Breast cancer, the most common cancer diagnosis among women worldwide, occurs in greater incidence among Caucasian American (CA) women compared to African American (AA) and Latina American (LA) women. However, AA and LA women are more likely to develop an aggressive form of cancer (i.e., triple- negative breast cancer (TNBC)) at an early age. Both ancestral admixture and genetic susceptibility have been linked to breast cancer risk, but most of the genetic variants identified so far are located in nonprotein encoding regions with undetermined functional significance. Thus, alternative approaches are critically needed to provide new insight to biological function and carcinogenesis. Intermediate phenotypes, such as metabolite profiles, may prove to be a useful alternative approach. Metabolites are end products of cellular regulatory processes, so their concentrations can be regarded as the ultimate net results of genetic and environmental interactions. We propose an innovative study of metabolomics involving a discovery component and a validation component to investigate the unexplained racial/ethnic disparity in breast cancer occurrence. The study is designed to address the overarching hypothesis that plasma metabolite profiles differ in CA, AA, and LA women and that metabolic signatures can explain some of the racial/ethnic differences in breast cancer risks. We further hypothesize that the metabolite-related breast cancer risks are modified by genetic susceptibility and ancestral admixture. In the discovery component, 900 breast cancer cases (300 each of CA, AA, and LA cases) and 900 race/ethnicity, age, and residence matched controls will be enrolled with collection of a pre-treatment blood sample for metabolite analysis and genotyping. Metabolomic profiling will be conducted at Metabolon, Inc. using GC/MS and LC/MS platforms. Genotyping will be conducted using the Illumina Custom GoldenGate OPA 1536-plex platform. Clinical data and tumor characteristics, including ER, PR, and HER2 expression status, will be extracted from the MD Anderson clinical database. Epidemiological data, including height and weight history, alcohol drinking, and family cancer history, will be extracted from the Patient History Database. This component will evaluate the role of metabolic profiles in racial/ethnic disparity in breast cancer risk and the extent to which this disparity may be explained by ancestral admixture and previous GWAS-reported genetic susceptibility. We will also evaluate whether the metabolite-breast cancer associations differ for TNBC and non-TNBC. The most promising 5-10 metabolites will be included for validation in two population-based prospective cohort studies (PLCO and Mexican American Cohort). The validation component, including 450 breast cancer cases (150 each of CA, AA, and LA cases) and 450 controls, will evaluate the associations of these metabolites with breast cancer risk and whether length of follow-up could modify the breast cancer associations. The proposed research will enable a critical narrowing of the gap in understanding the underlying biology of racial/ethnic differences in breast cancer.