivated through homozygosity or compound heterozygosity (12). These numbers include mostly common variants, which are more likely to be phenotypically neutral given the effect of purifying selection (13). Limiting to variants with a minor allele frequency (MAF) <0.5%, the 1000 Genomes Project estimated that there are 10–20 LoF variants per individual (8). LoF variants are usually defined as variants that truncate protein sequences [nonsense and frameshift insertion-deletion (indel)] or that abrogate splice sites or stop codons (stop-loss) (12). Using this definition of LoF variant, a (...truncated)