Breast cancer is a heterogeneous disease and its consequent complexity is a major challenge for physicians and biologists. Notwithstanding its potential curability due to the availability of treatment modalities which are effective in the presence of favourable clinical or pathobiological features, there is still a great deal of controversy over its clinical management. In recent decades, tumour biomarkers that are indicative of or related to cell traits characterising malignancy--that is self-sufficiency in proliferative growth signals, insensitivity to growth inhibitory signals, evasion of apoptosis, limitless replicative potential, and activation of pathways leading to neo-angiogenesis, invasion and metastasis--have provided information that has been proven to be associated with disease progression. However, when these biomarkers have been analysed individually, their prognostic relevance has been found to be modest, the only remaining clinically useful biomarkers being cell proliferation and plasminogen activation-related factors for prognosis, and steroid hormone receptors and the oncogene HER2/neu for prediction of response to hormonal therapy or to the novel targeted anti-HER2/neu therapy. It therefore remains necessary to reduce the intrinsic complexity of breast cancer in order to improve its clinical outcome. One way to achieve this objective derives directly from the concept that cancer is a genetic disease at the somatic level and from the recent availability of high-throughput post-genomic analytical tools such as gene and protein expression techniques for global gene expression analysis. Following these novel approaches, a number of recent studies have produced gene expression profiles in breast cancer that are markedly associated with disease progression and directed to answer different clinical and biological questions. However, the outcome of these novel studies still needs to be validated, which will entail cooperation between different specialists and integration of all the different skills involved in translational research in oncology.