Our lab has investigated the strengths and limitations of current biomarkers for risk prediction and cardiovascular screening. Our findings have contributed to a re-appraisal of approaches to the identification and statistical evaluation of cardiovascular biomarkers. We proposed that limitations of conventional biomarkers could be overcome in part by applying “less biased” methods and developing cardiac-specific profiles. We have applied molecular profiling to large populations for cardiovascular biomarker discovery. We have also led the application of metabolomics and other profiling methods to large-scale clinical studies. In 2011, we identified an amino acid-based metabolite profile that predicts type 2 diabetes mellitus more than a decade before its onset, a finding that has generated significant interest around the potential for metabolite profiling to identify predictive biomarkers. We subsequently identified additional biomarkers of cardiometabolic risk, studied the biological role of several biomarkers, and applied novel statistical methodologies for metabolomic and proteomic studies.