Skip to main content

Table 1 Comparison among different normalization methods based on simulated normal microarray data with the Balagurunathan's method. The data sets are simulated without or with dye bias and include 1000 genes in 10 slides with 5% differentially expressed genes. The ratio of up-regulated to down-regulated genes is 1:1. The additional file 3 contains the full data sets.

From: Using Generalized Procrustes Analysis (GPA) for normalization of cDNA microarray data

Method

Without dye bias

With dye bias

 

ν (1)

β (1)

ν

β

Raw

0.1182

0.004627

0.1231

0.004586

Global

0.1179

0.004732

0.1223

0.00473

Lowess

0.1143

0.005368

0.1182

0.004976

Scale

0.1113

0.004666

0.1201

0.004515

Quantile

0.1149

0.005226

0.1212

0.004716

VSN

0.06294

0.002581

0.06314

0.002541

GPA

0.02122

0.00117

0.02069

0.00119

Global+Scale

0.1117

0.004579

0.1191

0.004595

Global+Quantile

0.1149

0.005226

0.1212

0.004716

Global+GPA

0.02122

0.001174

0.02068

0.001156

Lowess+Scale

0.1088

0.005161

0.1153

0.005067

Lowess+Quantile

0.1133

0.005085

0.1198

0.004812

Lowess+GPA

0.02215

0.001279

0.02205

0.00134

  1. (1) ν and β : the median of variance and bias, respectively, of MSE.