Results for Bayesian model cross-validation
Results show 5-fold and LOO validation with Bayesian models generated with Discovery Studio and open models implemented in the MMDS mobile app.
| Model | Number Active/Total | Reference | Discovery Studio FCFP6 Plus 8 ADME Descriptors | Discovery Studio ECFP6 Only | Open Model ECFP6 Only |
|---|---|---|---|---|---|
| MATE1 | 81/841 | Wittwer et al. (2013) | 0.65 (LOO 0.72) | 0.62 (LOO 0.68) | 0.69 (LOO 0.67) |
| MATE1 | 24/45 | Astorga et al. (2012) | 0.75 (LOO 0.79)a | 0.63 (LOO 0.72) | 0.70 (LOO 0.74) |
| MATE2K | 21/45 | Astorga et al. (2012) | 0.59 (LOO 0.72)a | 0.60 (LOO 0.81) | 0.75 (LOO 0.73) |
| OCT2 | 243/904 | Kido et al. (2011) | 0.80 (LOO 0.81) | 0.76 (LOO 0.79) | 0.79 (LOO 0.79) |
| OCTN2 | 30/49 | Diao et al. (2009, 2010) | 0.85 (LOO 0.89)a | 0.71 (LOO 0.77) | 0.85 (LOO 0.88) |
| ASBT | 17/38 | Zheng et al. (2010) | 0.75 (LOO 0.91)a | 0.73 (LOO 0.87) | 0.74 (LOO 0.76) |
| NTCP | 50/107 | Dong et al. (2013, 2014) | 0.61 (LOO 0.73) | 0.56 (LOO 0.68) | 0.49 (LOO 0.53) |
ADME, absorption, distribution, metabolism, excretion; LOO, leave one out.
↵a Previously published by the authors of this article.