Table 1.
The number of correctly identified genes as a function of both subspace location and dimension
| k | Δk = 0 | Δk = 1 | Δk = 2 | Δk = 3 | Δk = 4 |
| 1 | 4 (0.85) | ||||
| 2 | 6 (0.53) | 4 (0.85) | |||
| 3 | 5 (0.71) | 6 (0.53) | 5 (0.71) | ||
| 4 | 3 (0.94) | 6 (0.53) | 10 (5.9 × 10-2) | 5 (0.71) | |
| 5 | 8 (0.22) | 14 (1.3 × 10-3) | 13 (3.8 × 10-3) | 13 (3.8 × 10-3) | 8 (0.22) |
| 6 | 12 (1.1 × 10-2) | 13 (3.8 × 10-3) | 17 (2.7 × 10-5) | 17 (2.7 × 10-5) | 5 (0.71) |
| 7 | 9 (0.12) | 14 (1.3 × 10-3) | 17 (2.7 × 10-5) | 14 (1.3 × 10-3) | |
| 8 | 9 (0.12) | 15 (3.9 × 10-4) | 13 (3.8 × 10-3) | ||
| 9 | 8 (0.22) | 10 (5.9 × 10-2) | |||
| 10 | 8 (0.22) |
Arraylets ranging from k-Δk to k+Δk were used to define the subspace. The genes used to define the GSVD are indicated by asterisks in the online Additional Material [see Additional file 1]. The numbers in brackets are the corresponding p-values.