We present a methodology for the detection of SNP-SNP interactions in case-parent trios. Inference is based on an extension of the logic regression methodology (Ruczinski et al., 2003), embedded in a conditional logistic regression framework. Missing data are addressed using a haplotype-based imputation algorithm. We document open source software to generate high-order SNP-SNP interactions and to impute missing genotypes among case-parent trios using a novel and efficient haplotype and mating table based algorithm. The proposed trio logic regression is applied to data from a Schizophrenia study (Fallin et al., 2005). In addition, for case-control studies, we propose a framework to incorporate genotype uncertainties for any genetic model. Via a simulation study, we demonstrate that the power to detect a true association can be substantially increased for difficult to call genotypes. |