of the metabolome

cellular sample remain unidentified. Furthermore, the biochemical connectivity between such novel metabolites has been impossible to assess. Even well known and intensely studied biosynthetic pathways have proven difficult to survey comprehensively. For example, the last biosynthetic enzyme of mammalian cholesterol biosynth-Opinion TRENDS in Biotechnology Vol.24 No.12to gene expression microarrays, difficult [1–3]. The lack of unifying principles (such as the genetic code) that would assist molecule identification, comparison and causal connection is another important challenge. Hence, although whole-genome information has been used to infer the global metabolic network structure within entire organisms, such efforts have been limited to well-studied areas of metabolism. Moving beyond the traditional biochemical map requires new experimenta