Metabolism provides a highly amplified reflection of the biochemistry of living creatures. Careful analysis of the complex metabolic signatures and temporal changes may provide essential information about disease onset, progression, and may provide a possibility of early prediction of the outcome. In addition, assessment of component-level contributions to this information allows identification of biomarkers and molecular-level analysis of the relevant biological processes.

We processed previously collected 1H NMR data of blood sera of pregnant mares grazing on endophyte-infected fescue compared to those grazing on endophyte-free fescue pastures. Complete spectra (with water suppression only), also relaxation filtered and diffusion filtered spectra to highlight small and larger molecular components, respectively, of the same samples were all evaluated. Principle component analysis (PCA) reveals specific clustering of the samples according to disease and its outcome. Loading vectors along selected PCs help us to identify contributing components, including both small metabolites and larger species, such as various lipids.