Table 2.
The effect of filtering method on different α-diversity metrics.
| Filtering method | Observed OTUs Mean (SE) |
Chao1 Mean (SE) |
Shannon Mean (SE) |
Inverse Simpson Mean (SE) |
|---|---|---|---|---|
| No filtering | ||||
| All OTUs included | 266 (4.5) | 380.9 (7.6) | 3.243 (0.02) | 13.289 (0.49) |
| Based on OTU abundance in whole dataset | ||||
| Filter OTUs with <0.1% abundance in dataset | 65 (0.6)** | 67.9 (0.7)** | 2.922 (0.03)** | 11.299 (0.38)* |
| Filter OTUs without >10 reads in at least one sample in dataset | 189 (2.1)** | 219.0 (2.6)** | 3.218 (0.03) | 13.196 (0.49) |
| Based on OTU abundance in individual sample | ||||
| Filter OTUs with one read | 188 (2.7)** | 188 (2.7)** | 3.2 (0.03) | 13.2 (0.5) |
| Filter OTUs with < 10 reads | 106 (1.3)** | 106 (1.3)** | 3.164 (0.03)* | 12.955 (0.48) |
| Based on OTU abundance in triplicate | ||||
| Filter OTUs without ≥10 reads in at least one of the three replicates | 121 (1.4)** | 121.4 (1.4)** | 3.180 (0.03) | 13.029 (0.48) |
| Filter OTUs if not present in all three replicates | 169 (2.2)** | 182.2 (2.6)** | 3.206 (0.03) | 13.141 (0.49) |
*p<0.05 vs All OTUs included, **p<0.001 vs All OTUs included.