Bayesian oligogenic analysis of quantitative and qualitative traits in general pedigrees

A Bayesian method for multipoint oligogenic analysis of quantitative and qu alitative traits is presented. This method can be applied to general pedigr ees, which do not necessarily have to be "peelable" and can have large numb ers of markers. The number of quantitative/qualitative trait loci (QTL), th eir map positions in the genome, and phenotypic effects (mode of inheritanc es) are all estimated simultaneously within the same framework. The summari es of the estimated parameters are based on the marginal posterior distribu tions that are obtained through Markov chain Monte Carlo (MCMC) methods. Th e method uses founder alleles together with segregation indicators in order to determine the genotypes of the trait loci of all individuals in the ped igree. To improve mixing properties of the sampler, we propose (1) joint sa mpling of map position and segregation indicators, (2) omitting data augmen tation for untyped or uninformative markers (homozygous parent), and (3) up dating several markers jointly within a single block. The performance of th e method was tested with two replicate GAW10 data sets (considering two lev els of available marker information). The results were concordant and simil ar to those presented earlier with other methods. These analyses clearly il lustrate the utility and wide applicability of the method. Genet. Epidemiol . 21:224-242, 2001. (C) 2001 Wiley-Liss, Inc.