Table 7.
Comparison with Literature Results on the Ames Toxicity Benchmark
| Encoding | AUC ROCa |
|---|---|
| ECFP | 0.87 ± 0.01 |
| LSTAR | 0.86 ± 0.01 |
| SVM-dragonXb | 0:86 ± 0.01 |
| RAD2D | 0.85 ± 0.02 |
| RAD3D | 0.85 ± 0.01 |
| ASP | 0.85 ± 0.02 |
| GP-dragonXb | 0.84 ± 0.01 |
| AT2D | 0.83 ± 0.01 |
| RF-dragonXb | 0.83 ± 0.01 |
| AT3D | 0.81 ± 0.01 |
| kNN-dragonXb | 0.79 ± 0.01 |
| DFS | 0.78 ± 0.01 |
| AP2D | 0.78 ± 0.03 |
| AP3D | 0.76 ± 0.02 |
Benchmarks for the large Ames toxicity benchmark using LIBLINEAR (nested 5-fold cross-validation on defined splits). The features were hashed to 214 bit sparse binary fingerprints.
a Area under the ROC curve b Reference classifier from the study of Hansen et al. using dragonX 1.2, SVM: support vector machine, GP: Gaussian processes, RF: random decision forest, kNN: k-nearest neighbor