The development of therapies for Amyotrophic Lateral Sclerosis (ALS) has been hindered by the lack of biomarkers for both identifying early disease and for monitoring the effectiveness of drugs. The identification of ALS biomarkers in presymptomatic individuals might also provide clues to the earliest biochemical correlates of the disease. Previous attempts to use plasma metabolites as biomarkers have led to contradictory results, presumably because of heterogeneity in both the underlying genetics and the disease stage in the clinical population. To eliminate these two sources of heterogeneity we have characterized plasma amino acids and other metabolites in the SOD1(G93A) transgenic mouse model for ALS. Presymptomatic SOD1(G93A) mice have significant differences in concentrations of several plasma metabolites compared to wild type animals, most notably in the concentrations of aspartate, cystine/cysteine, and phosphoethanolamine, and in changes indicative of methylation defects. There are significant changes in amino acid compositions between 50 and 70days of age in both the SOD1(G93A) and wild type mice, and several of the age-related and disease-related differences in metabolite concentration were also gender-specific. Many of the SOD1(G93A)-related differences could be altered by treatment of mice with methionine sulfoximine, which extends the lifespan of this mouse, inhibits glutamine synthetase, and modifies brain methylation reactions. These studies show that assaying plasma metabolites can effectively distinguish transgenic mice from wild type, suggesting that one or more plasma metabolites might be useful biomarkers for the disease in humans, especially if genetic and longitudinal analysis is used to reduce population heterogeneity.