Abstract Background: Identifying the missing hereditary factors underlying the familial risk of breast cancer could have a major and immediate impact on managing the breast cancer risk for these families. Methods: We identified candidate breast cancer predisposition genes through whole exome sequencing of BRCAx families and subsequently sequenced up to 1325 genes, along with 76 common low penetrance variants associated with breast cancer, in index cases from 6,000 BRCAx families and 6,000 cancer free women (ethnically matched on principal component analysis). Results: The role of recently described (PALB2) or suspected (MRE11A) moderately penetrant genes was confirmed. Conversely, the size of the cohort means that the absence of enrichment for loss of function (LoF mutations) provides strong evidence against other reported breast cancer genes (BRIP1, RINT1, RECQL). For further moderate risk variants (in CHEK2, ATM, BRCA2) we observed significant risk modification based on the polygenic risk score (PRS - calculated from the common variant data), with the risk restricted to the co-occurrence of the rare variant and high PRS. Novel candidate genes were identified based on LoF mutations, including NTHL1 (38 cases versus 15 controls, OR 2.5 p=0.002): a member of the base excision repair (BER) pathway. DNA sequencing of the breast carcinomas from 17 heterozygous NTHL1 mutation carriers revealed a strong bias towards a C:G>T:A (C>T) transitions, consistent with a BER defect, which confirmed the recent findings in colorectal carcinomas from bi-allelic NTHL1 mutation carriers. This data extends the cancer predisposition phenotype of NTHL1 to heterozygous carriers. In addition to NTHL1, there are a large number of candidate genes where the ratio of LoF mutations in cases versus controls indicates that they may convey an actionable level of risk; 46 genes (519 families) meet the basic criteria of multiple LoF variants and an OR >2 for cases versus controls – including previously proposed breast cancer genes MRE11A, BLM, MLH1, MYH, FANCD2 and functionally plausible candidates such as MLH3, PARP2 and ATR. Collectively the OR of breast cancer for LoF mutations in this group of genes is 3.3 (95% CI 2.7-3.9, P=3.5x10-41). Conclusion: Our data shows that the effect of rare variation in established and novel breast cancer genes, along with consideration of the background polygenic risk, together explains a substantial component of the heritable risk of breast cancer in our cohort. Citation Format: Campbell IG, Li N, Rowley S, Goode D, Devereux L, McInerny S, Grewal N, Lee A, Trainer A, Wong-Brown M, Scott R, Gorringe K, James P. The contribution of rare variants, polygenic risk, and novel candidate genes to the hereditary risk of breast cancer in a large cohort of breast cancer families [abstract]. In: Proceedings of the 2017 San Antonio Breast Cancer Symposium; 2017 Dec 5-9; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2018;78(4 Suppl):Abstract nr PD1-04.