Identification of breast cancers which are deficient in homologous recombination (HR) is clinically important, as these patients are more responsive to specific therapeutic interventions. The homologous recombination deficiency (HRD) score, consists of the sum of three DNA-derived measures of genomic instability (HRD-LOH, LST and NtAI). A high HRD score (greater or equal to 42) has been developed as a predictor of HR deficiency and has the potential to be utilised as a predictor of sensitivity to platinating agents and PARP inhibitors. The HRD score has mostly been derived using array-based platforms; however, given the increasing use of next generation sequencing, in particular whole genome sequencing (WGS), in the research and clinical setting, it is important to characterise how the HRD score derived from array and WGS compare. Array and WGS data was obtained from a familial breast cancer cohort (n=78). The ASCAT algorithm was then used to derive copy number data. Finally, the scarHRD package, within the R statistical computing environment, was used to calculate the HRD-LOH, LST, NtAI and HRD score from copy number data. Overall, we observed a good overlap between the distribution of HRD score and its components derived from array or WGS, with the exception of the LST; this component’s distribution was significantly different between array and WGS (p < 0.05). A high correlation was observed between the array- and WGS-derived HRD score and HRD score components (Pearson’s correlation coefficient: 0.79 for the LST,0.80 for the HRD-LOH,0.88 for the NtAI and 0.90 for the overall HRD score). The data generated demonstrates that a good agreement exists between array- and WGS-derived HRD score and HRD score components, suggesting that these methods are interchangeable when deriving the HRD score.