Imputation from 328 African Ancestry genomes reveals new associations with asthma in DPP10
Masuko H, Rafaels NM, Huang L, Chavan S, Wilson JG, Williams KL, Ware LB, Ober C, Meyers DA, Hartert TV, Foreman M, Ford JG, Gonza Burchard E, Bleecker ER, Dunston G, Taub M, Beaty TH, Ruczinski I, Mathias RA, Barnes KC. Imputation from 328 African Ancestry genomes reveals new associations with asthma in DPP10. 2015; 135(2 Suppl):AB162.
RATIONALE: The gene encoding dipeptidyl-peptidase 10 (DPP10) was originally identified as an asthma candidate gene through positional cloning and subsequently a genome-wide association study (GWAS) meta-analysis However, the preciseDPP10locus conferring risk to asthma is unclear.
METHODS: We performed whole genome sequencing (WGS) in 328African Americans in the Consortium on Asthma among African-ancestry Populations in the Americas (CAAPA). We imputed DPP10 genotypes (Chr2:114735246-117781491) in 447 asthmatic and 459 non-asthmatic subjects, previously genotyped on Illumina’s 650Y GWAS panel using Minimac on the CAAPA WGS data. We also used CalcMatch to check concordance with genotype data. Logistic regression models adjusting for the first two principal components were used to test for association. Haplotype tagging coverage was determined using Haploview’s tagger for common SNPs (MAF > 0.01) with r2 > 0.8.
RESULTS: We included 25,115 DPP10 variants after removing monomorphic and singleton SNPs. The genotype-mismatch error rate with WGS data was 1.66%, and the allele-mismatch error rate was 0.84%. Intronic variant rs76969515 was most strongly associated with asthma risks (P = 2.3 x 10-4; OR = 2.87). The top GWAS peak previously reported was also significantly associated with risk of asthma (rs1435879, P = 5.0 x 10-4, OR = 2.16). These variants were not in LD with each other. For the GWAS peak +-10kb, the 650Y panel tags 5 of 37 haplotypes.
CONCLUSIONS: Leveraging CAAPA WGS data, we established a pipeline for imputation using existing GWAS marker data to identify novel variants in DPP10 associated with asthma. We will continue using this pipeline to identify additional risk variants in candidate genes for asthma in populations of African descent.