We also would like to emphasize that we are dealing with the extreme situation when cases and controls are genotyped on different platforms. The elevated false positive rates are not explicitly reported in the literature, as most groups do not have this problem because of the study design. But many groups have noticed it. In a recent paper by de Bakker[18], the author noted “the dangers of combining cases genotyped on one platform and controls genotyped on another” (Page 124). In the GENEVA consortium, there is a consensus that genotypes imputed from one array should not be combined with imputed genotypes from another array.
