Fig. 2From: An empirical Bayes approach to normalization and differential abundance testing for microbiome dataComparison of recall and precision with data from DM across different θ. To detect differentially abundant taxa, we simulated 100 data sets from the DM model with β=0.25 and θ∈{0.05,0.1,0.15,0.2,0.25,0.3}. a and b Recall of t-test and Wilcoxon rank sum test with various normalization methods. c and d Recall and precision of DESeq2, ANCOM, metagenomeSeq, Wrench, and those of t-test and Wilcoxon rank sum test, both applied after counts were normalized by eBayBack to article page