Exome-wide analysis of copy number variation shows association of the human leukocyte antigen region with asthma in UK Biobank

Abstract

Background The role of copy number variants (CNVs) in susceptibility to asthma is not well understood. This is, in part, due to the difficulty of accurately measuring CNVs in large enough sample sizes to detect associations. The recent availability of whole-exome sequencing (WES) in large biobank studies provides an unprecedented opportunity to study the role of CNVs in asthma. Methods We called common CNVs in 49,953 individuals in the first release of UK Biobank WES using ClinCNV software. CNVs were tested for association with asthma in a stage 1 analysis comprising 7098 asthma cases and 36,578 controls from the first release of sequencing data. Nominally-associated CNVs were then meta-analysed in stage 2 with an additional 17,280 asthma cases and 115,562 controls from the second release of UK Biobank exome sequencing, followed by validation and fine-mapping. Results Five of 189 CNVs were associated with asthma in stage 2, including a deletion overlapping the HLA-DQA1 and HLA-DQB1 genes, a duplication of CHROMR/PRKRA, deletions within MUC22 and TAP2, and a duplication in FBRSL1. The HLA-DQA1, HLA-DQB1, MUC22 and TAP2 genes all reside within the human leukocyte antigen (HLA) region on chromosome 6. In silico analyses demonstrated that the deletion overlapping HLA-DQA1 and HLA-DQB1 is likely to be an artefact arising from under-mapping of reads from non-reference HLA haplotypes, and that the CHROMR/PRKRA and FBRSL1 duplications represent presence/absence of pseudogenes within the HLA region. Bayesian fine-mapping of the HLA region suggested that there are two independent asthma association signals. The variants with the largest posterior inclusion probability in the two credible sets were an amino acid change in HLA-DQB1 (glutamine to histidine at residue 253) and a multi-allelic amino acid change in HLA-DRB1 (presence/absence of serine, glycine or leucine at residue 11). Conclusions At least two independent loci characterised by amino acid changes in the HLA-DQA1, HLA-DQB1 and HLA-DRB1 genes are likely to account for association of SNPs and CNVs in this region with asthma. The high divergence of haplotypes in the HLA can give rise to spurious CNVs, providing an important, cautionary tale for future large-scale analyses of sequencing data.

Exome variants associated with asthma and allergy

“…Although we can assume from the large UK Biobank that there are no genotyping errors, that phenotypes have been correctly ascribed and data correctly handled there is still some risk of the known “Weisburd’s paradox”, the difficulty in maintaining quality control as studies get larger 16 . Non-exon and some ultra-rare exonic variants are not included in the current analysis, neither are copy number variants 17 . There are also two more paper from the UK Biobank published during revision of the current paper including the association of eosinophil counts with ALOX15, CSF2RB, IL17RA, IL33, JAK2, S1PR4, SH2B3, NPAT and RMI1 18 as well as an abandoned preprint 19 from the Scandinavian Asthma Genetic Study that includes also UK Biobank participants highlighting protein-truncating variants in FLG and IL33.…”

DeepPheWAS: an R package for phenotype generation and association analysis for phenome-wide association studies

“…Human genomic variation includes variants which have more categories than SNPs. For example, the diploid human copy number of CCL3L1 ranges from 0 to 8 in UK Biobank participants ( Fawcett et al , 2022 ) ( Supplementary Fig. S5 ).…”

The complexity in DNA methylation analysis of allergic diseases

“…Another EWAS was carried out in two cohorts of an English population ( n = 1456), with DNAm measured at age 10 and spirometry measurements at three time points 10, 18 and 26 years old [26 ▪ ]. A total of 44 CpGs were associated with lung function trajectories, annotating to 73 distinct genes, of which 6 were linked to gene expression: LMF1 is involved in protein maturation, SMAD2 , of the important in development TGF-β-SMAD-signalling pathway, BTNL9, involved in T-cell receptor-signalling pathway and cytokine production, FBRSL1 , a duplication in which have been associated with asthma [27], and has been found to be differentially expressed in asthmatic patients [28]. Using the same cohorts, another analysis estimated the association between DNAm at birth and lung function at three points until age 26 [29].…”