Abstract
We performed linkage and family-based association analysis across chromosomes 1-22 in Replicates 1-5 of the Genetic Analysis Workshop 15 simulated data. Linkage analysis was performed using the Kong and Cox allele-sharing test as implemented in the program Merlin. Association analysis was performed using the transmission/disequilibrium test (TDT). A region on chromosome 6 was consistently highlighted as showing significant linkage to and association with the disease trait. We focused in on this region and performed fine-mapping using stepwise regression approaches using the case/control and family-based data. In this region, we also applied several new methods, implemented in the computer programs LAMP and Graphminer, respectively, that have recently been proposed for association analysis with family and/or case/control data. All methods confirmed the highly significant associations previously observed. Differentiating between potentially causal single nucleotide polymorphisms (SNPs) and other non-causal loci (associated with disease merely due to linkage disequilibrium) proved to be problematic. However, in most replicates we did identify two SNPs (either SNPs 3437 and 3439 from the dense SNP set, or SNPs 153 and 3437 from the combined non-dense/dense SNP set) that together explain most of the observed disease association in the DR/C locus region, and an additional SNP (3931 or 3933) that accounts for the association 5 cM away at locus D.