Skip to main content

Table 2 Simulation study.

From: Time-course analysis of genome-wide gene expression data from hormone-responsive human breast cancer cells

 

Noise model N

Noise model T5

Noise model T3

Method

Rej.4

Corr.5

FDR6

FNR7

Rej.4

Corr.5

FDR6

FNR7

Rej.4

Corr.5

FDR6

FNR7

EDGE1 0.01

383.8

382.4

0.004

0.064

405.6

403.4

0.005

0.062

462.6

460.8

0.004

0.057

EDGE1 0.001

207.8

207.4

0.002

0.081

183.6

183.6

0.000

0.083

187

187

0.000

0.083

timecourse 2

775.6

733.4

0.054

0.029

794.4

690.6

0.131

0.034

869.8

733.4

0.157

0.029

BATS3 N, 9

775.6

775.6

0.000

0.024

794.4

782.6

0.015

0.024

869.8

803.4

0.076

0.022

BATS3 N, 12

775.4

775.4

0.000

0.024

794.2

782.4

0.015

0.024

869

802.6

0.076

0.022

BATS3 D, 9

753.2

753.2

0.000

0.027

774.8

762.8

0.015

0.026

875.4

793.4

0.094

0.023

BATS3 D, 12

745.8

745.8

0.000

0.027

768.4

756.4

0.016

0.026

871.6

789.2

0.095

0.023

  1. 1q-value threshold.
  2. 2Number of rejected chosen equal to the case of BATS (N,9), for comparison purpose.
  3. 3Error model N = normal, D = double-exponential, the indicated number is the value of λ.
  4. 4Rej. (Rejected) = average number of genes declared differentially expressed.
  5. 5Corr. (Correct) = average number of the correctly rejected hypotheses.
  6. 6FDR (False Discovery Rate) = average proportion of falsely rejected hypotheses over the total number of rejected hypotheses.
  7. 7FNR (False Negative Rate) = average proportion of false negatives over the total number of not rejected hypotheses.
Datasets generated with 1000 true signals, with three different noise models. Results were averaged over 5 datasets.
Back to article page

BMC Bioinformatics

ISSN: 1471-2105

Contact us