ABSTRACT Metabolomics technique holds promise for better understanding the role of diet in cardiovascular disease (CVD) epidemiology and prevention. In particular, a multi-fluid multi-metabolite approach combining urinary and plasma metabolomics will likely enhance the accuracy and precision of nutritional biomarkers, with strong potential to develop novel tools for dietary assessment in precision nutrition research. This competing renewal is aimed to examine the effects of the randomized Mediterranean diet (MedDiet) interventions on urinary metabolite levels and to assess dietary biomarkers' predictive ability for future CVD risk in the landmark PREDIMED trial. We will assess urinary metabolomics in relation to CVD through a case-cohort design, including a representative subcohort randomly sampled from all PREDIMED participants (10%, n=779), and incident CVD cases (including myocardial infarction (MI), stroke, heart failure, and cardiovascular deaths) through December 2017 (n=529). The proposed Specific Aims are: (1) To identify specific urinary metabolites and multi-metabolite signatures of the MedDiet pattern and its main food components; (2) To examine the effects of the randomized dietary interventions on changes in the urinary multi-metabolite signatures of the MedDiet pattern and its major food components from baseline to year 1 (post intervention); (3) To investigate whether baseline and 1-year changes in urinary multi-metabolite signatures of the MedDiet pattern and of individual foods are associated with subsequent risk of CVD; and (4) To develop a multi-fluid (urine and plasma) multi-metabolite signature that robustly predicts risk of CVD and assess whether 1-year changes in the signature mediate the effects of the dietary interventions on the composite CVD outcome. As a sub-aim, we will compare the predictive roles of plasma and urinary metabolites of the MedDiet pattern on CVD risk. This ongoing grant started in 2013 and was renewed in 2017. To date, 36 papers have been published or submitted (including 20 papers in the current grant cycle). The PREDIMED metabolomics data have been used as a resource for replicating metabolomics analyses of CVD outcomes in several US cohorts. This competing renewal application will extend our long-standing research on MedDiet and CVD to multi-fluid multi- metabolite profiling in a large randomized intervention trial. This project leverages numerous strengths of the PREDIMED trial, a multi-disciplinary and highly productive team, cutting-edge metabolomics technologies, and high-dimensional data analytics. The identified dietary biomarkers will enhance the quality and rigor of nutrition research and strengthen the evidence base for developing dietary recommendations and nutrition policies. The proposed work will facilitate the precision nutrition research agenda outlined in the 2020-2030 Strategic Plan for NIH Nutrition Research.