We tested individual SNPs for association by Modified Quasi-Likelihood Score (MQLS) (Thornton and McPeek, 2007) because MQLS accepts genotypes in post-imputation dosage format and can account for subject relatedness by using a kinship matrix calculated from pedigree data. Unscreened Biobank controls were coded as phenotype unknown. We included an estimated sex-weighted 8.9% population AD prevalence derived from population (Hasin et al., 2007) and unpaid Dutch blood donor (Atsma et al., 2011) data to account for lack of control screening. Varying this estimate from 0% to 12% gave a similar p-value distribution for all prevalence estimates. MQLS cannot include covariates. We used a threshold of p≤5×10−8 for genomewide significant (GWS) results. Odds ratios were not calculated due to the non-independence of related case alleles. Secondary analytic approaches for gene-based, network and geneset analyses are described in Supplementary Methods.
