When pooling genotype data from different platforms, investigators could impute the SNPs missing on each platform to get a data set with comparable variables. This approach has been suggested as a way of combining study cases and controls with publicly available controls genotyped on a different chip (Zhuang et al. 2010). Fallin et al. (2010) used imputation to combine their case-control study, genotyped on Illumina, with a publicly available case-control study genotyped on Affy. A number of imputation methods exist, and they have been shown to be very accurate in the typical setting where cases and controls are genotyped together on the same platform (Li et al. 2010; Howie et al. 2009). However, their performance in the setting we are discussing here, when cases and controls have been genotyped on different platforms, has been largely unexplored.
