Despite recent intensive research efforts in functional genomics, the functions of only a limited number of Arabidopsis (<i>Arabidopsis thaliana</i>) genes have been determined experimentally, and improving gene annotation remains a major challenge in plant science. As metabolite profiling can characterize the metabolomic phenotype of a genetic perturbation in the plant metabolism, it provides clues to the function(s) of genes of interest. We chose 50 Arabidopsis mutants, including a set of characterized and uncharacterized mutants, that resemble wild-type plants. We performedmetabolite profiling of the plants using gas chromatography-mass spectrometry. To make the data set available as an efficient public functional genomics tool for hypothesis generation, we developed the Metabolite Profiling Database for Knock-OutMutants in Arabidopsis (MeKO). It allows the evaluation of whether a mutation affects metabolism during normal plant growth and contains images of mutants, data on differences in metabolite accumulation, and interactive analysis tools. Nonprocessed data, including chromatograms, mass spectra, and experimental metadata, follow the guidelines set by the Metabolomics Standards Initiative and are freely downloadable. Proof-of-concept analysis suggests that MeKO is highly useful for thegeneration of hypotheses for genes of interest and for improving gene annotation.