Krumsiek, J. et al. PLoS Genet. 8, e1003005 (2012).

The genome influences metabolite levels, and for a limited number of known metabolites, genome-wide association studies (GWAS) have shed light on this connection. A large number of metabolites, however, have not been chemically identified even though they are reproducibly detected by mass spectrometry methods. Their interaction with the genome is thus not investigated in classical GWAS. Krumsiek et al. seek to use the GWAS approach to identify unknown metabolites. They mapped the associations between SNPs, known metabolites and over 200 unknown metabolites to elucidate the pathways likely involved and to derive a hypothesis as to a candidate metabolite's identity that could then be experimentally tested. This approach allowed them to generate hypotheses for over 100 metabolites and confirm the identity of nine of them.