The results of further SNP and sample exclusions are presented in Table 1. As in the primary report from the IMAGE GWAS, 5 the majority of additional marker exclusions were made on the basis of call rate (i.e. the proportion of samples successfully genotyped for that SNP) conditional on minor allele frequency (MAF; i.e. the frequency of the least common allele for the SNP) (Figure S1, available online). Because rarer markers have a greater possibility of misclassification that would be more likely to bias tests of association, 11–13 we included SNPs with 0.01 ≤ MAF < 0.05 and call rate >99%; 0.05 ≤ MAF <0.1 and call rate >97%, MAF ≥ 0.1 and call rate >95%. Any SNPs found to be out of Hardy-Weinberg Equilibrium (p<1.0E-6) in founders were excluded from further consideration. We checked for sample duplication by examining identity-by-state for all pairs of individuals and found none. Probably owing to the extensive filtering of SNPs by Genizon and removed in the first step of data management, there were few additional marker or sample exclusions based on Mendelian errors,
