Analytical methods

We analyzed the leaves of 79 soybean accessions that could be grown from 80 accessions provided by the World Soybean Core Collection (The Research Center of Genetic Resources, National Agriculture and Food Research Organization, NARO). The accessions numbered GmWMC 011, 018, 024, 159, 163, 176, and 183 were sampled twice, while the remaining strains were sampled once, yielding a total of 86 samples were analyzed. The leaves dried in a ventilation drying device were ground in a mortar. Then, 75% methanol containing 1 μM of 7-hydroxy-5-methylflavone as internal standard was added at a ratio of 40 times the volume of the dry weight.
Other extraction, instrumental analysis, and data analysis methods are described in the help page of Thing Metabolome Repository and the following paper.

The Thing Metabolome Repository family (XMRs): comparable untargeted metabolome databases for analyzing sample-specific unknown metabolites.
Sakurai N, Yamazaki S, Suda K, Hosoki A, Akimoto N, Takahashi H, Shibata D and Aoki Y
Nucleic Acids Research Database Issue) 51 (D1): D660-D677 (2023)
DOI: 10.1093/nar/gkac1058

Note: There was an error in the mass spectrometry condition described on the help page and the paper, as follows:
Regarding the Release condition for Active Exclusion, 0.2 min (corrected) was applied instead of 0.3 min (incorrect).

Data compatibility with the related databases

The data provided on this website (Soybean Metabolome Repository, SoyMR) were obtained under the same conditions as Thing Metabolome Repository (ThingMR). Therefore, users can compare peak data by mass values and retention times in LC. On the other hand, the data in Food Metabolome Repository (FoodMR) and Plant Metabolome Repository (PlantMR) were obtained with different instruments and condition settings. Therefore, the mass accuracy and retention time of a compound differ. Nevertheless, users can compare the data between the databases by using a larger mass tolerance (20 ppm) applied in SoyMR/ThingMR in default, and a converted retention time that was approximated by a linear regression equation estimated by us (see the above-mentioned paper for the details). In the detailed information page of a peak, you can search for the existence of the same/similar peaks in other databases by clicking the search button for them.