The challenge of developing an atlas that catalogs all the functionally important genomic changes associated with the development of luminal-type breast cancer is discussed in this article. The development of genome-wide techniques such as expression profiling, array-based comparative genomic hybridization and unbiased sequencing have put a cancer genome atlas within reach. However these techniques have revealed that the somatic DNA alterations associated with the development of a common solid tumor such as breast cancer are extremely complex. For example, large scale tumor DNA resequencing projects, focused on a small number of cell lines and the analysis of many genes, suggest that as many as 100 somatic mutations may have accumulated by the time a diagnosis is made. Similarly, array comparative hybridization experiments have uncovered multiple gene amplification and deletion events. Dealing with this complexity requires access to tumor and matched normal DNA from a large number of cases, with sufficient material to complete a spectrum of analytical techniques. Second, an acceptable approach to patient consent or sample de-identification must be in place if DNA sequencing traces are to be entered into public databases. Third, samples must be linked to detailed information on disease outcomes in order to identify lesions associated with aggressive clinical behavior. We conclude that samples from neoadjuvant endocrine therapy clinical protocols offer the best sample sets to initiate a luminal breast cancer genome atlas because these studies are amongst the few in which investigators have obtained high quality frozen tumor samples associated with both short term information on the estrogen dependence of individual ER+ tumors, as well as conventional data on long-term cancer survival.