Table 2.
Sample means (standard deviations in parentheses) of the total number of selected OTUs (Total), and of the numbers of true positives and false positives
| Method | Total | TP | FP | |
|---|---|---|---|---|
| X ⊥ ⊥Z | fitZig | 157.0 (49.4) | 12.3 (4.7) | 144.6 (45.7) |
| DESeq2 | 20.4 (4.8) | 3.4 (1.5) | 17.0 (4.4) | |
| KW | 25.8 (6.1) | 3.5 (1.6) | 22.3 (5.7) | |
| aMiSPU with α 1=.7 | 234.1 (241.5) | 42.6 (18.5) | 191.6 (234.8) | |
| aMiSPU with k 1=1 | 58.5 (76.9) | 27.2 (20.6) | 31.3 (78.1) |
Based on 1000 simulation replications under scenario 1, by fitZig [32], DESeq2 [31], KW test, aMiSPU with α 1=.7, or aMiSPU with k 1=1. For fitZig, DESeq2 and KW test, cutoff 5×10−5, 0.05, 0.05 were chosen, respectively
FP number of false positives, KW Kruskal–Wallis test, OTU operational taxonomic unit, TP number of true positives