Genetically and pathologically accurate mouse models of leukemia and lymphoma have been developed in recent years. Adoptive transfer of genetically modified hematopoietic progenitor cells enables rapid and highly controlled gain- and loss-of-function studies for these types of cancer. In this Commentary, we discuss how these highly versatile experimental approaches can be used as biological filters to pinpoint transformation-relevant activities from complex cancer genome data. We anticipate that the functional identification of genetic 'drivers' using mouse models of leukemia and lymphoma will facilitate the development of molecular diagnostics and mechanism-based therapies for patients that suffer from these diseases.