A Bayes Factor Approach to Discriminate between Linked or Pleiotropic Qtl

INTRODUCTION Most of the experiments to map QTLs involve recording of several traits, often genetically correlated between them. Several authors have tried to solve the problem of discriminating between linked and pleiotropic QTLs (Cheverud et al., 1997 ; Almasy et al., 1997 ; Lebreton et al., 1998 ; Knott and Haley, 2000). From a bayesian point of view, the comparison between alternative models is solved by the calculation of Bayes Factors (Kass and Raftery, 1995). The Bayes Factor is the ratio between the marginal probabilities of data under both models after integrating out all the parameters. The objective of this paper is to propose a Bayes Factor approach for testing between linked and pleiotropic QTL. First, we present the general Bayes Factor, and, secondly, two particular examples from an experimental F2 cross between outbred lines of pigs are presented.