The information contained in the draft sequence of the human genome offers a solid basis to study gene function. The DNA microarray technology is one of the most important tools enabling the performance of such studies. The global expression analysis using microarrays enables the construction of a simultaneous expression pattern of thousands of genes and thus an unprecedented opportunity to obtain molecular signatures of the state of activity of diseased cells. Microarray analysis may thus provide invaluable information on disease pathology, progression, resistance to treatment, and response to cellular microenvironments. Ultimately such analyses may lead to improved early diagnosis and innovative therapeutic approaches. Yet, by employing this advanced technology in order to identify new target molecules for diagnosis or therapy, one is faced with the necessity to focUs on one or several genes and to investigate them further. A huge number of differentially expressed genes are usually identified when 2 cell samples are comparatively analyzed in an individual assay. Several factors, mainly a possible genetic diversity between the samples, minimize the capability of reaching an educated conclusion as to which gene or genes to select for further studies. Gene expression arrays were employed to identify genes whose expression characterizes leukemic leukocytes. Many different genes were differentially expressed by "leukemic-phase leukocytes," derived from three untreated acute myelogenous leukemia (AML) patients and by "remission-phase leukocytes," obtained from the same patients following induction of remission. However, only two of these genes, the dual-specificity MAPK phosphatase PYST2 and the tryptophan 5-hydroxylase, were found to be more highly expressed by the leukemic-phase than by remission-phase leukocytes of all three patients. The microarray results of only the PYST2 gene could be verified by RT-PCR. By focusing on genes that had a similar expression pattern in cells from multiple donors we overcame the problem of genetic diversity and identified, out of 60 differentially expressed genes, a single candidate target gene in AML. Pooling specimens in order to overcome the problem of genetic diversity (e.g., clinical material obtained from treated patients and from untreated ones), is not recommended because it may dilute the treatment effects. Therefore, it is essential to perform each experiment using individual preparations.