New criteria for linkage detection have been developed based on the fact that a true linkage peak is longer than a false peak. We have shown that requiring two """"""""significant"""""""" linkage statistics in a 10cM region has better power to detect linkage than requiring one significant statistic at a lower p-value. Nuclear family structures may be more powerful for detecting linkage to common disease genes than extended pedigree structures. For rare genes, extended pedigree analyses are better. Thus, it may be advantageous to break pedigrees into nuclear family structures when analyzing them for linkage. Multiplex sibships contain additional information about linkage disequilibrium compared to single offspring when the disease gene is common. We are comparing different statistics for linkage disequilibrium detection and different sampling schemes in order to determine the best way to narrow down regions containing disease susceptibility genes detected by linkage analysis.
