Skip to main content
Fig. 5 | BMC Bioinformatics

Fig. 5

From: Robust differential expression analysis by learning discriminant boundary in multi-dimensional space of statistical attributes

Fig. 5

Compare the curves of the true positives vs. the target FDR (6 vs. 6 and G 500500 ). The x- and y- axes indicate the target FDR cutoff and the average number of true positives, respectively. The solid curve with blue circle markers represents DCT+R+NB and other curves represent non-DC methods. The result of a method at a particular target FDR is shown in this plot if (1) its average true FDR does not exceed the target FDR by 10%; and (2) its average number of true DEFs is ≥0.5 (rounds up to 1). DC was able to meet all target FDR cutoffs. The results of voom and vst/limma are almost the same

Back to article page