To explore genes that determine the sensitivity of cancer cells to anticancer drugs, we investigated using cDNA microarrays the expression of 9216 genes in 39 human cancer cell lines pharmacologically characterized on treatment with various anticancer drugs. A bioinformatical approach was then exploited to identify genes related to anticancer drug sensitivity. An integrated database of gene expression and drug sensitivity profiles was constructed and used to identify genes with expression patterns that showed significant correlation to patterns of drug responsiveness. As a result, sets of genes were extracted for each of the 55 anticancer drugs examined. Whereas some genes commonly correlated with various classes of anticancer drugs, other genes correlated only with specific drugs with similar mechanisms of action. This latter group of genes may reflect the efficacy of each class of drugs. Therefore, the integrated database approach of gene expression and chemosensitivity profiles may be useful in the development of systems to predict drug efficacies of cancer cells by examining the expression levels of particular genes.