Background Genetic testing for BRCA1/2 germline mutations in hereditary breast/ovarian cancer patients requires screening for single nucleotide variants, small insertions/deletions and large genomic rearrangements (LGRs). These studies have long been run by Sanger sequencing and multiplex ligation-dependent probe amplification (MLPA). The recent introduction of next-generation sequencing (NGS) platforms dramatically improved the speed and the efficiency of DNA testing for nucleotide variants, while the possibility to correctly detect LGRs by this mean is still debated. The purpose of this study was to establish whether and to which extent the development of an analytical algorithm could help us translating NGS sequencing via an Ion Torrent PGM platform into a tool suitable to identify LGRs in hereditary breast-ovarian cancer patients. Methods We first used NGS data of a group of three patients (training set), previously screened in our laboratory by conventional methods, to develop an algorithm for the calculation of the dosage quotient (DQ) to be compared with the Ion Reporter (IR) analysis. Then, we tested the optimized pipeline with a consecutive cohort of 85 uncharacterized probands (validation set) also subjected to MLPA analysis. Characterization of the breakpoints of three novel BRCA1 LGRs was obtained via long-range PCR and direct sequencing of the DNA products. Results In our cohort, the newly defined DQ-based algorithm detected 3/3 BRCA1 LGRs, demonstrating 100% sensitivity and 100% negative predictive value (NPV) (95% CI [87.6–99.9]) compared to 2/3 cases detected by IR (66.7% sensitivity and 98.2% NPV (95% CI [85.6–99.9])). Interestingly, DQ and IR shared 12 positive results, but exons deletion calls matched only in five cases, two of which confirmed by MLPA. The breakpoints of the 3 novel BRCA1 deletions, involving exons 16–17, 21–22 and 20, have been characterized. Conclusions Our study defined a DQ-based algorithm to identify BRCA1 LGRs using NGS data. Whether confirmed on larger data sets, this tool could guide the selection of samples to be subjected to MLPA analysis, leading to significant savings in time and money.