Figure 4.

FDR estimation comparing different methods. Results were averaged across 100 datasets simulated based on Horbinski data (left column) and Mori data (right column) with three, five or eight replicates.