Data were excluded from one individual (an ATR control) whose DNA did not genotype successfully. Initial quality control found 23 SNPs for which genotyping failed (i.e., GenCall score=0). An additional 9 SNPs that had a call rate below 95% were removed from further analyses. Genotypic data revealed the presence of 32 duplicate samples; further scrutiny found that these individuals had participated in the project more than once. For individuals who had participated both in the pilot and funded phases of the project, interview data from the funded phase were retained. For those who participated multiple times in the funded project, data from the first interview were used for analyses (case-control status was consistent throughout). Data from an additional 3 participants were excluded on the basis of mismatch of genotypic and phenotypic gender. Data from an additional 47 SNPs with minor allele frequency (MAF) less than 2% were not included in analyses. 27 SNPs were removed from further analyses because deviations from Hardy–Weinberg equilibrium in ATR controls exceeded Bonferroni correction for the total number of remaining SNPs (0.05/1457 = 3.43 ×
