Table 2.
Metabolites prioritised ≥10 times in the training set for type 2 diabetes incidence differentiation
| Name | HMDB-ID | Selected times (out of 100 runs) |
|---|---|---|
| SM C24:0 | HMDB11697 | 100 |
| DAG C36:1 | HMDB07216 | 100 |
| TAG C58:11 | HMDB10531 | 89 |
| 5-Hydroxyindoleacetic acid | HMDB00763 | 86 |
| PC C36:4 | HMDB07983 | 85 |
| 3-Methyladipic acid | HMDB00555 | 81 |
| D-Glucose | HMDB00122 | 76 |
| 2-Aminodipate | HMDB00510 | 71 |
| Isocitrate | HMDB00193 | 64 |
| L-Phenylalanine | HMDB00159 | 49 |
| LPC C18:2 | HMDB10386 | 47 |
| Glycine | HMDB00123 | 46 |
| TAG C52:1 | HMDB05367 | 39 |
| LPC C18:1 | HMDB02815 | 22 |
| TAG C48:1 | HMDB05359 | 21 |
| TAG C48:0 | HMDB05356 | 21 |
| CE C20:3 | HMDB06736 | 13 |
| Taurine | HMDB00251 | 11 |
| TAG C54:8 | HMDB10518 | 11 |
Nineteen metabolites were selected ten or more times out of 100 iterations in the training set with tenfold cross validation at minimum mean cross-validated error
CE, cholesteryl ester; DAG, diacylglycerol; HMDB-ID, Human Metabolome Database identification; LPC, lysophosphatidylcholine; SM, sphingomyelin; PC, phosphatidylcholine