Table 8. Ranks and statistical significance on BAliBASE and PREFAB.
| MUSCLE | MUSCLE-p | T-Coffee | NWNSI | FFTNS1 | CLUSTALW | |
|---|---|---|---|---|---|---|
| MUSCLE | +0.001 | (0.15) | + 0.005 | +2 × 10–6 | +0.0002 | |
| MUSCLE-p | –6 × 10–6 | (0.3) | (0.7) | +0.0002 | +0.02 | |
| T-Coffee | –0.0002 | (0.4) | (0.55) | +7 × 10–5 | +0.01 | |
| NWNSI | –<10–10 | –<10–10 | –10–7 | +0.0001 | (0.06) | |
| FFTNS1 | –<10–10 | –<10–10 | –<10–10 | –<10–10 | –0.04 | |
| CLUSTALW | –<10–10 | –<10–10 | –<10–10 | –<10–10 | –0.008 |
Each entry in the table contains the P-value assigned by a Friedman rank test to the difference between a pair of methods. The upper-right corner of the matrix is obtained from Q scores on BAliBASE, the lower-left corner from Q scores on PREFAB. If the method to the left is ranked higher than the method above, the P-value is preceded by +. If the method to the left is ranked lower, the P-value is preceded by –. If the P-value is >0.05, the difference is not considered significant and is shown in parentheses. So, for example, MUSCLE ranks higher than T-Coffee on PREFAB with P = 0.0002 and MUSCLE-p higher than CLUSTALW on BAliBASE with P = 0.02.