Table 2.
oCPLS2 prediction errors of models based on previous (Chighine et al., 2023) and extended dataset validated with a test set of 6 independent samples. (*) selected metabolites excluding Maltose, Mannose, Methanol, creatinine
| PMI range (h) | training set (samples) | SDEP (hours) |
|---|---|---|
| 16–170 | Proof-of-concept dataset (n = 18) | 34.0 |
| Proof-of-concept dataset (n = 17) | 34.2 | |
| Extended dataset (n = 17) | 41.0 | |
| Proof-of-concept dataset (n = 17) * | 39.5 | |
| Extended dataset (n = 17) * | 42.1 | |
| 16–100 | Proof-of-concept dataset (n = 18) | 13.0 |
| Proof-of-concept dataset (n = 17) | 10.5 | |
| Extended dataset (n = 17) | 12.0 | |
| Proof-of-concept dataset (n = 17) * | 10.1 | |
| Extended dataset (n = 17) * | 11.7 |