Large scale meta-analysis of genome-wide association studies for height in multiple ancestries. S. Vedantam 1,2; A. Locke 3; E. Marouli 4; S. Berndt 5; L. Yengo 6; AR. Wood 7; T. Ferreira 8; S. Graham 9; on behalf of the Genetic Investigation of ANthropometric Traits (GIANT) Consortium

Research output: Contribution to conferenceAbstract

Abstract

Human adult height is a highly heritable, classic polygenic trait. Previous genome wide association study (GWAS) meta-analyses and exome chip studies by the GIANT consortium and others have identified >900 independent common variant association signals in >400 near-independent genetic loci that explain >25% of the heritability in height. To identify additional associations with height, we performed a GWAS on 456,426 European-ancestry participants from UKBiobank (UKB) using BOLT-LMM and combined these results with available GWAS summary statistics from GIANT (N~250,000; Wood et al. 2015). We identified 3,290 genome-wide significant associations in 712 loci, explaining >30% of heritability. In parallel, we are performing a meta-analysis of >250 studies from 6 different ancestries (not including UKB, total N = 906,589). All have genotype data imputed from the 1000Genomes Phase3 reference panel; many European-ancestry samples were also imputed to the Haplotype Reference Consortium or population-specific reference panels. Linear mixed effects models (implemented in Rvtest) and the first four principal components were used to adjust for relatedness and population structure. Association testing used inverse normally transformed age-adjusted residuals, stratified by sex and (if relevant) disease status. A preliminary meta-analysis using raremetal in 140 of these studies with European ancestry samples (total N = 618,362), imputed to 1000Genomes Phase3, identifies 1186 independent loci (defined as 1Mb apart from each other), including novel signals. Of the 1000 lead variants where UKB results were available, 918 strongly replicated in UKB (p<5x10-5, which is p<0.05 corrected for 1000 tests). Encouragingly, despite the large sample size, little evidence of association was observed at a highly stratified variant known to be associated with lactase persistence (rs4988235, p-value = 0.002). We will be adding data from cohorts from multiple ancestries and from the UKBiobank, providing a multiethnic discovery meta-analysis sample size of >1.5 million; large additional replication cohorts will be available. Secondary analyses will include fine mapping, aggregate testing of low frequency variants, and pathway analyses to delineate likely causal genes and biological mechanisms. Through our collaborative efforts, we have assembled a unique, large resource for understanding the genetic architecture of height and, more generally, polygenic traits
Original languageEnglish
Publication statusAccepted - 2018
EventAmerican Society of Human Genetics 2018 - San Diego, United States
Duration: 16 Oct 201820 Oct 2018
http://www.ashg.org/2018meeting

Conference

ConferenceAmerican Society of Human Genetics 2018
Abbreviated titleASHG
CountryUnited States
CitySan Diego
Period16/10/201820/10/2018
Internet address

Fingerprint

Genome-Wide Association Study
Multifactorial Inheritance
Meta-Analysis
Exome
Genetic Loci
Haplotypes
Population
Genotype
Genome
Genes

Bibliographical note

As a member of the GIANT consortium

Cite this

@conference{df4aadc137b84c59b326d07bb0cac9c5,
title = "Large scale meta-analysis of genome-wide association studies for height in multiple ancestries.: S. Vedantam 1,2; A. Locke 3; E. Marouli 4; S. Berndt 5; L. Yengo 6; AR. Wood 7; T. Ferreira 8; S. Graham 9; on behalf of the Genetic Investigation of ANthropometric Traits (GIANT) Consortium",
abstract = "Human adult height is a highly heritable, classic polygenic trait. Previous genome wide association study (GWAS) meta-analyses and exome chip studies by the GIANT consortium and others have identified >900 independent common variant association signals in >400 near-independent genetic loci that explain >25{\%} of the heritability in height. To identify additional associations with height, we performed a GWAS on 456,426 European-ancestry participants from UKBiobank (UKB) using BOLT-LMM and combined these results with available GWAS summary statistics from GIANT (N~250,000; Wood et al. 2015). We identified 3,290 genome-wide significant associations in 712 loci, explaining >30{\%} of heritability. In parallel, we are performing a meta-analysis of >250 studies from 6 different ancestries (not including UKB, total N = 906,589). All have genotype data imputed from the 1000Genomes Phase3 reference panel; many European-ancestry samples were also imputed to the Haplotype Reference Consortium or population-specific reference panels. Linear mixed effects models (implemented in Rvtest) and the first four principal components were used to adjust for relatedness and population structure. Association testing used inverse normally transformed age-adjusted residuals, stratified by sex and (if relevant) disease status. A preliminary meta-analysis using raremetal in 140 of these studies with European ancestry samples (total N = 618,362), imputed to 1000Genomes Phase3, identifies 1186 independent loci (defined as 1Mb apart from each other), including novel signals. Of the 1000 lead variants where UKB results were available, 918 strongly replicated in UKB (p<5x10-5, which is p<0.05 corrected for 1000 tests). Encouragingly, despite the large sample size, little evidence of association was observed at a highly stratified variant known to be associated with lactase persistence (rs4988235, p-value = 0.002). We will be adding data from cohorts from multiple ancestries and from the UKBiobank, providing a multiethnic discovery meta-analysis sample size of >1.5 million; large additional replication cohorts will be available. Secondary analyses will include fine mapping, aggregate testing of low frequency variants, and pathway analyses to delineate likely causal genes and biological mechanisms. Through our collaborative efforts, we have assembled a unique, large resource for understanding the genetic architecture of height and, more generally, polygenic traits",
author = "Amy McKnight",
note = "As a member of the GIANT consortium; American Society of Human Genetics 2018, ASHG ; Conference date: 16-10-2018 Through 20-10-2018",
year = "2018",
language = "English",
url = "http://www.ashg.org/2018meeting",

}

TY - CONF

T1 - Large scale meta-analysis of genome-wide association studies for height in multiple ancestries.

