The goal is to understand the critical events in tumour development and to apply this understanding to new approaches to diagnosis, prevention and treatment. It is clear that breast cancer is a heterogeneous disease at the molecular level, raising the possibility of a future functional classification based on mechanisms rather than morphology. These molecular phenotypes will also confer predictive value on the potential of the tumour to invade, metastasise and respond to or resist new therapeutic strategies. Studies of the genome in individuals are predicted also to enable the identification of polymorphisms that are associated with increased susceptibility to environmental factors, in addition to possibly explaining de novo variations in responses to drugs and radiation. The difficulty is how to identify which, of the approximately 30,000 genes expressed by a typical cancer cell alone or in combination, are the ones involved in these processes. The majority of breast cancers have such a multitude of molecular changes that it is difficult to distinguish between those that are critical to tumour progression and those that are epiphenomena of genetic instability and abnormalities in DNA repair. The identification of the earliest events in carcinogenesis must be the best hope, as it will then be possible to target the events that predispose to other secondary changes before they occur. Genomics and proteomics is the current hope to take us forward. This involves the application of a number of new technologies to facilitate the profiling of individual tumours, including laser-guided microdissection of microscopic lesions, comparative genomic hybridisation and loss of heterozygosity analysis of DNA using microarray technology to study DNA and expressed RNAs and protein profiling using 2D gel mass spectroscopy. With over 100,000 mRNAs and proteins to examine in complex tissues and in various combinations, there is obviously going to be a requirement for a large investment in computing power (bioinformatics) to facilitate the analysis of these data in relation to the clinical characteristics of the individual tumour and the patient.