Technologies such as transcriptional profiling, genome sequencing, epigenetic, proteomic or metabolomic profiling have resulted in large sets of genome wide data. However, most of these datasets are of correlative. Functional and phenotypic annotation of the genome is thus a key challenge for a fundamental understanding of physiology and disease pathogenesis. We combine genetic model organisms and repairable mutagenesis with in vivo mouse genetics and human cohort studies to functionally characterize candidate breast cancer genes. Using mouse genetics, we showed that RANKL and its receptor RANK are critical regulators of sex hormone and BRCA1 mutation-driven breast cancer, providing a rationale for cancer prevention trials.