T2 - S. Vedantam 1,2; A. Locke 3; E. Marouli 4; S. Berndt 5; L. Yengo 6; AR. Wood 7; T. Ferreira 8; S. Graham 9; on behalf of the Genetic Investigation of ANthropometric Traits (GIANT) Consortium

AU - McKnight, Amy

N1 - As a member of the GIANT consortium

PY - 2018

Y1 - 2018

N2 - Human adult height is a highly heritable, classic polygenic trait. Previous genome wide association study (GWAS) meta-analyses and exome chip studies by the GIANT consortium and others have identified >900 independent common variant association signals in >400 near-independent genetic loci that explain >25% of the heritability in height. To identify additional associations with height, we performed a GWAS on 456,426 European-ancestry participants from UKBiobank (UKB) using BOLT-LMM and combined these results with available GWAS summary statistics from GIANT (N~250,000; Wood et al. 2015). We identified 3,290 genome-wide significant associations in 712 loci, explaining >30% of heritability. In parallel, we are performing a meta-analysis of >250 studies from 6 different ancestries (not including UKB, total N = 906,589). All have genotype data imputed from the 1000Genomes Phase3 reference panel; many European-ancestry samples were also imputed to the Haplotype Reference Consortium or population-specific reference panels. Linear mixed effects models (implemented in Rvtest) and the first four principal components were used to adjust for relatedness and population structure. Association testing used inverse normally transformed age-adjusted residuals, stratified by sex and (if relevant) disease status. A preliminary meta-analysis using raremetal in 140 of these studies with European ancestry samples (total N = 618,362), imputed to 1000Genomes Phase3, identifies 1186 independent loci (defined as 1Mb apart from each other), including novel signals. Of the 1000 lead variants where UKB results were available, 918 strongly replicated in UKB (p<5x10-5, which is p<0.05 corrected for 1000 tests). Encouragingly, despite the large sample size, little evidence of association was observed at a highly stratified variant known to be associated with lactase persistence (rs4988235, p-value = 0.002). We will be adding data from cohorts from multiple ancestries and from the UKBiobank, providing a multiethnic discovery meta-analysis sample size of >1.5 million; large additional replication cohorts will be available. Secondary analyses will include fine mapping, aggregate testing of low frequency variants, and pathway analyses to delineate likely causal genes and biological mechanisms. Through our collaborative efforts, we have assembled a unique, large resource for understanding the genetic architecture of height and, more generally, polygenic traits

AB - Human adult height is a highly heritable, classic polygenic trait. Previous genome wide association study (GWAS) meta-analyses and exome chip studies by the GIANT consortium and others have identified >900 independent common variant association signals in >400 near-independent genetic loci that explain >25% of the heritability in height. To identify additional associations with height, we performed a GWAS on 456,426 European-ancestry participants from UKBiobank (UKB) using BOLT-LMM and combined these results with available GWAS summary statistics from GIANT (N~250,000; Wood et al. 2015). We identified 3,290 genome-wide significant associations in 712 loci, explaining >30% of heritability. In parallel, we are performing a meta-analysis of >250 studies from 6 different ancestries (not including UKB, total N = 906,589). All have genotype data imputed from the 1000Genomes Phase3 reference panel; many European-ancestry samples were also imputed to the Haplotype Reference Consortium or population-specific reference panels. Linear mixed effects models (implemented in Rvtest) and the first four principal components were used to adjust for relatedness and population structure. Association testing used inverse normally transformed age-adjusted residuals, stratified by sex and (if relevant) disease status. A preliminary meta-analysis using raremetal in 140 of these studies with European ancestry samples (total N = 618,362), imputed to 1000Genomes Phase3, identifies 1186 independent loci (defined as 1Mb apart from each other), including novel signals. Of the 1000 lead variants where UKB results were available, 918 strongly replicated in UKB (p<5x10-5, which is p<0.05 corrected for 1000 tests). Encouragingly, despite the large sample size, little evidence of association was observed at a highly stratified variant known to be associated with lactase persistence (rs4988235, p-value = 0.002). We will be adding data from cohorts from multiple ancestries and from the UKBiobank, providing a multiethnic discovery meta-analysis sample size of >1.5 million; large additional replication cohorts will be available. Secondary analyses will include fine mapping, aggregate testing of low frequency variants, and pathway analyses to delineate likely causal genes and biological mechanisms. Through our collaborative efforts, we have assembled a unique, large resource for understanding the genetic architecture of height and, more generally, polygenic traits

M3 - Abstract

ER -