Rare copy number variants in over 100,000 European ancestry subjects reveal multiple disease associations

Yun R Li ,Alexandria Thomas,Debra Abrams,Munir Khan,Michael March,Cecilia Kim,Frank Mentch,Bradley P Coe,Cuiping Hou,Renata Pellagrino,John M Maris ,Struan F.a Grant ,Xiao Chang,Fengxiang Wang,George Otieno,Marcella Devoto,Mingyao Li,Andrew Bridglall-Jhingoor,Marina Bakay,Meckenzie Behr,Zhi Wei,Maria Lemma,Maedeh Mohebnasab ,Dexter Hadley,Joseph Glessner ,Charlly Kao,Christopher J Cardinale ,Robert W Grundmeier ,Lawrence Shih Hsin Wu ,Shanell Harrison,Brendan J Keating ,Josephine Elia,Elizabeth Goldmuntz,Evan E Eichler,Haijun Qiu,John J Connolly ,Patrick Sleiman ,Jonathan P Bradfield ,Rosetta M Chiavacci ,Jin Li,Hákon Hákonarson

doi:10.1038/s41467-019-13624-1

Abstract

Copy number variants (CNVs) are suggested to have a widespread impact on the human genome and phenotypes. To understand the role of CNVs across human diseases, we examine the CNV genomic landscape of 100,028 unrelated individuals of European ancestry, using SNP and CGH array datasets. We observe an average CNV burden of ~650 kb, identifying a total of 11,314 deletion, 5625 duplication, and 2746 homozygous deletion CNV regions (CNVRs). In all, 13.7% are unreported, 58.6% overlap with at least one gene, and 32.8% interrupt coding exons. These CNVRs are significantly more likely to overlap OMIM genes (2.94-fold), GWAS loci (1.52-fold), and non-coding RNAs (1.44-fold), compared with random distribution (P < 1 × 10−3). We uncover CNV associations with four major disease categories, including autoimmune, cardio-metabolic, oncologic, and neurological/psychiatric diseases, and identify several drug-repurposing opportunities. Our results demonstrate robust frequency definition for large-scale rare variant association studies, identify CNVs associated with major disease categories, and illustrate the pleiotropic impact of CNVs in human disease.

Highlights

Copy number variants (CNVs) are suggested to have a widespread impact on the human genome and phenotypes
We identified a total of 11,314 deletion, 5625 duplication, and 2746 homozygous-deletion CNV regions (CNVRs), defined as a contiguous cluster of non-singleton single-nucleotide polymorphism (SNP) that spans
We identified candidate genes in DA-CNVRs in a number of wellestablished disease-associated loci, including chr2p24.3 (MYCN amplification in cancer)[24], chr22q11.21 (COMT and TBX1 deletion in neuropsychiatric disease and congenital heart conditions)[25,26,27], and chr17q21.1 (NR1D1, deletion and duplication associated with response to lithium in bipolar disease and major depressive disorder)[28,29]

Summary

Introduction

Copy number variants (CNVs) are suggested to have a widespread impact on the human genome and phenotypes. We uncover CNV associations with four major disease categories, including autoimmune, cardio-metabolic, oncologic, and neurological/psychiatric diseases, and identify several drugrepurposing opportunities. Our results demonstrate robust frequency definition for large-scale rare variant association studies, identify CNVs associated with major disease categories, and illustrate the pleiotropic impact of CNVs in human disease. Division of Oncology, The Children’s Hospital of Philadelphia, Philadelphia, Pennsylvania 19104, USA. Copy number variation regions (CNVRs) are significantly more likely to overlap OMIM genes (2.94-fold), genome-wide association studies (GWAS) loci (1.52-fold), and non-coding RNAs (ncRNAs; 1.44fold), compared with random distribution (P < 1 × 10−3). We uncover strong CNV associations with four major disease categories, including autoimmune, cardio-metabolic, oncologic, and neurological/psychiatric diseases, several of which impact genes that represent potential drug targets for future validation

Methods

Results

Conclusion