Figure 2.

Dataset generation process. (A) Original count data. (B) Samples are randomly assigned to class A or B. (C) nH1 taxa are randomly selected among the most prevalent ones. (D) Their abundances are multiplied by the fold-change to produce the final count table.