A Novel Algorithm to Test Significant Differences in Microarray Experiments

We present a novel algorithm to detect significant gene expression differences in microarray experiments. The algorithm was examined with the reference to the well-known t-test with duplicate control experiments and duplicate chemical-stimulation experiments. These methods are shown to be comparable. The standard deviation (SD) estimates of expression which is used to judge the significant differences is derived from the repeated experiments in the t-test, but in our algorithm, the estimate is given a priori as a function of expression levels. Although our algorithm requires the probabilistic model of SD, it is also applicable to a single pair of experiments (one control and one treatment).