Differential Expression Analysis in RNA-Seq by a Naive Bayes Classifier with Local Normalization

Table 3

AUC values of six methods on the training data set with different number of replicates.

Method	1	2	3	4	5	6	7
False positive test, no-call as NDE
edgeR	0.8535	0.8607	0.8595	0.8591	0.8584	0.8574	0.8567
DESeq	0.6621	0.8626	0.8622	0.8618	0.8613	0.8610	0.8602
Cuffdiff	0.5963	0.7904	0.7772	0.7904	0.7773	0.7748	0.7740
NOIseq	0.8334	0.8392	0.8425	0.8445	0.8452	0.8456	0.8460
Gfold	0.8870	0.8334	0.8871	0.8875	0.8874	0.8886	0.8886
GExposer	0.8968	0.9016	0.9024	0.9024	0.9028	0.9032	0.9030
False negative test, no-call as DE
edgeR	0.7845	0.7903	0.793	0.7934	0.7936	0.7934	0.7945
DESeq	0.5800	0.7871	0.7895	0.7902	0.7912	0.7905	0.7909
Cuffdiff	0.5894	0.7674	0.7645	0.7674	0.7623	0.7606	0.7610
NOIseq	0.7269	0.7342	0.7335	0.7308	0.7288	0.7276	0.7267
Gfold	0.7905	0.6702	0.7498	0.7670	0.7753	0.7753	0.7790
GExposer	0.7942	0.8015	0.8045	0.8051	0.8053	0.8051	0.8054