Asthma is a chronic inflammatory disease of the airways characterized by various pulmonary complications. It affects approximately 10% of Canadians and is a multifactorial disease with both genetic and non-genetic risk factors. Earlier studies estimated that genetics contribute to 35% to 95% of asthma heritability, yet only a small portion of this heritability was accounted by known loci. We hypothesize that the missing heritability of asthma may lie within rare variants, interactions among genes and interactions between genes with environmental exposures. In this study, we investigate the effects of rare variants as well as the combined effects of genes and modifiable exposures on asthma and asthma related outcomes such as recurrent wheeze in early childhood.

We ascertained genomics data (Illumina HumanCoreExome BeadChip) from the largest birth cohort in North America known as the Canadian Healthy Infant Longitudinal Development study (CHILD; 3455, trios). After quality control and imputations based on haplotype information from the 1000 Genomes Project, 2830 children and 23 million variants were included for our analysis. Status of recurrent wheeze, a clinical outcome correlated with asthma, reported from age 2 to 5 is used as a primary phenotype.

Through conventional univariate genome-wide association analysis (GWAS), we identified 32 significantly associated variants within 4 well-known asthma genes from chromosome 17. This validates prior findings and demonstrates that genetic correlations with asthma can be detected among children suffering from recurrent wheeze as early as age 2. Next, we performed a weight-based association analysis of rare variants (SNP-set kernel association test; SKAT) that combines the effect of multiple rare variations by genomic regions. We focused this analysis on 111 well-known asthma genes, which identified that the zinc-finger binding protein IKZF3 gene consists of 479 rare variants that were significantly associated with childhood wheeze. Moreover, gene-environment interaction analysis identified a numerous variations where interaction effect of genotypes combined with exposure to prenatal smoking posed strong correlations with recurrent wheeze. Finally, interactions among genes (epistasis) were also studied by constructing a network of intercorrelated genetic variants via hierarchical clustering, a conventional machine learning method for grouping elements based on similarity. This analysis identified 8 clusters of genetic variants that are associated (p<0.05) with childhood wheeze.

In this study, we identified the effects of rare variants, gene-environment interactions, and gene-gene interactions on childhood recurrent wheezing in a large Canadian birth cohort. In future studies, we will be expending our analyses to additional asthma-related phenotypes including positive skin tests and lung function measures. Moreover, we will integrate a broader range of environmental variables, including indoor and outdoor pollution exposure, pet exposure, breastfeeding, diet, family medical history, and medication usage. The study has also collected transcriptomics, epigenomics, metabolomics and microbiome data, which will enable integrative -omics analysis to identify connections between various levels of high-dimensional biological data.