A Distribution Free Summarization Method for Affymetrix Genechip Data Preprocessing Analysis

Motivation: Affymetrix GeneChip brand arrays require a summarization step in order to combine the information in a probe set into one value representing the expression level of the corresponding gene. Here we present a new summarization method, Distribution Free Weighted (DFW) fold change, that uses the information of fold change but does not make any distributional assumptions for the data. Results: Based on spikein data sets, we compare DFW with several popular methods, via both our own calculations and the ‘Affycomp II’ competition. The results show that DFW outperforms other methods when sensitivity and specificity are considered simultaneously. In fact, the area under the Receiver Operating Characteristic (ROC) curve for DFW is nearly 1.0 (a perfect value). Furthermore, DFW can obtain all the true positives with a small number of false positives. It is also computationally faster than most methods in current use. Availability: The R package for DFW is available upon request. Contact: [email protected]