Table 1 Comparison of t-statistics and Welch statistics in identifying significant genes under different values of Δs
From: Assessment of differential gene expression in human peripheral nerve injury
Δ | Significant genes | False positives | FDR |
|---|---|---|---|
t-statistics | |||
2 | 622 | 312 | 50% |
4 | 223 | 77.3 | 35% |
6 | 131 | 35.1 | 27% |
8 | 88 | 21.2 | 24% |
12 | 73 | 13.3 | 18% |
16 | 62 | 9.8 | 16% |
20 | 57 | 7.7 | 14% |
25 | 46 | 5.8 | 13% |
40 | 30 | 3.4 | 11% |
60 | 27 | 2.4 | 9% |
90 | 19 | 1.5 | 8% |
Welch statistics | |||
2 | 1348 | 648.3 | 48% |
4 | 525 | 155.9 | 30% |
6 | 264 | 64.5 | 24% |
8 | 97 | 24.0 | 25% |
12 | 50 | 10.4 | 21% |
16 | 20 | 4.0 | 20% |
20 | 12 | 2.5 | 21% |
25 | 7 | 1.0 | 14% |
40 | 7 | 1.0 | 14% |
60 | 6 | 0.4 | 7% |
90 | 6 | 0.4 | 7% |