Table 1 A summary of recent models used to predict aqueous solubility
From: A unified approach to inferring chemical compounds with the desired aqueous solubility
S. no | Model | # datasets | Descriptor information (size) | Software | \(\text{R}^2\) Min, Max |
|---|---|---|---|---|---|
1 | RF [3] | 1 | Deterministic 2D, 3D (200) | MOE | 0.89 |
2 | MLR, RF, SVM [4] | 1 | Non-deterministic (21) | Hybot, Dragon, Sybyl, VolSurf | 0.701, 0.736 |
3 | RF [5] | 1 | Non-deterministic (16) | PaDEL-Caret package | 0.82 |
4 | LASSO, PLS, RF, LightGBM [6] | 1 | Non-deterministic (317) | Dragon | N/A |
5 | MLR, RF [7] | 1 | Deterministic 2D, 3D | Mordred package | 0.80, 0.98 |
6 | MLR [8] | 5 | Deterministic 2D, 3D (58) | Sybyl, Amber | 0.4, 0.9 |
7 | MLR [9] | 7 | Deterministic (2, 3, 8)a | Gaussian09 program, Sybyl, BioPPSy | 0.47, 0.87 |
8 | PLS, BPN, SVR [10] | 1 | Deterministic (28) | Dragon | 0.69, 0.735 |
9 | CNN, RNN, DNN, SNN [11] | N/A | Non-deterministic (N/A) | N/A | N/A |
10 | GNN [12] | 1 | Non-deterministic 2D, 3D (839) | Mordred, Pybel, RDKit | 0.76 |
11 | BCSA [13] | 5 | Non-deterministic | Within model | 0.83, 0.88 |
12 | STN [14] | 5 | Deterministic (25) | RDKit | 0.65, 0.89 |
13 | GCNN [15] | 1 | Non-deterministic 2D, 3D (839) | Mordred, Pybel, RDKit | 0.86 |
14 | CS-LightGBM [16] | 1 | Non-deterministic | RDKit | 0.8575 |
15 | SAMPN [17] | 1 | Deterministic | MPN | N/A |