Gravar-mail: Bayesian Variable Selection in Searching for Additive and Dominant Effects in Genome-Wide Data