the microsatellite genotypic information from families for linkage analysis together with SNP data from the offspring who are genotyped for both linkage and GWAS, we can “infer” the transmission of SNPs for the rest of the family members who have not been genotyped on SNPs. We call these family members the “missing individuals”. In detail, first we generate the combined pedigrees in which each individual has both the linkage and GWAS genotype data filled in. Genotypes that those individuals do not have will be taken as missing data in the combined pedigrees. Then we use the program MERLIN [55] to read these combined pedigrees as input and infer the dosage probabilities of dense SNP genotypes for these missing individuals (see section Genotype Inference of Familial Individuals for more details). All the trio combinations from the inferred pedigrees are extracted on the condition that the children were affected and at least one parent in the trio was genotyped on microsatellite markers. The dosage-TDT that we have developed in this study is applied on these trio pedigrees using their inferred dosage probabilities. By incorporating the family linkage information into the GWAS data, we can potentially have higher power to detect association between
