# NCBI dbGaP analysis accession: pha002853 # Name: A genome-wide association study identifies alleles associated with risk of sporadic postmenopausal breast cancer # Description: The primary analysis of the CGEMS breast GWAS study explores the association between a single SNPs and breast cancer susceptibility in a group of 1,145 breast cancer patients and 1,142 controls. This exploration is done one SNP at a time, sequentially for each of the 483,123 SNPs maintained in the study. The analytic approach assumes no structure to the risk across the 3 possible genotypes at each locus. This approach maintains power to detect recessive or over-dominant alleles at the cost of a small decrease in power relative to a Cochrane-Armitage trend test for the detection of alleles with multiplicative risk effect. By maximizing genome coverage with a large number of SNPs and adopting an ‘agnostic’ approach to the analysis which does not take gene function or prior information on breast cancer or other phenotypes into consideration, we increase the opportunity to pursue different working hypotheses and different regions of interest now and in the future. # Method: The primary analysis consisted of fitting and testing of adjusted genotype-specific effect for each SNP marker. Dichotomous unconditional logistic regression was performed on breast cancer case control status. The regression was performed on two-indicator variables for heterozygote and rare homozygote genotype states if more than 15 copies of the rare homozygote were observed; otherwise a single indicator variable was used to contrast the common homozygote genotype from either of the other possible genotypes. The model was adjusted for: 4 indicator variables for age group at randomization (ages categories 74, with range 55-59 as the referent), an indicator for known hormone replacement therapy status at blood draw, an indicator for unknown hormone replacement therapy status at diagnosis; this variable is included because it was a matching factor, three sets of coefficients to adjust for estimated population stratification corresponding to the top three eigenvectors identified by the principal component analysis. p-value was obtained from a score test with 2 degrees of freedom. # Human genome build: 37 # dbSNP build: 132 # SNP ID: Marker accession # P-value: testing p-value # Chr ID: chromosome # Chr Position: chromosome position # ss2rs: ss to rs orientation. +: same; -: opposite strand. # rs2genome: Orientation of rs flanking sequence to reference genome. +: same orientation, -: opposite. # Allele1: genomic allele 1 # Allele2: genomic allele 2 # pHWE (case): p-value from HWE testing in cases # pHWE (control): p-value from HWE testing in controls # Call rate (case): Call rate for cases # Call rate (control): Call rate for controls # CI low: the lower limit of 95% confidence interval # CI high: the higher limit of 95% confidence interval