Multi-interval mapping of correlated trait complexes

Employing Monte Carlo simulations, we demonstrate the advantages of multitrait analysis in detection of linked QTL effects within the framework of mixture models. In spite of an increased number of parameters to be estimated, compared to the single-trait formulation, the proposed method allows for an improvement of detection power and estimation precision of linked QTLs in both adjacent or nonadjacent intervals, with coupling and repulsion effects. The results obtained are illustrated by examples based on data of the North American Barley Genome Mapping Project.