Table 3.
Evaluation of RNAi screens as a test set
| Datatype_kernel | Signal-to-noise ratio |
|---|---|
| micro_linear | 1.44 |
| micro_poly | 1.00 |
| micro_RBF | 3.91 |
| chip_binary_linear | 1.36 |
| chip_binary_poly | 1.84 |
| chip_binary_RBF | 1.88 |
| chip_contin_linear | 1.57 |
| chip_contin_poly | 1.86 |
| chip_contin_RBF | 1.91 |
| weight_binary_linear | 1.80 |
| weight_binary_poly | 1.96 |
| weight_binary_RBF | 3.82 |
| weight_contin_linear | 1.84 |
| weight_contin_poly | 1.65 |
| weight_contin_RBF | 2.35 |
| simple_binary_linear | 1.80 |
| simple_binary_poly | 3.43 |
| simple_binary_RBF | 3.80 |
| simple_contin_linear | 1.84 |
| simple_contin_poly | 1.38 |
| simple_contin_RBF | 3.21 |
Ranked methods based on signal-to-noise ratio performance of predicting the percentage of genes as positive from the positive test set (self-renewal screen) and as positive from the negative test set (insulin-pathway screen).