Skip to main content
. 2015 Jun 13;16(1):455. doi: 10.1186/s12864-015-1676-0

Fig. 3.

Fig. 3

Type I error and false discovery rate. Data were simulated from large-scale TCGA lung cancer sequencing studies, a miRNA and b mRNA. Two different types of data transformation, “shift” and “scaling & shift” were applied. Boxplots summarize type I error and false discovery rate of different statistical methods for DE detection under a wide range of simulations. Methods in red font are those do not have correct type I error and/or false discovery rate