Linkage disequilibrium can be used for identifying associations between traits of interest and genetic markers. This study used mapped diversity array technology (DArT) markers to find associations with resistance to stem rust, leaf rust, yellow rust, and powdery mildew, plus grain yield in five historical wheat international multienvironment trials from the International Maize and Wheat Improvement Center (CIMMYT). Two linear mixed models were used to assess marker-trait associations incorporating information on population structure and covariance between relatives. An integrated map containing 813 DArT markers and 831 other markers was constructed. Several linkage disequilibrium clusters bearing multiple host plant resistance genes were found. Most of the associated markers were found in genomic regions where previous reports had found genes or quantitative trait loci (QTL) influencing the same traits, providing an independent validation of this approach. In addition, many new chromosome regions for disease resistance and grain yield were identified in the wheat genome. Phenotyping across up to 60 environments and years allowed modeling of genotype 3 environment interaction, thereby making possible the identification of markers contributing to both additive and additive 3 additive interaction effects of traits.A useful new tool for crop genetic improvement is the identification of polymorphic markers associated with phenotypic variation for important traits by means of linkage disequilibrium (LD) between loci ( Thornsberry et al. 2001;Flint-Garcia et al. 2003). A major advantage of this approach over conventional linkage mapping is that it does not require the timeconsuming and expensive generation of specific genetic populations. LD is determined by the physical distance of the loci across chromosomes and has proven useful for dissecting complex traits because it offers fine-scale mapping due to the inclusion of historical recombination (Lynch and Walsh 1998). However, false positive correlation between markers and traits can arise in the absence of physical proximity due to population structure caused by admixture, mating system, and genetic drift or by artificial or natural selection during evolution, domestication, or plant improvement ( Jannink and Walsh 2002). False associations can also be caused by alleles occurring at very low frequencies in the initial population (Breseghello and Sorrells 2006a,b). These factors create LD between loci that are not physically linked and cause a high rate of false positives when relating polymorphic markers to phenotypic trait variation. Thus, separating LD due to physical linkage from LD due to population structure is a critical prerequisite in association analyses.Population structure can be quantified using Bayesian analysis, which has been effective for assigning individuals to subpopulations (Q matrix) using unlinked markers (Pritchard et al. 2000). Other multivariate statistical analyses such as classification (clustering) and ordination (scaling) can al...