(1)H NMR measurements on protein-free extracts from wheat leaf and stem tissue were used to investigate the biochemical correlates of partial resistance to fungal species implicated in the Fusarium head blight (FHB) disease complex. The wheat genotypes included five commercial wheat cultivars and 116 wheat genotypes, from the CIMMYT international FHB breeding programme in Mexico, that had been bred for FHB disease resistance, utilizing exotic, typically Chinese, resistance sources. Principal component analysis of the nuclear magnetic resonance (NMR)-derived metabolite profiles revealed distinct separation of the commercial wheat cultivars from the majority of the CIMMYT wheat genotypes with the commercial cultivars exhibiting markedly higher carbohydrate concentrations. A cross-validated partial least squares (PLS) regression model of the metabolite profile against the partial disease resistance component latent period (delay in sporulation of the fungus on the plant tissue) predicted latent periods that were significantly correlated with the experimentally determined values (R(2) = 0.34, P < 0.001). Identified metabolites that were found in plants with shorter latent periods (higher disease susceptibility) included choline, the single most influential metabolite in the PLS model, betaine, the amino acids glutamine, glutamate and alanine, trans-aconitate and sucrose. Metabolites related to increased disease resistance included glucose and unassigned resonances in the carbohydrate and aromatic regions of the NMR spectra. The current study has demonstrated the potential of metabolite profiling as a tool for marker-assisted selection in commercial breeding for resistance to FHB in wheat.