Abstract
This study established an Orthogonal Partial Least Squares (OPLS) model combining (1)H-NMR and GC-MS data to identify characteristic metabolites in complex extracts. Both in metabolomics studies, and natural product chemistry, the reliable identification of marker metabolites usually requires laborious isolation and purification steps, which remains a bottleneck in many studies. Both ginger (GR) and processed ginger (PGR) are listed in the Japanese pharmacopeia. The plant of origin, the rhizome of Zingiber officinale Roscoe, is differently processed for these crude drugs. Notably, the quality of crude drugs is affected by genetic and environmental factors, making it difficult to maintain a certain quality standard. Therefore, characteristic markers for the quality control of GR and PGR are required. Metabolomic analysis using (1)H-NMR was able to discriminate between GR and PGR, but there were unidentified signals that were difficult to distinguish based on NMR data alone. Therefore, we combined (1)H-NMR and GC-MS analytical data to identify them by OPLS. As a result, αr-curcumene was found to be a useful marker for these identifications. This new approach enabled rapid identification of characteristic marker compounds and reduced the labor involved in the isolation process.