Experimental validation of methods for differential gene expression analysis and sample pooling in RNA-seq

BMC Genomics

Table 1 Validation of four differential gene expression analysis methods for RNA-Seq

Parameters^a	edgeR	Cuffdiff2	TSPM	DESeq2
Total number of identified DEGs^b	82	136	8	1
Number for DEGs selected for qPCR validation	51	79	8	1
Sensitivity (True positivity rate) (%)	76.67	51.67	5.00	1.67
Specificity (True negativity rate) (%)	90.91	12.73	90.91	100.00
False positivity rate (%)	9.09	87.27	9.09	0.00
False negativity rate (%)	23.33	48.33	95.00	98.33
Positive predictive value (%)	90.20	39.24	37.50	100.00
Negative predictive value (%)	78.13	19.44	46.73	48.25
Positive likelihood ratio	8.43	0.59	0.55	∞
Negative Likelihood ratio	0.26	3.80	1.05	0.98
Overall agreement (%)	83.48	33.04	46.09	48.70

^aReplication of differential expression by quantitative Polymerase Chain Reaction (qPCR) was the reference standard
^bDifferentially Expressed Genes, after Benjamini-Hochberg false discovery correction at 5 %; TSPM: Two-stage Poisson Model

ISSN: 1471-2164