Abstract
Biomarkers hold promise for increasing success rates of clinical trials. Biomarker discovery requires searching for associations across a spectrum of data. The field of biomedical data integration has made strides in developing management and analysis tools for structured biological data, but best practices are still evolving for the integration of high-throughput data with less structured clinical data. Integrated repositories are needed to support data analysis, storage and access. We describe a data integration strategy that implements a clinical and biological database and a wiki interface. We integrated parameters across clinical trials and associated genetic, gene expression and protein data. We provide examples to illustrate the utility of data integration to explore disease heterogeneity and develop predictive biomarkers.