Gene therapy has often been described as a “black box” technique; that is, it is not known what happens to the cell after vector-mediated gene transfer. Although expression levels of the transgene are monitored, very little is known about how cellular gene expression is affected by the transduction process or transgene expression. Perturbations of cellular gene expression may be deleterious, even harmful, in several ways and could decrease the efficacy of therapy by altering the expression of the therapeutic gene. Studying the global changes in gene expression of cells transduced with different viral vectors is necessary for understanding the biology of gene therapy, and should yield safer, more effective therapies. To analyse these potential changes in gene expression more comprehensively, we are using the recently described cDNA microarray to look at cells transduced with a neoR-expressing retroviral vector. The cDNA microarray consists of numerous cDNAs representing genes and ESTs spotted individually on a microscope slide. Total RNA from transduced and untransduced cell populations are reverse transcribed with a fluorescently labelled nucleotide to produce differently labelled cDNA probes. These probes are mixed and hybridized to the cDNA targets on the microarray, which is then scanned to record fluorescence intensities for each probe at each target. A ratio of these intensities is used to identify expression differences between the cell populations, and signal intensity can approximate RNA quantity. This technique using the 1.4 K array allows us to look at the expression of more than 1,200 host genes at one time, and in so doing, screen for alterations caused by transduction or the expression of this commonly used bacterial reporter gene. Preliminary data indicate that the expression of several genes is significantly altered when cells are analysed immediately after selection. When these same cell populations are removed from growth in G418-containing selection medium and cultured for several months, significant changes in gene expression were not observed. These studies will help to address potential safety concerns as well as provide information to be used as a basis for designing better vectors.