Shared loci but distinct variants underlie genetic architecture of allergic diseases

Abstract

Asthma, allergic rhinitis, and atopic dermatitis are common, complex traits that are frequently co-morbid and have strong genetic correlation. However, the extent to which genome-wide genetic correlation between traits reflects shared causal variants or risk genes remains unclear. To address this question, we used functional fine-mapping. We generated genomic annotations from primary cells treated with immunomodulatory stimuli, then used these data to identify likely causal variants mediating genetic risk for allergic diseases including adult-onset asthma, childhood-onset asthma, allergic rhinitis, and atopic dermatitis. After identifying likely causal variants, we combined our functional annotations with expression quantitative trait loci and activity-by-contact modeling to predict effector genes. We confirmed a high degree of genetic correlation between GWAS loci for allergic diseases, but on the local level very few of the hundreds of likely causal variants identified by functional fine-mapping were shared between diseases. Instead, we found that each allergic disease was associated with a set of mostly unique variants. Nonetheless, nearly 40% of effector genes predicted to be the regulatory targets of these variants were shared between more than one allergic disease. When we tested candidate regulatory elements containing likely causal variants, we found that regulatory elements demonstrated variable allele-specific enhancer activity depending on the cell type in which they were tested. Overall, our findings suggest a highly pleiotropic gene regulatory network underlying allergic diseases, wherein disease-specific risk variants affect different regulatory elements that converge on the same set of target genes.