Lipidomics is a relatively new area, and there is still no large-scale database for lipid profiles derived from biological experiments.  We are building a relational lipidomics database called LipidomeDB.  It contains extensive metadata as well as data on many lipid metabolites as a function of treatment or genotype. Much additional information about lipid molecular species, their mass spectral properties, and their interrelationships is also being stored in the database.  LipidomeDB also provides for continuous annotation of large datasets containing spectral information about many identified and yet-to-be-identified lipid metabolites.  Having lipidomics data in a database will allow comparison of data from many experiments; it will also allow adaptation of statistical applications to lipidomics data. We plan to have the database online by the end of 2007 and then to continue to update and to add applications.