Abstract 5607: Fine-mapping of GWAS loci in African American breast cancer cases and controls from the Black Women's Health Study

Abstract

Abstract Genome-wide association studies in European and Asian ancestry populations have identified more than 10 SNPs associated with breast cancer, but to date there have been no reports from GWAS in African ancestry (AA) populations. We carried out fine-mapping of the associated loci in an AA population by genotyping a dense set of tagSNPs in each of 12 regions among 906 breast cancer cases and 1,111 controls from the Black Women's Health Study (BWHS). The BWHS is a prospective cohort study of 59,000 U.S. black women who have been followed since 1995; 50% of participants gave a saliva sample by the mouthwash-swish method. Controls were selected from among participants who had never reported breast cancer and were matched to cases on geographic region (Northeast, South, Midwest, West), and year of birth (plus or minus one year). Genotyping was carried out at the Broad Institute on the Sequenom iPLEX genotyping system. For each GWAS locus, we selected all tagSNPs with a minor allele frequency ≥ 5% and r2 ≥ 0.9 in the linkage disequilibrium (LD) block that contained the index SNP or, if a LD block was not evident, +/- 100 kb around the index SNP. We downloaded SNPs from the Hapmap Yoruba database (public release #26) and then selected tagSNPs using Tagger software implemented in the Haploview program version 4.1. We also selected the top 30 ancestral informative markers (AIMs) from a list of validated SNPs with allele frequency differences of at least 0.75 for Africans versus Europeans so that we could estimate and control for individual European admixture proportions An average reproducibility of 99.4% was obtained from 98 blinded duplicates and all SNPs showed concordance of >95%. Mean calling rate in the final data set for both SNPs and samples was 99.0%. We used PLINK software version 1.06 to calculate summary statistics and to perform Cochran-Armitage trend tests of association with breast cancer. Odds ratios and 95% confidence intervals were estimated with logistic regression implemented in the PROC LOGISTIC in SAS. ORs were adjusted for age, region of residence, birthplace (US, foreign country), and % European admixture. Only one of the 12 index GWAS SNPs was replicated in our sample and it has been reported separately. For six of the loci (including chr 8q24 and chr 2q35), other SNPs in the same LD block as the index SNP were significantly associated with breast cancer (empirical p for trend <0.05). Haplotypes of the index SNP and “best” SNP from these data were constructed. Results will be presented for breast cancer overall and by estrogen receptor status. The results indicate that many of the regions first identified in agnostic scans of European or Asian ancestry populations are also important for AA populations. In some instances, fine-scale mapping has helped to narrow the likely causal regions, as would be expected given the smaller LD blocks in AA persons. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 102nd Annual Meeting of the American Association for Cancer Research; 2011 Apr 2-6; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2011;71(8 Suppl):Abstract nr 5607. doi:10.1158/1538-7445.AM2011-5607