Extended pedigree resources ascertained for disease, and previously used for linkage analysis, contain individuals with high likelihood of being genetic in nature. Sampling individuals from these pedigrees for association or other linkage-disequilibrium (LD) based methods therefore increases the likelihood that genetic cases are selected and thus increases the power to detect such genetic factors involved in the disease. Unfortunately, the relatedness of these individuals invalidates standard statistical analysis techniques, which can lead to inflated type I errors. Methods that appropriately account and correct for the inherent bias in using correlated data will allow for maximal and powerful use of already ascertained pedigree resources for LD-based analyses. It is already clear that many small effect genes, in addition to environmental factors, are involved in common disease, and that both inter-genic and intra-genic epistatic interactions exist. Strategies to efficiently test complex interaction hypotheses, in conjunction with appropriate corrections for multiple testing are needed. Knowledge of underlying haplotype blocks will help minimize the number of tests, but efficient sequential methods will still be required to maximize power, especially when a-priori knowledge of interactions is absent, as is usually the case.
We aim to develop a flexible analysis tool and distribute a user-friendly, freely available software package that incorporates a broad range of statistical tests and strategies to test complex interactions in pedigree data. With the availability of such software it is anticipated that many researchers, including our own Genetic Epidemiology group at the University of Utah, with already ascertained resources will be able to begin new analyses and that the resources will gain new, previously unrealized, value.
