Fig. 1: Associations with specific polygenic risk score (PRS) percentiles.

The PRS percentile thresholds were determined in the sets of unaffected carriers for the disease under assessment. Table 2 shows the estimated hazard ratios (HRs). The black curve represents the expected HRs assuming the per standard deviation HR estimates in BRCA1 and BRCA2 carriers based on the continuous PRS models (Table 1). (a) PRS_ER- percentile-specific associations with breast cancer risk for BRCA1 carriers. The red curve represents the expected HRs over the PRS percentile distribution, assuming the per SD odds ratio (OR) estimate from the population-based validation studies from Mavaddat et al.¹² (OR = 1.45 per PRS_ER- standard deviation). (b) PRS_BC percentile-specific associations with breast cancer risk for BRCA2 carriers. The red curve represents the expected HRs over the PRS percentile distribution, assuming the per SD OR estimate from the population-based validation studies from Mavaddat et al.¹² (OR = 1.61 per PRS_BC standard deviation). (c) PRS_HGS percentile-specific associations with ovarian cancer risk for BRCA1 carriers. (d) PRS_HGS percentile-specific associations with ovarian cancer risk for BRCA2 carriers. The gray curve (c and d only) represents the theoretical HRs across the PRS distribution, calculated by assuming external single-nucleotide polymorphism (SNP) effect sizes and allele frequencies for SNPs contributing to the PRS. CI confidence interval, ER estrogen receptor, HGS high-grade serous.