Germline mutations in BRCA1/2 increase the lifetime risk for breast and ovarian cancer dramatically. Identification of such mutations is important for optimal treatment decisions and pre-symptomatic mutation screening in family members. Although current DNA diagnostics is able to identify many different mutations, it remains unclear, how many BRCA2-associated breast cancer cases remain unidentified as such. In addition, mutation scanning detects many unclassified variants (UV) for which the clinical relevance is uncertain. Therefore, our aim was to develop a test to identify BRCA2-association in breast tumors based on the genomic signature. A BRCA2-classifier was built using array-CGH profiles of 28 BRCA2-mutated and 28 sporadic breast tumors. The classifier was validated on an independent group of 19 BRCA2-mutated and 19 sporadic breast tumors. Subsequently, we tested 89 breast tumors from suspected hereditary breast (and ovarian) cancer (HBOC) families, in which either no BRCA1/2 mutation or an UV had been found by routine diagnostics. The classifier showed a sensitivity of 89% and specificity of 84% on the validation set of known BRCA2-mutation carriers and sporadic tumor cases. Of the 89 HBOC cases, 17 presented a BRCA2-like profile. In three of these cases additional indications for BRCA2-deficiency were found. Chromosomal aberrations that were specific for BRCA2-mutated tumors included loss on chromosome arm 13q and 14q, and gain on 17q. Since we could separate BRCA1-like, BRCA2-like, and sporadic-like tumors, using our current BRCA2- and previous BRCA1-classifier, this method of breast tumor classification could be applied as additional test for current diagnostics to help clinicians in decision making and classifying sequence variants of unknown significance.