BACKGROUND. High-impact (nonsense, splice or missense) variants have not been systematically studied for association with asthma risk.
AIM. Our aim was to screen all high-impact variants detectable in a small number of carefully selected asthmatic individuals (n=13) and prioritize for follow-up a set of uncommon variants likely to be related to asthma.
METHODS. We sequenced with >30X coverage the genomes of 13 doctor-diagnosed asthmatics, using Illumina GAII or HiSeq. We also sequenced both parents (discordant for asthma) of one asthmatic and an additional 18 unrelated controls. Asthmatics were selected based on seven criteria: (a) never smoker; (b) positive skin prick test; (c) positive bronchial hyperresponsiveness test; (d) childhood-onset asthma; (e) recent asthma treatment; (f) positive family history; and (g) a high ECRHS symptoms score. Sequence reads were analysed using the GATK pipeline, with variants called across the 33 samples jointly and annotated with ANNOVAR.
RESULTS. 8,769,721 SNPs passed QC, of which 92% were known (dbSNP 135) and 8% were novel; the transition to transversion ratio was 2.18 and 2.06, respectively. We identified 31,397 high-impact variants in 11,211 genes. To prioritize variants for follow-up, we selected variants: (a) called in the affected parent and the affected offspring, but not the unaffected mother of the trio (2,567 variants, 1,855 genes); (b) with a MAF<5% in the NHLBI GO Exome Sequencing and the 1000 Genomes Projects (328 variants, 310 genes); (c) more common in the 13 asthmatics than in the 18 controls (246 variants, 238 genes); and (d) located in genes with more high-impact variants in asthmatics than controls (96 variants, 89 genes). One variant per gene is currently being genotyped in 500 cases and 500 controls.
CONCLUSIONS. By sequencing 33 individuals, we identified 89 genes with high-impact uncommon variants that are plausibly related to asthma. These are being genotyped in a larger cohort.