Abstract : Longs segments ( 1 megabase) of homozygous DNA are common in the genomes of outbred human populations. Several lines of research indicate that elevated genome-wide homozygosity and runs of homozygosity (RoH) at specific loci may increase breast cancer risk. In this project, we determine if genome-wide RoH content and individual RoHs are more common in breast cancer cases than in controls, using logistic regression methods. Using genome-wide SNP data (525,000 SNPs) on 1,647 non-Hispanic white, early-onset premenopausal breast cancer cases and 1,556 matched controls we identified over 65,000 individual RoHs and 423 genomic regions harbor RoHs for at least 10 individuals in our dataset (i.e., RoH regions ). Overall RoH content was not associated with breast cancer status or with subtypes of breast cancer as defined by estrogen receptor status. Furthermore, analyses of each of the 423 RoH regions did not reveal any region in which RoH status was significantly associated with breast cancer risk or risk for a breast cancer subtype (after correction for multiple testing). Finally, comparing the RoH regions showing the strongest associations in our study to the regions with the strongest association in a prior study of RoHs in breast cancer did not reveal any common findings across studies. In this association study of RoHs and early-onset breast cancer risk, we have not implicated overall RoH content or any individual RoH in breast cancer risk. We are currently pursuing sequencing of candidate genes believed to be important for risk for ER-negative breast cancer.