The Framingham Heart Study simulated data set (Problem 3, replication 1, with answers) was used in this analysis. Founders and singletons from the original pedigree file (N = 849) were selected for this case-control analysis, using myocardial infarction (event 1) as the disease status, resulting in 133 cases and 716 controls. SNPs were eligible for selection if they were located in genomic regions not associated with myocardial infarction (null SNPs), of which 50 were selected by MAF and percent missing data. Ten SNPs, each at lower MAFs of 1% and 5%, and five SNPs, each at MAFs of 10%, 25%, and 50%, with similar low rates of missing data (approximately 0%) were selected. To examine effects of missing data, ten additional SNPs at MAF 5% and five additional SNPs each at MAF 50% with approximately 5% missing data were also analyzed (Table 1).
