1500 Background: Epigenetic regulatory pathways are intensely studied for their involvement in breast tumorigenesis, however little is currently known about the genetic variation in epigenome components contributing to the risk and/or prognosis of breast cancer. In this study we have tested how the novel germline genetic signatures identified in epigenetic regulatory genes (ERGs) may potentially contribute to clinical prediction of breast cancer risk. Methods: We have genotyped 711 SNPs tagging 87 ERGs in 1985 breast cancer cases and 1609 controls, using Sequenom i-Plex. The samples were of white European origin with the fraction of Ashkenazi Jewish (AJ) ancestry (n=1642). The association of SNPs with breast cancer risk was assessed using logistic regression, adjusted by age, AJ status and estrogen-receptor (ER) status. The predictive ability of SNP signatures was tested by ROC curves using logistic regression fitting the SNP/clinical covariate models, and the area under the curve (AUC) was used to assess their utility in the classification of breast cancer risk. Results: We have identified the signature of 20 SNPs tagging 13 ERGs, significantly associated with breast cancer risk. The strongest association has been observed for RUNX1 (rs7280097, OR=0.83, CI 95%: 0.71-0.94, p=0.006) and PRDM16 (rs12135987, OR=1.22, CI 95%: 1.06-1.42, p=0.007). The inclusion of predictor variables (age, AJ status, ER status) and 20 associated SNPs in logistic regression ROC curve analysis yielded in best fitting model involving 10 SNPs tagging 8 ERGs with AUC of 0.723, compared to 0.660 with predictor variables alone (p=0.003). Conclusions: We have identified a signature of 20 SNPs in epigenetic regulatory genes (20-SNP-ERG) significantly associated with breast cancer risk. In addition, the incorporation of 10 SNPs from 20-SNP-ERG into risk prediction model increases the ability to classify breast cancer risk in addition to other clinical and demographic covariates. The results suggest the promising clinical potential of 20-SNP-ERG signature in identification of high-risk individuals in the population, and point to possible biological implication of germline patterns in epigenetic enzymes in breast tumorigenesis.