Skip to main content
. 2018 Oct 31;13(10):e0206312. doi: 10.1371/journal.pone.0206312

Fig 5. Percentage of simulated datasets with the correct number of SVs detected, were significant SVs were determined using SVA (“BE” and “Leek”) and PCA (7%, 8% and 10% percent of variation).

Fig 5

Simulation with (A) binary (two-group) batch effect, (B) continuous batch effect, (C) a two-group batch effect and continuous batch effect.