The translational research enterprise requires bi-directional sharing of data, knowledge, and information between researchers in the biosciences and those in clinical disciplines. Informatics efforts in translational research have focused largely on developing automated methods to correlate the results of genomics, proteomics, or mechanistic assay studies with available data on diagnosis treatment, and outcomes. Often, the latter set of measures involves crude categories, such as ‘cancer’ or ‘no cancer,’ because further details about a patient’s health or performed interventions are lacking. Such limitations create problems of specificity for findings in translational research. There are recent efforts to gather clinical information through standardized Electronic Medical Record representations that include genetics and genomics data [Hoffman 2007]. Such passive observations may yield biological insights into the mechanism of human disease and therapeutics. However, formalized controlled experiments, particularly human clinical trials, are necessary to address potential biases in biomarker analysis [Ransohoff 2005]. As a result, researchers are proposing and undertaking trial designs that include adjunct high throughput assays or that directly evaluate biological hypotheses. We refer here to such studies as translational clinical trials.
