Differences in mechanisms of action between paclitaxel, doxorubicin and cisplatin in breast cancer may be accompanied by a corresponding diversity in gene expression patterns. We analyzed gene expression patterns in the MCF-7 cell line at different times (6, 24, 48 and 72 h) after treatment with paclitaxel, doxorubicin or cisplatin, using complementary microarrays representing 1,308 human complementary DNAs. The microarrays were created from six suppression subtractive hybridization libraries, representing complementary DNAs to messenger RNAs that were either up- or downregulated in MCF-7 cells after treatment with either paclitaxel, doxorubicin or cisplatin (IC90 concentrations). We used a hierarchical clustering method to group genes on the basis of similarity between expression patterns in each treatment. Two striking aspects of the patterns were evident. First, the expression pattern of each drug was unique for that drug. Fewer than 2% of the 1,308 human cDNAs present on the microarray were differentially expressed for all three drugs. Second, within each unique drug expression pattern, distinct clusters could be detected. It was clear from time-course hybridizations that clusters could be divided into early-, intermediate- and late-response genes. Only a small percentage of the fragments are common to the cellular injury produced by any of these drugs. The majority of the changes in mRNA expression observed with any one drug were unique to that drug, providing the foundation for a clinically applicable technology to develop a breast cancer response chip.