We chose to adopt the posterior mode approach and make discrete imputed genotype calls. By setting a suitably stringent threshold, we were able to reduce the effect of genotype uncertainty. Under this approach, it is straightforward to test for Hardy-Weinberg equilibrium as a quality control filter for genotype-specific imputation failure at untyped SNPs. We reiterate the important requirement that replication of significant association for imputed SNPs should always include experimental genotyping in an independent data set [Browning, 2008]. This requirement addresses the possibility that unaccounted for uncertainty in imputed calls may inflate the false positive error rate of association testing.
