Generally speaking, different populations have different linkage disequilibrium structures and different allele frequencies that lead to different IQS values. A mixture of different populations will make the IQS sensitive to the ratio of population mixture. Therefore, as in general association studies, a mixture of different populations should be avoided. However, African Americans have a unique and relatively stable genetic structure. The IQS score from African Americans is stable in our study and is useful to filter out poorly imputed SNPs.
