Assessment of performance and error metrics (in minutes) for RT prediction models for existing chromatographic methods (CMs) used in the RT-Pred server.


CM Name Size of Dataset Training R2 Training MAE Training MedAE Training RMSE Testing R2 Testing MAE Testing MedAE Testing RMSE |Void Volume| |Retained| % Void volume + retained compounds 3 State Classification Accuracy
AjsUoB 953 0.98 0.16 0 0.33 0.97 0.17 0 0.41 208 5 22.35 0.97
Fem_Long 418 0.99 1.31 0.68 2.16 0.98 1.66 0.51 2.69 84 6 21.53 0.96
Fem_Lipids 70 0.98 0.61 0.1 0.98 0.95 1.09 1.15 1.5 9 4 18.57 0.97
Life_Old 193 0.95 0.15 0.1 0.23 0.92 0.18 0.06 0.31 32 1 17.1 0.89
Life_New 184 0.87 0.32 0.09 0.54 0.88 0.34 0.18 0.5 31 1 17.39 0.95
RIKEN 470 0.85 0.5 0.09 1.15 0.91 0.4 0 0.98 154 1 32.98 0.87
Eawag_XBridgeC18 363 0.96 0.51 0.3 0.79 0.9 0.9 0.76 1.26 17 2 5.23 0.96
BfG_Nts_RP1 912 0.99 0.28 0.19 0.41 0.93 0.78 0.44 1.11 7 3 1.1 1
Hilic_BDD_2 592 0.97 0.51 0.14 1.03 0.92 0.67 0 1.56 180 3 30.91 0.91
UniToyama_Atlantis 143 0.97 0.68 0.52 0.92 0.86 1.52 0.54 2.38 2 2 2.8 0.94
BDD_C18 671 0.98 0.19 0 0.48 0.97 0.2 0 0.48 345 6 52.31 0.93
RPMMFDA 1676 0.98 0.18 0.12 0.27 0.93 0.37 0.24 0.58 75 7 4.89 0.98
SNU_RIKEN_POS 672 0.96 0.29 0.21 0.43 0.94 0.35 0.26 0.52 36 10 6.85 0.89
UFZ_Phenomenex 394 0.97 0.78 0.37 1.27 0.91 1.5 0.49 2.32 12 6 4.57 0.97
MTBLS87 147 0.9 0.85 0.49 1.37 0.8 1.11 0.78 1.78 6 2 5.44 1
KI_GIAR_zic_Hilic_pH2_7 538 0.97 0.32 0.15 0.52 0.9 0.61 0.2 1.06 48 5 9.85 0.93
Meister zic-pHilic 295 0.89 0.56 0.32 0.89 0.87 0.75 0.54 1.02 32 5 12.54 0.9
AjsTestF 995 0.96 0.29 0.1 0.56 0.91 0.43 0.06 0.86 183 17 20.1 0.85
IJM_TEST 349 0.97 0.45 0.16 0.8 0.89 1.11 0.69 1.74 7 3 2.87 1
IPB_Halle 82 0.96 0.34 0.16 0.55 0.93 0.48 0.35 0.71 6 1 8.54 1
Average ± Standard Deviation 505 ± 616 0.95 ± 0.01 0.46 ± 0.13 0.22 ± 0.11 0.78 ± 0.16 0.91 ± 0.03 0.73 ± 0.22 0.36 ± 0.25 1.19 ± 0.21 73.7 ± 90.3 4.5 ± 3.71 14.9 ± 12.86 0.94 ± 0.04



* 10% of data from the datasets collected from PredRet were held out for testing.
Abbreviations: MedAE - Median Absolute Error, MAE - Mean Absolute Error, and RMSE - Root Mean Squared Error