OPTIMAL STRATEGIES FOR MAPPING COMPLEX DISEASES IN THE PRESENCE OF MULTIPLE LOCI

Recent advances in genome technology have led to mapping and subsequen t isolation, by positional cloning, of a number of genes for common an d/or complex human diseases. It therefore will be possible to utilize information about a known locus in the search for additional, perhaps less penetrant, genes for a particular disease. It is also unclear, un der these situations, what the optimal sampling strategy should be. To address these questions, we have calculated the expected LOD score fo r localizing one locus in a variety of two-locus models of disease, fo r four different pedigree structures, and under three different scenar ios regarding knowledge/testing of one of the two loci. These design c onsiderations are evaluated by use of a cost function that incorporate s the costs of ascertaining different family structures, the relative costs of genotyping and mutation testing family members, and the amoun t of information provided by each family structure and testing scenari o. The results indicate that, in most cases, affected sib pairs are a particularly poor strategy, especially when linkage or mutation data a re available at the known locus. We also demonstrate that prescreening the sample of families for mutations at known susceptibility loci is, in general, a cost-effective strategy.