|
Sample size = 600
| | | | | |
|---|
| |
Type I error
|
Power
|
|---|
| |
Null
|
Setting 1
|
Setting 2
|
Setting 3
|
|---|
|
α-level
|
0.05
|
0.01
|
0.05
|
0.05
|
0.05
|
|---|
|
Rarefaction
| | | | | |
|
Linear regression
|
0.0547
|
0.0084
|
0.9949
|
0.6928
|
0.1247
|
|
ZIP
|
0.7387
|
0.6622
|
1.0000 ∗
|
0.9742 ∗
|
0.7720 ∗
|
|
ZINB
|
0.2019
|
0.0771
|
0.9812 ∗
|
0.7321 ∗
|
0.2398 ∗
|
|
ZINQ-MinP
|
0.0526
|
0.0106
|
0.9994
|
0.8557
|
0.1508
|
|
ZINQ-Cauchy
|
0.0580
|
0.0110
|
0.9991
|
0.8346
|
0.1493
|
|
TSS
| | | | | |
|
Linear regression
|
0.0536
|
0.0088
|
0.9970
|
0.7425
|
0.1320
|
|
ZIB
|
0.0110
|
0.0017
|
0.9964 +
|
0.6255 +
|
0.0305 +
|
|
Tobit
|
0.0543
|
0.0099
|
0.9989
|
0.8041
|
0.1467
|
|
ZIlogN
|
0.6992
|
0.6872
|
1.0000 ∗
|
1.0000 ∗
|
0.9999 ∗
|
|
ZIG
|
0.0548
|
0.0102
|
0.9961
|
0.7264
|
0.1196
|
|
ZINQ-MinP
|
0.0501
|
0.0101
|
0.9995
|
0.9096
|
0.1669
|
|
ZINQ-Cauchy
|
0.0503
|
0.0103
|
0.9994
|
0.8981
|
0.1555
|
|
CSS
| | | | | |
|
Linear regression
|
0.0527
|
0.0113
|
0.9995
|
0.8934
|
0.1733
|
|
Tobit
|
0.0526
|
0.0110
|
0.9985
|
0.8597
|
0.1628
|
|
ZIlogN
|
0.0475
|
0.0095
|
0.9996
|
0.8794
|
0.1464
|
|
ZIG
|
0.0494
|
0.0096
|
0.9998
|
0.8850
|
0.1474
|
|
ZINQ-MinP
|
0.0501
|
0.0103
|
0.9993
|
0.8852
|
0.1520
|
|
ZINQ-Cauchy
|
0.0505
|
0.0095
|
0.9991
|
0.8735
|
0.1524
|
- Setting 1: 100% from HBP edf for HBP samples;
- Setting 2: 80% from HBP edf and 20% from non-HBP edf for HBP samples;
- Setting 3: 60% from HBP edf and 40% from non-HBP edf for HBP samples.
- ∗: power of a method that inflates type I error
- +: power of a method that deflates type I error
- Results by the various methods on 10000 simulated datasets by generating samples from the edf of Anaerovorax’s normalized abundance, including type I error control and power under different settings with significance cutoffs 0.05 and 0.01
