One of the main purposes in analysis of microarray experiments is to identify differentially expressed genes under two experimental conditions. The Meta-analysis method, rank product meta-analysis approach, considered a powerful tool for identification of differentially expressed genes. However, rank product meta-analysis approach used the each dataset in the computation of the fold changes, which leaded to less computational efficiency. Here we modified the rank product meta-analysis approach to obtain an improved model for identifying different gene expression. The new model, grouping rank product approach, adds competitive classification of samples to group datasets before the computation of the fold changes. We used the grouping rank product approach on two simulated datasets and two breast datasets and showed that the grouping rank product approach is not only as accurate as the rank product meta-analysis approach, but also more computational efficient in identifying differentially expressed genes.
ON THE PERMUTATION DISTRIBUTION OF THE RANK PRODUCT-MOMENT STATISTIC
