Skip to main content

Table 3 Reproducibility of each test statistic when used on replicate datasets.

From: Improving the statistical detection of regulated genes from microarray data using intensity-based variance estimation

   Mavg t d90th percentile dSAM Z(Iavg) Z(Imin)
R 2 Dataset 4 #1 0.89 0.00 0.86 0.87 0.93 0.94
  Dataset 4 #2 0.90 0.00 0.87 0.87 0.93 0.93
  Dataset 4 #3 0.89 0.00 0.88 0.88 0.95 0.95
Slope Dataset 4 #1 0.90 0.00 0.97 1.63 1.00 1.00
  Dataset 4 #2 0.93 0.01 0.96 0.58 1.02 1.04
  Dataset 4 #3 1.00 0.07 0.96 0.74 1.09 1.08
  1. Linear regression slope coefficients and R2 coefficients are calculated between corresponding statistics from two replicate three-array subsets of Dataset 4. Columns represent (left to right): the average logged ratio Mavg, the t-statistic, the 90th percentile penalized statistic, the SAM penalized statistic, the Z-statistic using Iavg pooling metric and the Z-statistic using Imin for three different pairs of subsets. Although Mavg is not a statistical test, it is included in this table for comparison.