Metabolite profiling has been used to assess the potential for unintended composition changes in potato (Solanum tuberosum L. cv. Desirée) tubers, which have been genetically modified (GM) to reduce glycoalkaloid content, via the independent down-regulation of three genes SGT1, SGT2 and SGT3 known to be involved in glycoalkaloid biosynthesis. Differences between the three groups of antisense lines and control lines were assessed using liquid chromatography–mass spectrometry (LC–MS) and gas chromatography (GC)–MS, and data analysed using principal component analysis and analysis of variance. Compared with the wild-type (WT) control, LC–MS revealed not only the expected changes in specific glycoalkaloid levels in the GM lines, but also significant changes in several other metabolites, some of which were explicable in terms of known pathways. Analysis of polar and non-polar metabolites by GC–MS revealed other significant (unintended) differences between SGT lines and the WT, but also between the WT control and other control lines used.
