Skip to main content

Table 2 Generating Reference Standards with r-DEG methods in datasets with replicates Standards

From: Evaluating single-subject study methods for personal transcriptomic interpretations to advance precision medicine

Method

Yeast (n = 30 paired samples), genome size = 7126 genes

MCF7 (n = 4 paired samples), genome size = ~ 22,000 genes

FDR Threshold

Number of DEGs

Percent of Genome as DEG

FDR Threshold

Number of DEGs

Percent of Genome as DEG

edgeR

.05

4437

62%

.005

3231

14%

DESeq

.05

4594

64%

.001

3207

14%

DESeq2

.05

4802

67%

.0005

3255

15%

DEGseq

.05

5087

71%

3.56e-12

3351

15%

NOISeq

.05

3914

55%

0.078

3397

15%

Intersection of all methods

n/a

3118

44%

n/a

1173

5%

Union of all methods

n/a

6425

90%

n/a

6039

27%

  1. FDRs are adjusted to obtain lists of DEGs of the same length as reported in the original publications. As shown with the intersection of all DEGs predicted by distinct methods, determining a gold standard in RNA-Seq analyses of multiple biological isogenic replicates remains a challenge
  2. n/a not applicable