TABLE 2.
Top 15 significant pseudo-ions selected by statistical methods.
| Rank | Pseudo-ion | m/z | Mean difference of peak intensity ratio (10−4) | log2(fc) | Wilcoxon rank sum test | F-test | KS test | Observation times |
| 1 | Pseudo-ion 136 | (4,700, 4,720) | 10.22 | 0.88 | 1.24×10−38 | ≤0.01 | ≤0.01 | 3,200 |
| 2 | Pseudo-ion 5 | (2,080, 2,100) | −50.75 | −0.43 | 1.02×10−51 | ≤0.01 | ≤0.01 | 12,528 |
| 3 | Pseudo-ion 27 | (2,520, 2,540) | −13.63 | −0.41 | 1.73×10−64 | ≤0.01 | ≤0.01 | 10,704 |
| 4 | Pseudo-ion 95 | (3,880, 3,900) | −3.51 | −0.61 | 2.40×10−10 | ≤0.01 | ≤0.01 | 3,358 |
| 5 | Pseudo-ion 353 | (9,040, 9,060) | −1.03 | −0.77 | 5.60×10−05 | ≤0.01 | ≤0.01 | 478 |
| 6 | Pseudo-ion 1 | (2,000, 2,020) | −13.52 | −0.40 | 1.72×10−41 | ≤0.01 | ≤0.01 | 10,727 |
| 7 | Pseudo-ion 32 | (2,620, 2,640) | −13.09 | −0.40 | 1.01×10−34 | ≤0.01 | ≤0.01 | 10,453 |
| 8 | Pseudo-ion 293 | (7,840, 7,860) | −0.26 | −0.67 | 3.20×10−07 | ≤0.01 | ≤0.01 | 406 |
| 9 | Pseudo-ion 284 | (7,660, 7,680) | −4.76 | −0.72 | 1.87×10−03 | ≤0.01 | ≤0.01 | 2,873 |
| 10 | Pseudo-ion 11 | (2,200, 2,220) | −13.45 | −0.37 | 6.23×10−42 | ≤0.01 | ≤0.01 | 10,405 |
| 11 | Pseudo-ion 2 | (2,020, 2,040) | −20.51 | −0.38 | 9.96×10−34 | ≤0.01 | ≤0.01 | 11,433 |
| 12 | Pseudo-ion 4 | (2,060, 2,080) | −97.53 | −0.39 | 2.46×10−23 | ≤0.01 | ≤0.01 | 12,318 |
| 13 | Pseudo-ion 367 | (9,320, 9,340) | 0.01 | 1.11 | 3.86×10−04 | ≤0.01 | 0.01 | 285 |
| 14 | Pseudo-ion 50 | (2,980, 3,000) | −12.41 | −0.35 | 6.23×10−33 | ≤0.01 | 0.01 | 10,574 |
| 15 | Pseudo-ion 163 | (3,240, 3,260) | −3.71 | −0.84 | 2.46×10−09 | ≤0.01 | 0.01 | 768 |
Mean difference is calculated by CIRKP-CISKP;fc represents the fold change value; fold change is calculated by CIRKP/CSIKP; total number of training samples: 13,414. Bold type values means the statistical quality of these pseudo-ions are relatively lower than other pseudo-ions since less samples are observed.