Skip to main content

Table 2 AUC of different methods in simulations A and B. Methods considered include POST, TreeFDR (TF), Single-OTU test (SO) implemented by POST with c=0, DESeq2 (DE), ANCOM-BC (AB) and LinDA (LD), Wilcoxon rank-sum test with proportional data (WR-P) or CLR transformed data (WR-R) for binary outcomes, and Spearman correlation test with proportional data (SC-P) or CLR transformed data (SC-R) for continuous outcomes. The outcome values are simulated assuming no covariate effects and 5 different causal OTU scenarios. Scenarios 1 to 3 consider larger “causal hubs,” each containing about 7–10 causal OTUs; scenario 4 considers smaller causal hubs of 2–3 causal OTUs; scenario 5 considers causal OTUs with random positions in the phylogenetic tree. Methods with the highest AUC are shown in bold

From: Phylogeny-guided microbiome OTU-specific association test (POST)

Simulation Simulation A Simulation B
Outcome Binary outcome Continuous outcome Binary outcome
Method POST TF SO DE AB LD WR-P WR-R POST TF SO DE AB LD SC-P SC-R POST TF SO DE AB LD WR-P WR-R
Large effect size*
 Scenario1 0.80 0.78 0.62 0.69 0.54 0.56 0.51 0.51 0.64 0.52 0.58 0.61 0.57 0.61 0.57 0.61 0.58 0.53 0.55 0.52 0.57 0.59 0.54 0.55
 Scenario2 0.88 0.77 0.71 0.71 0.70 0.67 0.62 0.58 0.66 0.50 0.62 0.65 0.62 0.62 0.57 0.51 0.60 0.48 0.56 0.57 0.61 0.56 0.52 0.51
 Scenario3 0.88 0.78 0.72 0.70 0.64 0.64 0.60 0.60 0.70 0.49 0.62 0.63 0.61 0.62 0.58 0.51 0.63 0.50 0.59 0.55 0.58 0.58 0.57 0.51
 Scenario4 0.78 0.75 0.70 0.69 0.70 0.67 0.62 0.58 0.63 0.48 0.61 0.63 0.58 0.61 0.58 0.52 0.56 0.49 0.56 0.54 0.55 0.55 0.55 0.51
 Scenario5 0.66 0.72 0.67 0.64 0.65 0.63 0.57 0.55 0.59 0.48 0.60 0.61 0.59 0.60 0.54 0.51 0.54 0.50 0.54 0.52 0.57 0.55 0.53 0.50
Small effect size*
 Scenario1 0.72 0.65 0.59 0.62 0.57 0.57 0.53 0.51 0.62 0.48 0.56 0.58 0.56 0.59 0.55 0.56 0.56 0.49 0.53 0.52 0.54 0.57 0.52 0.52
 Scenario2 0.76 0.65 0.61 0.62 0.63 0.62 0.58 0.55 0.62 0.49 0.59 0.61 0.59 0.59 0.54 0.51 0.54 0.49 0.53 0.55 0.56 0.53 0.51 0.50
 Scenario3 0.76 0.65 0.63 0.62 0.62 0.60 0.57 0.55 0.66 0.49 0.60 0.60 0.58 0.60 0.56 0.50 0.59 0.50 0.56 0.55 0.55 0.56 0.54 0.50
 Scenario4 0.69 0.65 0.62 0.62 0.64 0.62 0.58 0.55 0.60 0.48 0.58 0.60 0.57 0.58 0.56 0.51 0.54 0.50 0.53 0.53 0.52 0.54 0.53 0.51
 Scenario5 0.59 0.58 0.59 0.58 0.60 0.58 0.54 0.52 0.56 0.50 0.56 0.57 0.56 0.57 0.53 0.50 0.53 0.50 0.52 0.52 0.53 0.53 0.52 0.50
  1. *For simulation A, small OTU effect size is from N(±1,1) and large OTU effect size is from N(±2,1). For simulations B, small OTU effect size is simulated from N(±0.2, 0.04) and N(±0.3,0.06) for continuous and binary outcomes, respectively; large OTU effect size is simulated from N(±0.5, 0.1) and N(±1,0.2) for continuous and binary outcomes, respectively