The dTDT makes it possible to process the inferred dosage probabilities of the un-genotyped SNPs for those missing individuals in a nuclear family. As a follow-up study of this generalized TDT approach, we develop a simulation to investigate how the power changes for association detection with different inputs. In this simulation, we generate multiple sets of trios under various settings. Each set has 1,000 trios. Each trio has one affected child. Because we focus on the interaction between SNP and microsatellite markers, only one SNP marker and one microsatellite marker are simulated. In each trio, the microsatellite markers are assigned to both the parents and the child. SNPs are only assigned to the child. The parents who do not have such SNP markers are considered as the missing individuals and their SNP genotypes are inferred by MERLIN. The dTDT is then used to process the inferred dosage probabilities and p-values are reported from the χ 2 statistics.
