Skip to main content
Fig. 1 | Genome Biology

Fig. 1

From: Topconfects: a package for confident effect sizes in differential expression analysis provides a more biologically useful ranked gene list

Fig. 1

Results of simulation 1. a Proportion of top genes correct by various ranking methods in the top 20, 100, and 500 genes. Where genes were correct only because the ranking method fell back to ranking by limma p value, this is shown in gray. b Achieved FCR and FDR for different LFC thresholds by the confect method for a target FDR or 5%. Below the graphs, “LFC threshold” is the threshold used to select a set of genes; “True genes” is the number of genes with LFC truly exceeding this threshold. “Correct coverage” is the number of genes discovered and with true LFC correctly covered by the confidence bound. “False coverage” is the number of genes discovered but with true LFC outside the confidence bound. The sum of these last two numbers is the total number of discoveries at that threshold. All results are averaged over 100 simulations

Back to article page