Skip to main content

Table 1 Simulated data error models

From: Comparison of small n statistical tests of differential expression applied to microarrays

Variance Model Expression Intensity (x i j ) Random Error Parameters
Common μ i + ϵ i ϵ i ~ N (0, c) c (.1, .05, .01)
Inverse Gamma μ i + ϵ i ϵ i ~ N (0, IG(3, b)) b (.2, .1, .05)
Local log2( μ i e ϵ p i MathType@MTEF@5@5@+=feaagaart1ev2aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacPC6xNi=xH8viVGI8Gi=hEeeu0xXdbba9frFj0xb9qqpG0dXdb9aspeI8k8fiI+fsY=rqGqVepae9pg0db9vqaiVgFr0xfr=xfr=xc9adbaqaaeGaciGaaiaabeqaaeqabiWaaaGcbaGaeqiVd02aaSbaaSqaaiabdMgaPbqabaGccqWGLbqzdaahaaWcbeqaambvLH1qn1uy0Hws0fgBPngaryWyT1wAXadaiqGacqWF1pGSdaWgaaadbaGaemiCaaNaemyAaKgabeaaaaaaaa@3D7C@ + ϵ ai ) ϵ ai ~ N (0, c)
ϵ pi ~ N (0, .1)
c (50, 25, 5)
  1. The Expression Intensity column shows how the replicate intensity values were generated for each transcript. The Random Error column shows the method of calculating specific ϵ values. The Parameters column shows the values that were used to generate high, medium, and low variances, respectively.