Skip to main content
. 2022 Jun 21;23:243. doi: 10.1186/s12859-022-04786-9

Table 1.

Block-level identification rate over 500 replications for four true blocks under three graph structures

Method Graph type Block 1 Block 2 Block 3 Block 4 Average (SD)
STB Random 0.8289 0.9918 1 0.8598 0.9201 (0.0885)
STC 0.9401 1 0.6736 1 0.9034 (0.1558)
ZIBR 0.8223 0.8595 0.9814 0.9835 0.9117 (0.0831)
ANCOM-BC 1 1 1 1 1 (0)
Hu et al. [14] 0.0020 0.4633 0.1388 0 0.1510 (0.2181)
STB Hub 0.7478 0.7917 1 0.9364 0.8690 (0.1188)
STC 0.9670 1 1 1 0.9917 (0.0165)
ZIBR 0.6865 0.7726 1 1 0.8648 (0.1601)
ANCOM-BC 1 1 1 1 1 (0)
Hu et al. (2020) 0.0082 0.2238 0.1458 0 0.0944 (0.1091)
STB Cluster 0.7868 1 1 1 0.9467 (0.1066)
STC 0.9510 1 0.9787 0.7975 0.9318 (0.0918)
ZIBR 0.7361 0.7961 0.7682 0.9506 0.8128 (0.0951)
ANCOM-BC 1 1 1 1 1 (0)
Hu et al. [14] 0.0040 0.3300 0.1360 0 0.1175 (0.1551)

Ω is set to reflect the correlations among OTUs. Random graph structure indicates that OTUs are correlated with each other randomly. The hub and cluster graphs capture some aspects of biological networks, such as highly connected nodes and community structure. More details can be found in Osborne et al. [20]