| Description |
Quantitative data for metabolite accumulation in the seeds or seed coats of the 14 accessions or species were obtained by UPLC-TQMS. Peaks that showed S/N ratios >30 were selected as the detected metabolites to be used in further analyses. Areas under the selected peaks were converted into logarithms (base 2) after missing values, which appeared when a metabolite was not detected in a sample, were replaced with 0.1. Data were normalized by z -score transformation using the software TM4 MEV (Chu et al. 2008). The resulting data matrix was analyzed using hierarchical clustering based on the Euclidean distance and visualized by MEGA4 (Tamura et al. 2007) as a dendrogram. The family-specifi c metabolites were identifi ed by BL-SOM analysis of the matrix in combination with a model data set consisting of one hypothetical metabolite specifi c to each of the families Brassicaceae, Gramineae and Fabaceae. |