We have compared the power of different methods for detecting linkage disequilibrium using the transmission disequilibrium test for multi- allelic loci. We have developed a method to use information from multiple affects in a family for association analysis. Since we carry out these analyses in regions of the genome where there is evidence for linkage, we developed a method to use simulation to correct for the bias of dependant observations. We determined that using more than one sib/family can improve the power of detecting association. These methods have been applied in order to detect linkage disequilibrium in the regions of chromosome 18 and chromosome 21 identified as being linked to bipolar illness. We have identified a few markers that show some evidence for association. We are attempting to improve the power of genome scans by considering the length of a region of increased haplotype sharing in affected-sib-pairs. We have shown that true linkage peaks are longer than false linkage peaks because there is a length bias due to the fact that the sibs are selected for being identical-by-descent (IBD) at the disease locus. We are attempting to find new criteria for significance based on length of IBD sharing that makes takes advantage of this difference. We are continuing to compare the power of different methods of non- parametric linkage analysis on simulated data. We collaborated in developing a new method that simulates the IBD statistic in all affected relative pairs. Association between schizophrenia and the SCA-1 locus on chromosome 6 is present. This is consistent with reports of linkage to that region.
