Count-variance plot for each column of A, where A is the data matrix from the mouse bone marrow scRNA-Seq experiment [25] and the analysis workflow of [26].

The color for each column represents whether the column is selected or not by k = 14, ϵ = 0.05 DCSS. The plot also shows the thresholds for count, variance, and index of dispersion with same number of selected columns as DCSS. The columns selected by DCSS are highly variable and have large counts, similar to Fig 3.