In studies on the onset and progression of cancer, molecular biology has proven to be very useful. Reasoning that tumorigenesis can be considered as a multistep process in which a normal cell gradually escapes from its delicately regulated growth pattern, one might describe this process in terms of gene expression. The group of genes that are good candidates for such studies are the proto-oncogenes. Recent progress in oncogene research, however, has made us believe that the group of oncogenes is much greater than the 50 so far characterized, since all growth factor-, growth factor receptor-, signal transducer- and transcription-regulating genes have a potency to become oncogenic. Alternatively, recessive genes, such as tumor suppressor genes, might be equally relevant to the onset and progression of cancer. Therefore, new approaches in molecular studies on the genesis and progression of malignancies are urgently needed. One of these methods is the so-called differential or subtraction hybridization analysis. Particularly, the rationale of differential hybridization and the first applications to prostate cancer research are discussed in this paper.